|
| Greg's diary |
| January 2018 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Monday, 1 January 2018 | Dereel | Images for 1 January 2018 |
| Top of page | ||
| next day | ||
| last day |
|
Goodbye Nemo
|
Topic: animals | Link here |
Over 5 years ago we gave away Nemo, our black German Shepherd Dog, mainly because we were concerned about him getting hurt in the wild. We didn't forget him, though, and Yvonne in particular has been following his progress.
No more. Last week he got bitten by a snake, and died on Friday. Goodbye, Nemo. At least he had a good life, and he was nearly 8 years old.
|
How to sleep, Leonid style
|
Topic: animals | Link here |
When dogs sleep, it's usually on their side, with feet, head and tail on the ground. But Leonid has a special issue, his tongue:
|
How to friend strangers
|
Topic: technology, general, opinion | Link here |
A few days ago I received a fairly typical message:
A tourist guide in México? I don't know her. But we have a mutual friend. Who is it? A very well-known member of the Internet community. How did he get to know her? So I sent him an email (yes, it occurred to me later that I could have sent a Facebook PM, but I don't go in for that sort of thing). The reply:
Sometimes I think this happens far too often. Spam 100 people with friend requests, and one might accept, for whatever reason. Then she's a friend of a friend, and the others are much easier to “friend”.
|
Annex Palestine!
|
Topic: politics, opinion | Link here |
One of the more disturbing news items of the day barely made it to the Western press: the Likud Central Committee voted to annex parts, probably all, of the occupied West Bank of the Jordan river, stated as Judea and Samaria. It's easy to guess why they're doing this now, and it shows the far-reaching effects of Donald Trump's stupid declaration last month. It also goes in line with what David Friedman said a couple of days ago.
But why hasn't this been reported more widely? The US news sources I read (Washington Post and New York Times) haven't mentioned it, and neither has any Australian news source. Al Jazeera did, of course.
I don't believe in the concept of “fake news”. But it's somehow concerning how one-sided press coverage is. It makes you wonder what's really going on with the protests in Iran, where Western journalism is paying a lot of attention, though I doubt that they understand the situation.
|
ISO invariance
|
Topic: photography, opinion | Link here |
Today I discovered a new term: ISO invariance. What's that? I spent a couple of hours reading mainly vague pages about the subject, and I still don't know. Some of the better pages are: Sony Alpha 7R II: Real-world ISO invariance study and Noise, Dynamic Range and Bit Depth in Digital SLRs, the latter quite old and only marginally related to the subject. What I have understood so far is:
Much of this makes sense, and the pages I have read go into obsessive detail about the sources of noise and what can be done about it. But none of the pages that I have seen address the most obvious issue with underexposure: you lose shadow detail. Assuming that you underexpose by 6 EV, something that they advocate, your signal levels will drop from a range of 0 to 4,095 to a range of 0 to 63.
There's also a comparison page that I haven't been able to fathom. I downloaded some surprisingly big raw files, over 110 MB for 12 MP images, only to find that DxO PhotoLab doesn't understand the format.
So apart from the issue of gradation, there are two main things I still don't understand:
In passing, I also came across “Unlocking Olympus”, which might be interesting in other contexts. It didn't add to this investigation.
|
Happy New Year!
|
Topic: general, opinion | Link here |
It's been seven years since we stayed up until midnight for the New Year. Instead we went on to New Zealand time and drank our champagne at 22:00 local time.
Problem: we needed Chris Bahlo to help drink the champagne, and she was here only the day before. So we didn't even drink champagne last night.
But the New Year is long, so we decided that this time we would celebrate the arrival of the New Year with the Californians, and 19:00 local time today. That's dinner time, so Chris came over again, and we finally got rid of the champagne.
| Tuesday, 2 January 2018 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Some garden work
|
Topic: gardening | Link here |
I've been meaning to mow the lawn for I don't know how long, and today I finally got round to doing it. It wasn't easy: firstly, our ride-on mower is really too big for the small (400 m²) garden area to the north of the house, and secondly there are so many wildflowers growing there through the grass that I didn't want to mow them all down. So in the end I did about two-thirds of it. The result looks messy, but maybe I'll think of a way of improving it.
And then the weeds! We're still waiting for Warrick Pitcher, but I hadn't expected them to take over so quickly in just a few weeks. I should spend some time every day fighting them.
|
More ISO invariance insights
|
Topic: photography, opinion | Link here |
I still don't understand “ISO invariance”, and today's research helped little. This page claims:
The short answer is that ISO invariance means that a camera will produce the exact same image quality by staying at ISO (or whatever the base ISO is on the camera) and dramatically underexposing the photo and then brightening it up again in Lightroom, as if you had shot the camera at the proper ISO in the first place.
That completely ignores the bit depth of the resulting image, as does just about everything that I've seen so far. And the reference to Lightroom, by no means the only one, shows a certain lack of perspective: the technique should be independent of the processing software.
This page at least comes up with a definition:
An iso-invariant camera is one where read noise remains constant over the range of ISO values...
It also refers to this page (from photographylife), which defines it differently:
ISO invariance is a property of your camera’s sensor. If your camera sensor is perfectly ISO invariant, there is no penalty in noise if you brighten a photo in post-production rather than increasing your ISO in-camera. They are functionally the same.
Well, firstly they all manage to ignore the issue of bit depth, closely related to dynamic range. And it's difficult to find what the alternative is. And then this page states:
With an ISO-variant camera (one whose read noise decreases with increased ISO), the benefit is in favor of brightening with added in-camera ISO, which will typically result in less read noise than "shooting dark" and pushing in raw processing.
This suggests that the ISO-variant camera is better. In general, the literature appears full of contradictions. One insight that is plausible and useful is (again the photographylife page):
Digital amplification in-camera is no different from digital amplification by brightening the photo in Lightroom, except that [sic] decreases your dynamic range and makes it harder to retain detail in the highlights.
This is partially true, assuming that the missing word is something like “the former”. Retaining detail in the highlights has nothing to do with amplification, of course: it can happen without, and it's a question of exposure. The best statement I have seen is (still the same page):
Most photographers have heard that there is a difference between the “native” and “extended” ISO values in certain cameras. For example, according to Nikon, my D800E has a native range from ISO 100 to ISO 6400, and an extended ISO range that stretches down to ISO 50 and up to ISO 25,600.
And that is ISO invariance, at least at the simplest level – the concept that, in some sense, your ISO setting “doesn’t actually matter.” Whether you set ISO 6400 in your camera and brighten it in Lightroom, or you set ISO 25,600 in camera, you’ll end up with two files that have exactly the same amount of noise (though, again, the ISO 6400 photo may have more latitude in highlight recovery than the simulated ISO 25,600).
OK, that makes sense, though it assumes, of course, that the output of the ADC is no wider than the raw image bit depth. So if somebody says to me “don't go beyond 6,400 ISO, it won't help”, I can live with that. But that's not what (many) people are saying: they're saying “don't go beyond base ISO”, as in the first quote above.
An example is this page, which includes a series of photos of a standardized test sheet, which can be downloaded. But they've done it wrong! They've compared images exposed correctly at standard ISO setting, and then 6 images progressively underexposed by 1 EV, always at the same ISO setting. That's not what this is about: it should have images taken at the same shutter speed and aperture, with the ISO varied by 1 EV from (say) 100 ISO to 6,400 ISO.
But in general, I need to read this stuff still more carefully. Maybe there's some value in the concept that I still haven't discovered. But so far I have:
Some people say that ISO invariance refers to the digitally amplified ISO settings, typically starting at round 6,400 ISO, and that it doesn't make sense to use the higher settings. Others say that it means that it makes no sense to use any setting beyond the base setting. And others again state that, with an ISO-variant camera, it's better to use the highest available ISO. And these people refer to each other's pages with apparently no recognition of the conflicts.
Even the sample images that I've seen appear to be based on incorrect premises that don't reflect either of these approaches.
I still haven't seen anything that relates underexposure to reduced dynamic range.
The concept of “extension” is tied to “digital amplification” in at least one place. But some cameras, including many of mine, have extensions in the other direction (typically simulated 100 ISO on a camera with base ISO 200), which would imply a digital reduction according to this definition. And that won't work: it would clip the highlights. This casts some doubt on the definitions.
In general, the terms “ISO-variant” and “ISO-invariant” seem inappropriate. We're really talking about how the ISO settings work in the camera: only analogue amplification, or digital extension?
There's a lot of material on the web on photography, and the quality varies from appalling (see Jen Miller's atrocity) to very good. But somehow I haven't found a very good page on this subject yet.
While looking at the topic, I arrived at this page. It doesn't add much to the discussion, but the web site may be of interest.
| Wednesday, 3 January 2018 | Dereel | Images for 3 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
DBDRV results
|
Topic: Stones Road house, opinion | Link here |
Today I got the report of the DBDRV assessment of last month. I wasn't expecting much, but this was an unmitigated disaster from start to finish. The DBDRV and the assessor should be ashamed of themselves.
In detail: there were four items.
The issues I raised here were the dangerous mounting of the cooktop and the inability to get a small flame.
The standard requires that if a burner is positioned within 200 mm from the wall, the wall must be non-combustible surface. As the wall behind the cook top is a non-combustible tiled surface, the item is not defective.
I had shown him that the wall behind the cooktop got charred by the heat. At the very least, he should have mentioned this. If that's all that the standard requires, it is severely deficient. Tiles can be destroyed by heat and potentially set fire to whatever is behind them. But he only quotes the permissive part of the standard (Clause 6.10.1.2 of the Australian Standard AS/NZS 5601.1:2013), which also states “Where the distance is less than the specified clearance then protection of the combustible surface is required as per AS/NZS 5601.1:2013”. He doesn't state what protection is required.
Flame control for the gas burners appeared normal at the time of assessment. Therefore this item is not defective.
What does he mean by “appeared normal”? It didn't match the installation instructions. He didn't mention this in his report.
My objection was that the range hood only extracted about 20% of the rated flow. It was clear when he was here that he did not understand the issue, and he stated explicitly that he did not know how to measure it. I told him. But there was none of this in his report. He showed two photos, of the range hood and a straight section of the flue.
Section 18.02 of the Guide to Standards and Tolerances 2007 (Guide) states:"During the documented maintenance period after completion, fittings are defective if they do not operate as intended by the manufacturer. If the maintenance period is not documented it is to be taken as three months. After the end of the maintenance period failure is not a defect unless it is caused by the builder's workmanship."
And then he goes on to the third item. What kind of nonsense is that? Finger trouble, it seems. He mixed up the sections, and the rest of this one is in the middle of the following section:
The assessment revealed that the rangehood was operating as intended by the manufacturer.
I can't determine what the manufacturer intended, and nor can the assessor. But it most certainly doesn't comply with the specifications! He was completely incapable of assessing it! No mention of my measurements, nor indeed what he did to come to this conclusion. And then he writes:
The assessment was carried out in December 2017 and the hand over date was April 2015, therefore the range hood is also outside the maintenance period.
Nonsense! He's confusing certificate of occupancy and handover date (28 April 2015 and 5 May 2015), and he must surely be wrong that my claims have lapsed because of the time. I first reported the issue at an on-site meeting on 28 May 2015, following up with an email message, and we've been fighting ever since.
This was a complete failure. As I mentioned, he didn't even measure the area I was talking about. He quoted:
Clause 14.08 of the Guide to Standards and Tolerances 2007 (Guide) states:
"Floor levels within a room or area are defective if they differ by more than 10 mm in any room or area or more than 4 mm in any 2 m length."The assessment revealed that when an automatic levelling laser level was applied to the Dining & Living floor and the levels were within 4 mm of each other in any 2000 mm direction. Within the whole area there was a difference of up to 5 mm. Under doors a difference was noted of up to 4 mm side to side. In accordance with the Guide, floors must not differ by more than 10 mm in any room or area. Therefore the item is not defective.
So basically, he didn't do this measurement at all.
This is a contractual issue, of course. But I still find it offensive when he writes:
The contract document drawings show windows with double glazing (DG) and sliding units with Low E (Low E) in accordance with the energy report as part of the building permit. The assessment revealed that the glazing to the sliding doors was single glazed Low E glass as required on the contract document drawings and energy report.
That's stupid. It took him himself 10 minutes to understand the abbreviations on the drawings, and there was nothing there to say that the doors should be single glazed. I'm sure there's a requirement somewhere that contracts be specified in English.
In summary:
So I'm done for. I'm now worse off than if I had never contacted DBDRV.
|
Trump: Mine is bigger than yours
|
Topic: politics, opinion | Link here |
A couple of days ago Kim Jong Un claimed to have a launch button for ICBMs on his desk. Just the bait for Donald Trump, who now almost literally replied “Mine is bigger than yours”. I'm left wondering whether that was an attempt at humour.
What an idiot!
|
ISO invariance exposed
|
Topic: photography, opinion | Link here |
Finally I've found a reference to what I've been saying all along about ISO invariance: the concept of underexposing and then pulling the image back up again is Just Plain Wrong. Yes, as I've discovered myself, it works remarkably well when things go wrong. But now I have it stated by somebody else:
Underexposing by 3 stops at base ISO (100?) takes advantage of only 1/8th of the sensor's physical capacity and in effect turns a 14-bit sensor into an 11-bit sensor. It uses the same exposure and generates the same S/N ratio as if you had exposed normally at ISO 800 in place of 100 and used all 14 bits.
But that's a forum post. Why don't other people write sensible articles? It's still not even clear that all people mean the same thing.
|
New kitchen roll holder
|
Topic: general, opinion | Link here |
Our old kitchen roll holder has broken. Not surprising: it's out of plastic, and we've had it for over 20 years. And it wasn't difficult to find a replacement on eBay. It arrived today, and I installed it immediately.
What a disaster! The aluminium foil cutter just doesn't work at all, and the thing is designed to deliberately restrict the width of the rolls, so standard Australian rolls (33 cm) don't fit, although the device is 37 cm wide:
Why do people make things like that?
But why did the aluminium foil cut so badly? Looking at the cutters shows that each device has different cutters for aluminium foil and plastic foil. Here the old unit on the left, first the plastic foil, then the aluminium foil. On the right I have my best guess at the way it's intended to work for the new holder:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's not really surprising that neither of the cutters work for aluminium foil.
|
Kangaroo trap
|
Topic: animals | Link here |
Over 3 years ago we found a kangaroo trapped in fencing wire:
|
|||||||||||||||||||||||
We called the wildlife rescue, which came along and cut it out of the wire, but the injuries were too severe, and they had to kill it.
At least that was a humane death. Today we saw what happens when nobody sees in time:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
The future of retailing
|
Topic: general, technology, opinion | Link here |
Four years ago I participated in a Coursera course about the history of the Internet. One of the assignments was an essay on what the Internet might be like in 20 years' time. I've been following it ever since—it's amazing to think that 20% of the time has already elapsed.
Today I read an article in the Washington Post about the shutdown of the last department store in Hermitage, PA. Hermitage has a population of about 15,000, and the nearest larger town is at least 50 km away. This will probably kill off the Shenango Valley Mall as well.
In passing, it's notable that the mall web site is at least 9 months out of date: it lists all three department stores that were there until March 2017, when Sears and Macy's closed down. Wouldn't you be happy if you went to a mall looking for one of three department stores, and found none at all?
Is this important? From a larger viewpoint, no. More and more people are buying “online”, and that will continue. Further outlets will shut down, and some big companies will go broke. What we really need now is cheaper postage (the USA is particularly bad in this respect) and better presentation of the goods online.
| Thursday, 4 January 2018 | Dereel | Images for 4 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Penang laksa again
|
Topic: food and drink, history | Link here |
Penang laksa is a happy recollection that I have of my youth, though now that I have my entire paper diary online, I discover that I have only one mention of what could have been Penang laksa, but I didn't specify; the rest were Sarawak laksa. The first explicit mention of Penang laksa in my diary is less than a year old, and was only partially successful.
Today I tried again with a different paste, which had instructions including the use of har ko. What's that? It says “prawn paste”, but the normal Malay word for that is belacan. Maybe that's the paste that is always in Penang laksa, and that I had thought was Tamarind. Went looking in the pantry, and how about that, I found something:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Hea ko? Har ko? Who cares? It must be the right stuff. It even comes from Penang. Further investigation takes me back to the belacan page on Wikipedia, which tells me that it's the Min Nan term, and that “hae ko” is the correct spelling, though the Cantonese haa¹ gou¹ (蝦膏) might be transliterated as har go. My recollection is that it much less solid than belacan.
Clearly my specimen is not the freshest. How old? No expiry date on it, of course, but a price label:
|
|||||||||||||||||||||||
That's in Dutch! It's not overly surprising: I used to buy Malaysian and Indonesian food in Venlo when I was living in Voerde, from 1979 to 1982, though on one occasion I bought some in Brussel. And when? Confirmed with Yvonne that I had never gone shopping with her in Holland. I was in Bruxelles in June 1987, and did some shopping before I returned, though I didn't think I would have bought this kind of food.
So which was it? This page suggests that it was Venlo. At an educated guess, I would have bought this stuff in 1979 or (more likely) 1980. I know we have some old stuff in the kitchen, but I think this takes the cake. The last time I ate laksa in Malaysia was over 48 years ago, and only about 10 years before I bought this paste, which is thus older than many people I know. I didn't use it.
The laksa itself was pretty much the same as last time. Once again I made it in two steps. First, the sauce.
| quantity | ingredient | step | ||
| 250 g | Richmond asam laksa paste | 1 | ||
| fish broth and water to 1500 ml | 1 |
The fish broth came from boiling the fish below.
| quantity | ingredient | step | ||
| 300 ml | sauce | 1 | ||
| 80 g | noodles | 1 | ||
| 90 g | hoki fillet, boiled and shredded | 2 | ||
| 40 g | cooked prawns | 2 | ||
| 1 | fish ball | 2 | ||
| 50 g | pineapple chunks | 3 | ||
| raw onions | 3 | |||
| tamarind paste | 3 | |||
| mint | 3 | |||
| Thai basil | 3 | |||
The tamarind paste was necessary; it's not sour enough otherwise. Here's the result:
|
|||||||||||||||||||||||
And it tasted quite good, much better than last time. Different brand of paste? That's what I thought, but it seems that it's the same:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In previous entries I referred to it as “Richmond” or “Oriental Merchant”, both of which are written on the back.
|
ISO Invariance: the proof
|
Topic: photography, opinion | Link here |
One of the big claims of the proponents of “ISO invariance” is that you can take photos at the same exposure and lower ISO sensitivity settings than “required”. But all the sample images I have seen are taken from a different point of view: same ISO setting, different exposure. OK, if I can't find anything online, I can always do it myself. According to this page the Olympus OM-D E-M1 Mark II should be “somewhat” ISO invariant. Let's try with a scene that has noticeable shadows. According to my previous claims, shadow detail should be the first to go. So I took the same image at 1/40 s and f/5.6, starting at 6400 ISO and dropping in steps of 1 EV to 200 ISO and “low”, whatever that might mean. Here are the results for 6400 ISO, 3200 ISO, 800 ISO, 200 ISO and “low”. Run the cursor over an image to compare it with the first (6400 ISO):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I took others at 1600 and 400 ISO, with expected results. See the raw photo page for all images. I used
DxO PhotoLab with the “Artistic HDR” profile preset. I didn't try any noise recovery. As
expected, the quality drops significantly as the ISO setting drops. Only the “low” setting
seems out of keeping.
Here's a comparison of the details at 6400 ISO, 800 ISO and 200 ISO. Again, run the cursor over an image to compare it with the first:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
As I've noted in the past, it's really quite impressive how well the camera deals with extreme underexposure. But as a deliberate technique? Useless.
|
Piccola regains territory
|
Topic: animals | Link here |
Yvonne drew my attention to this this morning:
|
|||||||||||||||||||||||
What's so strange about that? That was Rani's armchair, and Piccola wasn't allowed on it. It has taken 8 months for her to regain this territory.
| Friday, 5 January 2018 | Dereel | Images for 5 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
How long will my shutter last?
|
Topic: photography, opinion | Link here |
One of the most likely causes of failure on a modern system camera is the shutter. Some manufacturers rate them, others don't. I've heard various numbers about my cameras, generally round the 100,000 mark, but no firm statement from the manufacturer. But Yvonne's Olympus E-PM2 had a shutter failure after only 12,000 releases. I've taken more than that with almost all my cameras, but I don't know if one of them might fail too.
So it was interesting to find the shutter life database, unfortunately too small and too out of date to be of any use. Can it be revived?
|
Smart TV usage
|
Topic: multimedia, technology, opinion | Link here |
Finally got hold of Danny at The Good Guys in Ballarat, and told him of my pain with the TCL 60"(152cm) UHD LED LCD Smart TV . No, he didn't know that the thing didn't have a microphone. He also didn't know that Android TV relied on it. That's good for me: it means that nobody warned me that it was incomplete. Danny was happy for me to return it for a refund or exchange, so that's fine.
But why didn't he know? My guess is that nobody uses Android TV. Certainly I find it pretty useless, and the fact that the manufacturer offers a castrated version seems to suggest that it's basically a box to tick on the spec sheet, not something that somebody uses.
Now I have another problem: the ALDI 65" TV works fine, but next weekend they have another on offer with a 75" diagonal. Should I buy it? There's no problem returning the 65" model, but the 75" model costs more than twice as much ($1,699 against $799). Is it worth it for just another third as much screen area again?
|
Mobile phone fun
|
Topic: technology, general, opinion | Link here |
Petra Gietz along this morning with her phone. She tells me that despite the new radiation tower, her coverage was no better. Charged the phone for her and confirmed that it worked fine. It seems that she had thought that the Dereel tower would cover Enfield as well. That raises the question: where is the tower that services Enfield (13 km away)? They used to have better coverage than here.
While we were at it, got her to call my phone to confirm that we had coverage, and also to capture her mobile phone number. I didn't hear the phone ring. Then somebody answered. Wrong number. Looked at the number I dialed. It was correct. Huh? After she was gone, I checked: I had remembered my phone number incorrectly since I started using it. That doesn't exactly confirm my claim that I could remember it easily. A good thing I didn't have new business cards printed.
| Saturday, 6 January 2018 | Dereel | Images for 6 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Hot weather
|
Topic: general | Link here |
So far the summer has been relatively mild, especially compared to the last month of spring, but today was an exception. The temperature rose to 40.8° by 16:34. And then we had one of these sudden drops in temperature that happen so often in this part of the world:
|
|
|||||||||
Looking at the database entries, the temperature dropped by up to 2.2° per minute! In 12 minutes it dropped by over 8°:
Is that a record? It could be. I've had many such occasions before, but I can't be sure I whether any has been as extreme as today. On Black Saturday we had a similar situation, but the highest rate of change I calculated then was 1.3°/minute. I should keep better records.
|
Erase that link!
|
Topic: technology, opinion | Link here |
A while back I became interested in house prices round Newton Abbot in Devon, where I spent a lot of time in the early 1960s. And somehow I'm now on a real estate mailing list that gives minimal information about properties.
Today another one arrived. Things have changed a lot since I was in Devon. In May 1969 I considered a house—admittedly not in the best condition—for sale for £300. Now it seems that something (they didn't divulge what) in the Newton Abbot area sold for round £200,000, roughly what we paid for our current house. OK, enter the address into Google Maps and take a look.
Problem: how do you get rid of the pin that locates the address? It gets in the way, but I can't find a way to get rid of it. Not so today:
|
|
|||||||||
OK, where do I click? Nothing seemed to help. And then I discovered:
|
|||||||||||||||||||||||
|
|
|||||||||
It's “Erase That ink”, not “Erase That link”, and it's a business name. Aren't expectations funny things?
| Sunday, 7 January 2018 | Dereel | Images for 7 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
DBDRV: no thanks
|
Topic: Stones Road house, opinion | Link here |
Spent much of the day writing a letter to the DBDRV expressing my disgust at the assessment that I have received, now partially corrected. I asked for information on how to lodge a formal complaint, and suggested that they should investigate the matter from their end. I wonder how they'll respond.
|
Dosai? Dhosai? Thosai?
|
Topic: food and drink | Link here |
The prawn paste that I found on Thursday must rank as the oldest foodstuff we had in the pantry, but there's plenty there that has well exceeded its use-by date. I also have some idli mix (date unknown) and some dosai mix (from Sri Lanka, and labeled Thosai) which expired not quite 10 years ago in April 2008. Presumably it's at least 12 years old.
Time to continue tidying up. Which do I make? I decided on the dosai, something that I have never made before. Watching the people make them in Woodlands in Singapore rather unnerved me. So this time I decided to bite the bullet and make them.
The instructions were rather rudimentary. Put yeast in water until it foams a bit. Sorry, not this 12 year old, long dead yeast. I had to use some of my bread yeast, which presumably doesn't make any difference. Mix until a thin batter is made. How much? I still don't know. I made it thinner and thinner, but it still proved to be too thick. And the recipes I found on YouTube were completely different from mine, like this one for masala dosa:
But I didn't have the ingredients for that, so I made it without filling. Far too thick. But it didn't taste bad, and it's something to experiment with.
| Monday, 8 January 2018 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Spam writer's network?
|
Topic: general, technology | Link here |
Found while cleaning out my spam folder today:
I've seen that before, but the address was different. Same technique, including referring to a diary entry that is 50 years old, same wording. Even the name starts with J (I've seen at least three variants). And this was filtered as spam by Gmail, not me. There seems to be a whole network of them out there, and they must have annoyed enough other people to be filtered as spam.
|
Buying a new kitchen roll holder
|
Topic: general, technology, opinion | Link here |
The seller of the kitchen roll holder that I got last week is stringing me out, but the real issue is to get a new one. They're either expensive or ridiculously expensive in postage—the most expensive postage I have seen was € 99. And then I found one from a company called ibuys. Also not cheap, but in total round $55. So I tried to buy it, fighting as usual the stupid web forms that all these companies have.
This one had an additional trick: it refused my email address. Why? I don't know. I couldn't complete the purchase. So I tried their “contact” page. It, too, wanted an email address, and rejected mine. While pondering that, for the fun of it, entered president@whitehouse.gov. Yes, it accepted that willingly—so willingly that it sent the message with no further input from me.
Sometimes you'd think that people don't want to sell things.
|
Fighting the weeds
|
Topic: gardening | Link here |
Nice, almost windless day today, just what I needed to spray the weeds. That worked, but some of them were so high that it seemed more appropriate to pull them out. This must be some of the first real work I've done in the garden in a long time, and it didn't last long.
| Tuesday, 9 January 2018 | Dereel → Ballarat → Geelong → Dereel | Images for 9 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Returning the TV
|
Topic: general, multimedia | Link here |
To the Good Guys in Ballarat this morning to return the TCL TV. Despite my concerns, things went pretty smoothly. Danny confirmed to me that many customers asked about Android TV, but he suspected that few actually used it. And no, they didn't have an Internet connection to demonstrate the things, and they don't intend to: it would take far too much time.
In passing, it's interesting to note this statistic, from Statista:
Yes, it's about phones, but it applies in many cases to smart TVs. What interests me is the discrepancy between awareness and use of voice-operated functions. It seems that I'm not the only person to be disenchanted by them.
|
Food shopping and periodontics
|
Topic: health, general, food and drink | Link here |
Off earlier than expected to Geelong, along a very slow Midland Highway: January is the main holiday season, so they have chosen it to do roadworks on the main roads. It seemed that, for no good reason, they had restricted the speed limit to 40 km/h all the way from Lethbridge to Bannockburn, a distance of about 10 km. In the process, I had time to note that I had crossed the Midland Highway more often than I have driven along it, in about 8 different places from Buninyong to Bannockburn.
Arrived in Geelong 20 minutes before I had planned, and spent it looking for a parking place in front of the Vietnamese grocery in the middle of town. Normally I have been able to park directly in front of them, something that surprised me, but this time I had to park a block away.
Down through a shopping street, showing the signs of the times: many unoccupied shops, including an entire arcade. Those shops that were occupied seemed to be hairdressers or restaurants, though there was one shop for bridal equipment. This is very much in keeping with the expectations in the Future of the Internet paper I wrote 4 years ago.
Didn't find much at the Vietnamese place, and I was finished early enough to drive round town a bit and take some photos:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Nothing much of interest at the periodontist's. Gum gaps seem to be getting yet smaller, to the point that they're pretty much normal for a much younger person.
Then to Belmont to buy Indian and Malaysian food, notably the latter. Spoke to Jason Yap at the Gourmet Asian Grocery, predominantly Malaysian. Jason comes from KL, and he paints a particularly dismal picture of what life is like there now—that's why he's in Australia. I wonder what the future holds.
| Wednesday, 10 January 2018 | Dereel | Images for 10 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
eBay: 1 point up, 1 point down
|
Topic: general, technology, opinion | Link here |
Returning my terminally broken roll holder proved more difficult than expected: the seller strung me along and kept trying to compromise (“we'll refund $12 and you can keep it”). So this morning I asked eBay to “step in“. And that they did. I don't know how long it took them, since they don't give me copies of this kind of request, but it must have been less than 10 minutes. One point to eBay.
And of course the seller got negative feedback. Finally he reacted:
OK, but sorry, too late, and not very convincing.
But that wasn't the end of it. In the evening I got another message:
|
|
|||||||||
Huh? What the hell is that? And what was the “original outcome”? This makes no sense at all. OK, look at the case details:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
This is a particularly insidious case of reverse chronological reporting: it's impossible to recognize, since they don't give times. But it's telling me that I complained to eBay, the seller gave me a complete refund, and then I appealed. The appeal was closed.
What's wrong with this? Everything! I didn't appeal, and the conflicting information on the two pages left me wondering whether I got a refund or not. But clearly this is yet another case of complete and utter incompetence at eBay. One negative point to eBay.
|
Landscaping, finally
|
Topic: gardening, general | Link here |
Woken at about 7:50 this morning by Warrick Pitcher. He was due to do some earthworks in the garden at the beginning of last month, but something got in the way. Now he could do it, giving us barely time to get out of bed.
In the course of the day he transformed the garden: two new beds and a gravel pathway, also hiding the fact that one part was a drain:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
By evening he had used up all the gravel on less than half the area, but he had soil left over:
|
|||||||||||||||||||||||
Nothing for it: order more gravel, which is horribly expensive. We had thought that we would get by with 4 m³, but we had to order another 6 m³ for $620. By comparison, Warrick's work time was cheap.
In passing, it's interesting that my yellow flowers are doing quite well, and even seem to be suppressing the grass:
|
|||||||||||||||||||||||
Warrick tells me that they're periwinkles, but the links I've found there seem to refer to Vinca minor, which they most certainly are not.
|
Masala dosa, try 1
|
Topic: food and drink | Link here |
So it looks like the canonical spelling is dosa, not dosai nor thosai, and the best way to eat them seems to be masala dosa. It doesn't look difficult, but where are the recipes? I have a number of Indian cookbooks, but I couldn't find a recipe in any of them.
OK, to the web. Plenty of recipes. The good news is that the ingredients are all pretty similar: potato, onion, curry leaves, mustard seed, turmeric and a surprisingly small quantity of channa dal. In the end I started with this recipe, which surprised me by printing a full 47 pages. Here's the recipe as I tried it:
| quantity | ingredient | step | ||
| 500 g | potatoes | 1 | ||
| 200 g | onions | 2 | ||
| 5 g | green chili | 2 | ||
| oil for frying | 3 | |||
| 1 g | yellow mustard seed | 3 | ||
| 2 g | black mustard seed | 3 | ||
| 2 stems | curry leaves | 4 | ||
| 20 g | channa dal, cooked and crushed | 4 | ||
| 2.5 g | turmeric | 5 | ||
| 1.5 g | asafoetida powder | 5 | ||
| water | 5 | |||
| 10 g | salt | 6 | ||
And how did it work? The spice quantities looked a little on the low side, and that's the way they tasted. I'll try more next time. And I still need to understand how to mix dosa batter correctly. It should clearly be much thinner than I had expected.
| Thursday, 11 January 2018 | Dereel | Images for 11 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Completing the landscaping
|
Topic: gardening | Link here |
Warrick Pitcher along this morning at 7:50 to receive the gravel and spread it:
|
|||||||||||||||||||||||
He had a couple of other jobs to do, but he was gone by 11:00.
|
Sprinkler work
|
Topic: gardening | Link here |
One of the surprises we had this morning was that the ground that Warrick had gone over was quite wet. It hadn't rained yet this year (the first rain came in the evening). Clearly an issue with the sprinklers. I've been leaving that to Mick, but I have had more and more reason to think that they needed attention. Fired up the sprinklers and confirmed yes:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Spent several iterations fixing a large number of issues, partially leaks, partially blocked sprinkler heads. And I'm still not done.
|
Leaking sprinkler photos
|
Topic: photography | Link here |
My photos of the sprinkler leaks wasn't as dramatic as I would have liked. In particular, what went wrong here? Yes. there was a jet of water going across the middle of the photo:
|
|||||||||||||||||||||||
Too fast a shutter speed, I think. Looking at individual parts of the photo, the jet has turned into individual water droplets:
|
|
|||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
These were taken at automatically selected 1/250 s. Next time (hopefully there won't be a next time) I'll try a much slower shutter speed.
|
Internet future: on track?
|
Topic: technology, opinion | Link here |
I've seen a number of things lately related to my essay the future of the Internet. One thing I didn't get quite right (yet) is the demise of the fixed line telephone network. I claimed that it would be replaced by VoIP, but this statistic from Statista tells a different story:
That's definitely a perspective problem on my part. I don't see (have) a need for mobile phones, and my $15 recharge keeps me going for a year. On the other hand, calls to mobile phones make up the bulk of my call costs, so I try to avoid them. Yes, I've seen numerous people give up their fixed line phones, forcing me to spend Real Money to call them on their mobile phone. The real issue is the line rental for “fixed line” phones, which in this case even includes VoIP, even though it's much cheaper.
Mobile phones don't preclude using VoIP, of course: I have a VoIP app on my mobile phone, and if I could get it to work acceptably it might well be much cheaper than mobile call rates. But the real issue I have is that people are still moving round much more than the Internet requires.
|
Mystery flower exposed
|
Topic: gardening | Link here |
I've seen this particular wildflower in many places, starting in Kleins Road years ago. But this one is unusual in that it isn't in the middle of a lot of other vegetation. Also the flowers are some of the largest I've seen in this species, nearly 1 cm across:
|
|||||||||||||||||||||||
|
|
|||||||||
Time for some real close-ups.
| Friday, 12 January 2018 | Dereel | Images for 12 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
I am the very model of a modern stable genius
|
Topic: politics, opinion | Link here |
The more I watch Donald Trump's behaviour, the more I wonder if he is not secretly trying to take the Mickey out of the whole western world. His latest “stable genius” claims reminded many people of the Major-General's Song in the Pirates of Penzance. Here a number of them.
|
Everybody hates Twitter
|
Topic: technology, opinion | Link here |
It's no secret that I don't like social media. But I thought I was relatively alone until I read this statistic from Statista:
More people want to kill Twitter (that's the bluebird without text in first place) than keep it alive! I wonder if that has something to do with Donald Trump.
|
Clematis recovery
|
Topic: gardening | Link here |
It's been six weeks since I planted my expensive Clematis, and since then I've watched them try not to die. They're still not looking good, but things are improving. The “General Sikorski” has a couple of shoots, even if one lost its tip during a recent hot spell:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
And though the Edo Murasaki still looks very patchy, it's coming up with a couple of buds:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
| Saturday, 13 January 2018 | Dereel → Ballarat → Dereel → Ballarat → Dereel | Images for 13 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Yet another bigger TV
|
Topic: multimedia, general | Link here |
Off early this morning to ALDI to buy yet another TV. Yes, we just bought one, but has only a 65" (1.65 m) diagonal, and the new one measures 75" (1.90 m). This kind of special sells quickly, so we wanted to be there when they opened. We managed that with ease. I had left to be there in 32 minutes, but on the way, Yvonne told me that it took her 22 minutes to get there. It didn't even take that long: we were there in 19 minutes.
Finally we got in and found the TVs in the storage area: 6 of them in total. We got the first, and two others standing in line got another two.
Gradually we're coming to the limit of the size of a TV that we can transport in the car:
|
||||||||||||||||||||||||
Went past the Sebastopol used car lot on the way home. Very little available, and at prices that make it seem worthwhile to keep my current car. Then to the Delacombe Town Centre, where we were at the opening last September. It was still early, but there did seem to be some activity. And I had heard that they're planning to extend in four directions, so maybe something will come of it after all. Spent forever trying to get out: there's a completely superfluous traffic light that appears to be incorrectly programmed, and it didn't change for me.
Got the TV into the house and unpacked with little difficulty—we're getting used to this—and set it up quickly. They've improved the packaging: now it explains how to open the package:
|
|||||||||||||||||||||||
And then things just fell out the bottom, leaving just the packaging foam at the top of the box:
|
||||||||||||||||||||||||
Well, most things. Where was the remote control? Where were the feet? In the foam, of course, which wasn't easy to get out.
Not surprisingly, the TV is very similar to the other one except for the screen. The remote control is identical—we didn't even unpack it—the connections are all the same, and so is the software. Even the too-short power cable seems to be the same. The biggest issue was the size, the weight, and the location of the feet, which bent the plank that I had specially purchased for the last TV, to the point that I was concerned it could snap. We had to put some old gardening books underneath the middle:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
I had it running within 2 hours of setting off to town (run the cursor over an image to compare it with its neighbour):
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Is it worth it? It's a third more screen size than the previous one, and two-thirds more than the old TV that failed last month, and that does make a difference. But we didn't have time to watch TV today.
The hallway is getting narrow. Before we buy any more TVs, we'll have to get rid of some:
|
||||||||||||||||||||||||
|
Eliminating cars
|
Topic: photography, opinion | Link here |
What could I do while waiting for ALDI to open? A panorama, of course. There's quite a bit of stuff round the roundabout where ALDI is situated. The other side is the beginning of the Glenelg Highway, and there's Woolworths on the other side. And wouldn't it be nice to get the road without any cars on it?
The effort (14 shots) was only marginally successful. I got this far and gave up:
|
||||||||||||||||||||||||
Yes, I was able to get rid of the cars, and with more work I could have done better. But it wouldn't have been good: this is really a situation for a tripod, and getting the components to align was just too difficult. I should keep a tripod in the car.
|
No water!
|
Topic: general | Link here |
Just before midday, Yvonne came to me with the news that we had no water. Well, nothing came out of the taps. A bit of investigation showed that the water pump was tripping the RCD.
Why do these things always happen at weekends? Googled for the phone number of UPI, and was given the convenient information Closed now. Dammit! 12:02, and they shut at 12:00. Tried the Ballarat Pump Shop, which was engaged, and then tried calling UPI anyway. Yes, Luke was still there, and with a bit of begging he got Nick to agree to stay there until 12:30, when he could lend me a pump. Photos, please, to help choose a pump with similar fittings:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Off in a hurry to Ballarat again, and somehow managed to arrive there at 12:32. We got a loaner pump, which proved to be pretty much identical to the one we have, and we were out again by 12:47, only 45 minutes from when I tried to contact them.
Back home, installation was hampered by the lack of a second pipe wrench, but it worked. If only we had known that before we left for town the first time.
|
System down, won't come up, customer screaming for blood
|
Topic: technology, opinion | Link here |
When I came inside after installing the water pump, Yvonne had a message from Chris Bahlo: “The server is down, whatever that means”. Really? Yup, www.lemis.com was off the air. Off to investigate. The system is run by RootBSD in North Carolina, and there's VNC access to the system console. But that didn't work either! Something serious? I was able to access their ticketing system, which reported that all was well, and so I entered a ticket, at 2:55:12 UTC.
But then I looked at the existing tickets. One issued just yesterday:
Greg Lehey groggyhimself@lemis.com 2018-01-12 05:14:32You are receiving this message because you currently have services on our old VM infrastructure.
We are decommissioning our old infrastructure at the end of the month, and will need to migrate your services to our new platform before then. We have scheduled the following migration window for this event:
Scheduled Migration: Jan. 12th 6-9PM EST
During this window your VM will be temporarily shut down and moved to our new VM platform. Total downtime is dependent on amount of data on your VM's disk, but in most cases will be less than 30 minutes per 50GB of data. Your VM's IP address, usernames/passwords, and other data will remain unchanged.
If you have any questions regarding this migration event, or would like to reschedule for another window please let us know. Once the migration is complete we will update this ticket to let you know.
Staff RootBSD 2018-01-13 01:56:34
The migration is complete and your server is back online.
Huh? What's wrong with this picture (part 1)?
When I ran Tandem's European OS and hardware support in the 1980s, we had three levels of severity for problem reports, with 3 being the most serious. But we always had the feeling that there should be a level 4, “System down, won't come up, customer screaming for blood”. Somehow that seems appropriate here.
I got a reply to my ticket: closed without comment. The migration ticket was updated with information about how to use a web-based VNC service. I tried that and got a vague error message indicating that I couldn't connect to the system.
I later got an update to the ticket, just as I noted that the system had come back online:
Your server is online and pingable now. I was able to boot into single user mode and fix your rc.conf settings. The alias was incorrectly configured, and was preventing the primary interface from assigning a routable IP address.
The original VNC access details are no longer valid, you can now use our new portal to securely access the VNC console. This is a long-waited security improvement, as the original VNC access was unencrypted.The error 1006 when trying to access the VNC console indicates a web socket issue with your browser. If you receive this error please attempt to open the console again, or try a different browser. If the issue recurs please let us know so we can investigate further.
The good news is that the server was up again. But there are still things wrong with this:
I've been very happy with RootBSD up to now. This incident has made up for that in a big way. Should one big problem spoil the overall impression? Certainly I need to think about my future with RootBSD.
|
Shitholes of the world
|
Topic: politics | Link here |
A new word has gone across the airways in the last day or so: “shithole”, translated into German as the relatively mild „Drecksloch“. Yes, it's horrible that the President of the United States of America should use such words. But given his track record, I was surprised how badly people took it.
But then there are the satirists, of course:
And there's Facebook. For reasons I don't understand, I didn't save this image, and of course the original has vapourized.
|
Cooking Kassler
|
Topic: food and drink | Link here |
We had Kassler for dinner today. Another data point for my cooking time table: 500 g, starting at 8°, took 55 minutes to reach 74° in a non-fan-forced oven.
|
Shopping in the Internet era
|
Topic: general, technology, opinion | Link here |
At dinner this evening we discussed the Delacombe Town Centre with Chris Bahlo. Somehow the idea of building new shopping centres doesn't match what is going on elsewhere, as I noted last week. And even in Australia we have the recent failure of Masters Home Improvement to think about. Chris observed that there were still some things that people needed to do on site, like hairdressing—exactly what I had noted earlier this week. And we discussed the opportunities for people delivering food ordered “locally” (with the example of us buying in Ballarat and having it delivered to Dereel).
Yvonne considered that she would still go shopping, because she would want to choose her melons, grapefruit or papayas. That's a valid concern, but it it a show-stopper? Maybe there's an opportunity, in the short term at any rate, for people who can perform that function.
| Sunday, 14 January 2018 | Dereel | Images for 14 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Server outage, day two
|
Topic: technology | Link here |
Yesterday's server outage was finished by 15:00:
=== grog@www (/dev/pts/1) ~ 3 -> date; uptime
Or was it? Today we discovered that mail was not working. Why not? It was in fact working, just rejecting everything:
UNKNOWN? What's that? In any case, none of this was from systems I knew. OK, send a test message and see what happens. Nothing, not even a connect. This was from freefall.freebsd.org, and it's possible that the mail system might take a few minutes to send it. OK, use telnet and send the message directly. That worked.
OK, a second system? ffm.lemis.com? I don't have mail running there. ozlabs.org? Yes. And mailq showed that it couldn't connect to mx1.lemis.com, which is the same server.
After a bit more messing around, I noted the address of mx1: 192.109.197.81. That's in my old /24 net block. And how about that, the configuration fix that RootBSD made hadn't worked: my alias was gone. Fix that, and all was well.
Only indirectly RootBSD's fault: I had made a mess of the original config file, but the only way to test it was to reboot, and that's a no-no. But after fixing it, I would have noticed whether the alias address was there or not—if I had had the access I needed.
And that UNKNOWN? Almost by definition, it's spam, since it was going to the primary address and not the mx1 address. But it seems that it happens a lot, and I just didn't pay much attention before.
Later, while tidying up my Gmail folders, found the message about the upgrade. It had been identified as spam. But that alone doesn't explain it. I had looked for it yesterday and found messages going back to before Christmas (now in “Trash”), but not that one. Gmail is too polite to give times, but it seems that it must have arrived after I last looked.
|
Cloud storage
|
Topic: technology | Link here |
Spent some time looking at cloud storage options. This article looked promising, and so did some of the prices. I need about 250 GB at the moment, and it seems that I can get up to 1 TB for $10 or so a month, maybe cheaper. Or, as the article puts it for the first two, with adventurous exchange rates:
Price: 2GB free. 1TB for $10 a month (£6.58, around AU$11) with Dropbox PlusPrice: 15GB free. 100GB for $1.99 a month (£1.59, around AU$2.50). 1TB for $9.99 a month (£8, around AU$13)
But how? The article recommends the old favourites first: Dropbox and Google Drive. Off to read their description. Easy drag-and-drop interface! What's easy about that? How do I script it? Why can't I use standard tools for copying? It would be nice to use rsync, for example.
Part of the problem is that the documentation doesn't really describe what I want to know, only what somebody behind a mouse needs to know. And Google Drive's documentation suggests that it won't take images larger than 16 megapixels, and that they will be downsized. Is that correct? If it does, it kills it outright.
Why is this all so difficult?
|
TV problems
|
Topic: multimedia | Link here |
The new TV looks good, and the increase in size is a real boon. But since installing it, I've seen that some videos are jerky. It's particularly noticeable with Al Jazeera News, probably because of the text band that continually wraps across the bottom. It seems to jerk in roughly 1 second intervals.
Does this really have anything to do with the new TV? It wouldn't be the first time that the YouTube scripts were broken. OK, get out my field monitor and try using my newly arrived HDMI switch to move between them.
Multiple failure. Firstly, the switch has three inputs and one output. Secondly, it seems to have failed. Thirdly, just getting at this stupid HDMI connector behind the TV is a pain. HDMI is fiddly at the best of times, but with a socket pointed down towards the ground behind the TV it's not the best of times.
Finally I had done it, and how about that, on the field monitor it displayed correctly. So it does seem to be related to the TV. Maybe an artefact of scaling down from 3840×2160 to 1920×1080? Tried setting up a second server running at native resolution, which gave me other issues: for some reason I couldn't start an xterm on it.
OK, any information in the log file. You want information? I have information!
=== root@teevee (/dev/pts/5) /var/log 7 -> l Xorg.0.log
2.5 GB! What's in there? Lots of repetitive information about the EDID. OK, I had this server started with -logverbose 6, but it shouldn't have been that much. Is writing to the log file slowing things down?
Remove verbosity, restart server. Problem gone? No, just reduced:
The messages repeated once or twice a second. And yes, this TV also calls itself XXX TV. I've seen that before, and on that occasion I found a tweak that may or may not have worked. Unfortunately, the tweak was still in place, and it most definitely was not working.
So what's the problem? It looks as if the TV is sending some information, either status or a real reconnect, and that's causing the driver to do something that takes up time. It went away with the old TV, so it should be possible to fix it, possibly by telling the driver to ignore what the TV says altogether.
Why is this such a pain?
|
The positive side of HDMI
|
Topic: animals, technology | Link here |
I may have cursed the HDMI cables that I was playing around with this afternoon, but one person found them convenient:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Silly galahs
|
Topic: animals, photography | Link here |
In South Australia we had lots of galahs round the property, but since moving to Dereel we haven't seen many. But lately we've had a number
Came out the front door today and found a number sitting in the dead tree in the “nature strip”, beautifully illuminated by the setting sun. Back inside to get a camera and lens. They were still there when I came out again, but the sun wasn't:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
With any luck they'll hang around for a while and give me other chances.
|
Riding is fun
|
Topic: animals, general, health | Link here |
While I was watching the news today, Chris Bahlo came in and said “you're needed here”. Out to find Yvonne in anything but good condition. They had been out riding, Yvonne on La Tigre. La Tigre had tripped, falling on her face and scratching her nose a bit. Yvonne had gone flying over La Tigre's head, rolling over a couple of times, landing on her back and winding herself. She also managed what promises to be a superb black eye.
Took Chris home, who told me on the way what had happened, and back to find Yvonne having changed her mind about not needing medical attention. Off to the Ballarat Base hospital, for the first time getting people to come out and collect her out of the car. Parked, and then in to take a look:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Doctor in shortly after that, and confirmed that there was nothing extremely serious wrong with her, but that they would need to perform some scans. I came home and waited. Finally, round 21:45, I got the results: ribs 4 and 9 on the right broken, vertebra T6 crushed, vertebrae T7 and T10 both with broken transverse processes. Fortunately there was no impact on the spinal cord. They were still waiting for a surgeon to come and take a look, and there was a possibility that she would be operated on tomorrow, either in Ballarat or in Melbourne.
Then I spoke to Yvonne, who said much the same thing, but also that they were planning an MRI scan tomorrow, and that there was a possibility of a shoulder blade fracture.
Round 0:25 she called again. The surgeon had decided that there was no immediate need for surgery, and further decisions would be made after tomorrow's MRI scan.
What a weekend!
| Monday, 15 January 2018 | Dereel → Ballarat → Dereel | Images for 15 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Garden flowers in mid-summer
|
Topic: gardening | Link here |
Middle of the month, time for the monthly flower photos. There's certainly progress to be seen since last month. The Acer negundo (Box Elder) is growing happily. Here last month and this month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On the other hand, the birches seem to think that it's autumn.
|
|||||||||||||||||||||||
Not enough water? Too warm? Too much sun?
As expected, the Buddleja is flowering. Again the first image is from last month:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The Clematis are also gradually recovering:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
On the other hand, the Hebes are doing differently. The white-flowering ones are happy, but the purple-flowering ones less so. Otherwise they look happy enough, so maybe this is just a seasonal thing:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That's definitely the case for the Leucospermum cordifolium, which has reached the end of its flowering period:
|
|||||||||||||||||||||||
The roses weren't at all happy last month, but they're coming back in force:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I think that was related to the problems we had with the irrigation.
The salvias in the north bed looked pretty much dead in early October, but they've recovered well:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I'm somewhat ambivalent about this one, a thistle. On the one hand, it's pretty, but it's also a weed:
|
|||||||||||||||||||||||
The tree ferns are doing well enough, especially the one that Mick nearly killed off:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Indoors, despite my expectations, the Phalaenopsis is still flowering, though it's on its last legs:
|
|||||||||||||||||||||||
As always, more photos on the photo page.
|
Yvonne's progress
|
Topic: health | Link here |
Call from Lara, a registrar (what do they do? Wikipedia seems vague, but this page, from the AMA, suggest that's it a trainee specialist) this morning. The MRI scan had shown that Yvonne also has a fractured scapula (shoulder blade), as we had feared, and that the crushing of vertebra T6 had resulted in fragments intruding into the spinal canal. In addition, the anterior ligament had ruptured.
It seems that operation is a last-ditch option for this sort of thing. The spinal cord has not been compromised, so they planned to fit a brace later this week, which she will need to wear pretty much all the time for the next three months. More to be determined by Mr. Tom Morris, the neurosurgeon.
Into town later in the day, first to take the water pump in for repair, then to the hospital. The new main entrance is open, and the entrance hall takes up enough space for a whole small hospital. Got directions to the ward (“Ward 2 North”). Getting out of the lift showed that I was in the right place:
|
|||||||||||||||||||||||
But where did I go from there? Looked around in all directions, and finally found it:
|
|||||||||||||||||||||||
Hand cleaning station? Danger? Keep out? That's just to fool intruders. It's the little red sign in the distance.
Yvonne is not feeling well:
|
|||||||||||||||||||||||
Since I spoke to Lara, it seems that things have changed. They want to try fitting the brace today, and if everything goes well, there's a possibility that she will be able to come home tomorrow. I'm not holding my breath. They have also told Yvonne that it'll be a year before she can ride again, which does nothing to help her frame of mind. I'm just glad that she won't be in a wheelchair.
I had brought euroa, the Microsoft laptop with me, and sure enough, the hospital has an 802.11 wireless network for patients. And sure enough, I couldn't connect. No WPA key, just a disclaimer to accept. Or so they say: I never got that far. Part of it seems to be that I had one of four wireless interfaces set up with a static IP, but even after I fixed that, I still couldn't connect. Never mind, Yvonne's not really in a position to use a computer at the moment.
|
Image stabilization in practice
|
Topic: photography, opinion | Link here |
Some of the photos I took this morning were in very limited light. Never mind, I was using the ultimate combination of image stabilized camera and lens: the Olympus OM-D E-M1 Mark II with the M.Zuiko Digital ED 12-100 mm f/4.0 IS PRO. Together, they say, I should get roughly 6.5 stops of image stabilization. So the 1/6 s I used should have compared to about 1/500 s without stabilization.
Not quite:
|
|||||||||||||||||||||||
|
|
|||||||||
With flash, all was well. For once, the flash wasn't obtrusive:
|
|||||||||||||||||||||||
|
|
|||||||||
But why are the results that bad? I have relatively steady hands, and I was careful not to shake the camera. That's really quite disappointing.
| Tuesday, 16 January 2018 | Dereel | Images for 16 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Yvonne's progress
|
Topic: health | Link here |
Called up Yvonne in the early afternoon. No, they hadn't fitted the brace: they had decided that it wouldn't be necessary after all, and tomorrow they would let her get up and walk around. So she wouldn't be let out today. She was feeling quite nauseous, and didn't want to see me.
Later, as requested, I got a call from Emilia, who stressed that she was a junior doctor on the team, and for my liking a little too forcedly enthusiastic. The nausea came from the painkillers—they think—and there's a good chance that she'll be let out tomorrow. At least she was able to get up and walk around a little today. But I'm left wondering why they continually change their minds. Maybe it's the result of differing opinions from the myriads of people who have been looking at her.
|
Beer consumption and pricing
|
Topic: brewing, opinion | Link here |
I know that beer is expensive in Australia, and I know that they drink a lot of it in Czechia. But these statistics from Statista surprised me nevertheless:
That suggests that beer costs USD 1.08 per litre in Czechia and USD 6.30 per litre in Australia. That seems reasonable (well, plausible: the price itself is unreasonable).
|
More XXX TV X configuration
|
Topic: multimedia, opinion | Link here |
I still don't have this XXX TV thing working properly with X. It seems that it continually reconnects, causing a brief delay in the output. The log files show continual blocks of this, about 2 to 3 times a second:
What I've established so far:
It doesn't just happen to me. I have found a number of reports on web. Possibly there are more, but it's difficult to find a search term.
It's not operating system specific. All the reports I have found were for Linux.
Some claim that it's driver version specific. One person (where? lost the URL) claimed that going back to an older version avoided the problem. That's plausible, because I first found the problem last July after upgrading the driver for existing hardware.
One person claimed that the problem was an interaction with devd. This is Linux, of course, but FreeBSD also has a devd, so that's worth checking.
It is hardware specific. I have established that it happens with both the XXX TV models I have here, and I've seen it with the old, now non-functional Bauhn TV. All of these devices have similar hardware. But it doesn't happen on my field monitor.
Some reports on the web suggest that the HDMI cable might be to blame. That may have been the case at the time, but I've tried different HDMI cables here, and they all behave the same.
So: where do I go from here? At some level it's an Nvidia bug. Try using the standard driver, like I did with lagoon in November. But the results were different: I couldn't get more than 1024×768 Yet Again.
OK, back to the nvidia driver. What about the devd claim? Stopped devd, and sure enough, I got this in the log file about every 10 seconds:
But the connect messages continued. OK, RTFM time. One of the options is:
Option "ConnectedMonitor" "string"
Allows you to override what the NVIDIA kernel module detects is connected to your graphics card. This may be useful, for example, if you use a KVM (keyboard, video, mouse) switch and you are switched away when X is started. In such a situation, the NVIDIA kernel module cannot detect which display devices are connected, and the NVIDIA X driver assumes you have a single CRT.
OK, try that. I selected DFP because that's what's connected, but it seems that it wanted DFP-1. But that's incidental. It didn't do what I wanted and ignore anything else. It chose DFP-1, though I hadn't mentioned it, and hung for some reason that I don't want to investigate.
The next option is:
Option "UseDisplayDevice" "string"
The "UseDisplayDevice" X configuration option is a list of one or more display devices, which limits the display devices the NVIDIA X driver will consider for an X screen. The display device names used in the option may be either specific (with a numeric suffix; e.g., "DFP-1") or general (without a numeric suffix; e.g., "DFP").
At least it mentioned the suffixes here. It also mentioned the device TV, which sounded appropriate, though it didn't say what it meant. But the limitation didn't work:
Not surprisingly, nothing changed.
Then I looked at Option "IgnoreDisplayDevices" "string", but that didn't do anything either.
Then I noticed the URL of the documentation: http://us.download.nvidia.com/XFree86/Linux-x86/340.98/README/xconfigoptions.html. That's wrong for a couple of reasons. Firstly, it's for Linux, and less obviously, it's for an older version (340.98, while I have version 384.98). OK, let's find the right one.
That proved to be almost impossible. None of the implied directories in that URL exist, all the way up to the home page http://us.download.nvidia.com/. Why did they do that? But then, the logical URL https://www.nvidia.com/ doesn't work either, at least not for me at the time of writing: it redirects to https://www.nvidia.com/en-au/geforce/products/10series/store/, which is 404.
Finally went to /usr/ports/x11/nvidia-driver/pkg-descr, which gave me the URL http://www.nvidia.com/object/unix.html. Finally I was able to find the complete documentation for this driver at http://us.download.nvidia.com/XFree86/FreeBSD-x86_64/384.111/README/index.html (note the version mismatch), and from there I could find the options page. But by then I was out of time.
What a pain!
| Wednesday, 17 January 2018 | Dereel → Ballarat → Dereel | Images for 17 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Flaky networking
|
Topic: technology | Link here |
Into the office this morning to find a surprising status on display :0:
What's that? It's the raw output from my NBN statistics page, and it's bad. Typical output should look like this:
The output I had showed (third column) that on average I could only reach three of five systems that I ping on a regular basis. And even the other end of my National Broadband Network link (last column) had ping times of up to 262 ms instead of the normal 30 or so, with the exception of the two where the ping failed altogether. Clearly something very flaky, and apparently relating to my link. It had started at 6:40. Pings targeted at the other end of the IP link showed a very high rate of packet loss:
=== grog@eureka (/dev/pts/38) ~ 6 -> date; ping -c 20 180.150.112.1
Wireless link problems? The NTD showed normal status. OK, we've been there before. Power cycle the NTD, restart dhclient. No change. Enter a ticket (slowly) with Aussie Broadband, my RSP, and wait.
After another 45 minutes, during which Yvonne called me on the VoIP phone and managed to make herself understood (but wasn't able to understand me), the connection came back to normal.
|
Going shopping
|
Topic: general, opinion | Link here |
The good news of the day: Yvonne is coming out of hospital. But first I needed to do the weekly shopping.
I'm clearly out of practice. We do our menu and shopping planning together, so I know the drill, but by the time I left for town, I had a list only about half the size of normal. First to ALDI, where I found about half of what I needed, and managed to pay by credit card ($0.20 surcharge!) without wanting to: it seems that these POS terminals automatically book from credit card if you wave the card at them, while if you insert it, you get the choice. Easy for people who do this all the time, but that's not me.
The cashier refused to reverse the transaction, so I called for the manager. And he was friendly enough, and prepared to do so, but he was really only a trainee manager, so he called somebody more senior, who reversed the transaction and typed in all the items manually. And that for $0.20! I may be stubborn, but I did feel that this time I probably overdid it.
Then to Woolworths, where I had more fun with labelling. Which of these beans is the cheapest?
|
||||||||||||||||||||||||
Answer (not overly surprising): the loose ones at the bottom. Yes, the ones at the top are $5 as well, but for 340 g ($14.71 per kilogram), while the ones at the bottom are per kilogram:
|
|
||||||||||
|
|
||||||||||
And yes, aren't the photos from my mobile phone excellent?
Also looking for papaya. All I found were halves, called “Paw paw”, a name that suggests the deep, distant colonial past. And with a claim that seems absurd: “Cut today, sold today”. Where are papayas grown in Australia? I'd suspect Queensland, suggesting two to three days from tree to shop.
The off to fill the car up with petrol, not without fun with the self-service checkout. It was quite hot today, and there was little traffic. Even the petrol station was deserted, and I was able to observe somebody who drove in behind me, though other pumps were unoccupied. Sometimes I wonder if I belong in this world any more.
Picking up Yvonne was straightforward enough, but my intention of driving her straight back home didn't quite succeed. First we needed to find a toilet seat at Omni Healthcare, and then to UFS for a startling number of painkillers. anti-inflammatories and nausea suppressors.
And while we were there, Yvonne thought of some of the things that I had forgotten, notably avocados, so across the road to Woolworths again and bought them. Then it occurred to me that I had forgotten to collect the dog food, so off round town to pick up that. Finally got home nearly 2 hours after I had gone to pick her up.
|
Yvonne home again
|
Topic: health | Link here |
So Yvonne's home again. The good news is that she seems in better state than I had feared. She can walk relatively normally, though sitting down and getting up are still a problem, and will probably be the case for some days. Clearly it was the right decision to take her to the hospital, but I wonder what would have happened had we not. She could have died, but she could also have ended up in much the same state as now, just in more pain. And that's probably what happened to many people in the past, and also happens nowadays in countries where they can't afford the medical care.
Looking at her medication was instructive. The pharmacist spent quite some time telling me of the dangers of most of them, and I wrote notes on them (why don't we get the information in writing in the first place?). It wasn't until I got home that I found that two of them were very similar: brand name “Endone”, containing 2.5 mg oxycodone hydrochloride, and to be taken with care, not only because of constipation, and brand name “Targin”, for which I was given no specific warnings: and it also contains the same quantity of oxycodone hydrochloride, this time along with 2.5 mg naloxone hydrochloride. Why no warning? And why the Endone in the first place?
Put Yvonne in the guest room for the night, since the doctors had expressed concern that she might not handle getting up out of her water bed easily. But in the middle of the night I found her up and letting in the dogs, and she decided the guest bed was too hard, so she went back to the water bed, apparently without ill effects.
| Thursday, 18 January 2018 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
NBN outage follow-up
|
Topic: technology, opinion | Link here |
Call from Will at Aussie Broadband today, following up on my problem report yesterday—after 24 hours! I would have expected a more timely response, but as it was, of course, the problem “went away” without help from Aussie or me.
Interesting discussion, in which he asked for more information, including MAC address (why?) and confirmation that the interface was configured for autonegotiation. All that comes from the ifconfig output, and of course we were able to confirm that it was correct.
He also considered the possibility that the issue was congestion. That's possible, of course, but it would have had to be between the ODU and the POI (National Broadband Network-speak for Outdoor Unit, the antenna on our roof, and Point of Interconnect, where they connect to other networks). And that should be significantly overdimensioned. In addition, the times of day weren't exactly typical for high traffic.
Also discussed typical latency across the link, and he came up with pretty much what I experience: 15 ms to 50 ms. And that seems to happen at random. We considered that probably there are multiple internal connections between the tower and POI, and that they could be used at random. That could also have a bearing on the problem.
Before sending in the additional information, though, I needed to fix up my network statistics page, which has been broken for a long time. I didn't realize how long: I had commented out the update of the weekly stats in April 2015, before we moved to Stones Road. And the other ones no longer worked, because gnuplot has changed its understanding of dates—or so it seems. Somehow these plots are really hard to understand.
|
Carry a camera!
|
Topic: photography, opinion | Link here |
When I was a lad, some entity in the UK, probably Amateur Photographer, came up with the slogan “Carry a camera!”, and I very much went along with it. In general, I have done so again since I got my first digital camera nearly 20 years ago. But in the last couple of years I have changed my handbag and my cameras, and things no longer fit. Yesterday I was caught out without a camera when I took the photos of the vegetables, and it shows. Peter Jeremy asked me why I didn't carry one again.
Why, indeed? After all, I have the Olympus E-PM1, and with the Panasonic Lumix G 20 mm f/1.7 it seems smaller (though not lighter) than the old Nikon “Coolpix” 880 that I carried with me 15 years ago:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Isn't the difference in monitor size impressive? The CoolPix appears to be bulkier, but in fact, at 5.56×10×7.82 cm, it has a box volume of only 434.8 cm³, while the E-PM1, at 6.94×11×6.54 cm, has a volume of 499.3 cm³. Still, it's comparable, so I'll try it for a while. It's also a good use for the 20 mm Lumix, a lens that I hardly use.
In the process, it's interesting to compare the cameras from a photographic point of view:
| 880 | Samsung phone | E-PM1 | E-M1 Mark II | |||||
| Date of manufacture | August 2000 | 2011? | February 2010 | December 2016 | ||||
| Standard ISO | 18°/50 | 19°/64 | 24°/200 | 24°/200 | ||||
| Maximum ISO | 18°/50 | 19°/64? | 42°/12,500 | 45°/25,000 | ||||
| Resolution | 3 MP | 8 MP | 12 MP | 20 MP | ||||
The interesting thing is that the E-PM1 is not that much newer than the CoolPix—it's nearly half its age. But in the 8 years since the E-PM1 was made, the parameters haven't changed that much. The OM-D E-M1 Mark II has 20 MP and a maximum ISO of 45°/25,000, not that big a jump for 7 years. Progress has moved to other features.
|
Opioid crisis
|
Topic: health | Link here |
Why did the doctors prescribe Yvonne two different medications, both of them containing oxycodone hydrochloride? And why was the one with additional ingredients (Targin) slated for consumption twice a day, while the other (Endone) “only” every 4 hours? And why did the pharmacist warn against overuse of Endone, but not of Targin?
Went looking. That was instructive on many levels. Oxycodone is one of the key drugs in the current US opioid epidemic. And that's all that there is in Endone. Targin, on the other hand, contains naloxone hydrochloride, which, to quote Wikipedia, is a medication used to block the effects of opioids. And that explains the statement on the box, which I have never seen before: “POSSESSION WITHOUT AUTHORITY ILLEGAL”.
OK, that makes sense. And then I discovered that Yvonne had been taking the Endone, but not the Targin. Why? Unclear markings, unclear recollection of what the doctor said. She switched, but that difference could have been serious. Once again: why don't we get the information in writing in the first place?
|
Heat and smoke
|
Topic: general | Link here |
Another hot day today, with a top temperature of 42°. Ideal bushfire weather. And towards
evening we smelt (wood) smoke and saw haze. But where did it come from? The appalling
bushfire general emergency
map showed nothing within 30 km. And I've learnt better than to call the emergency
numbers unless I can actually see flames (by which time, of course, it would probably be too
late).
|
Trump of the week
|
Topic: politics | Link here |
Sadly this image has been lost in the mists of time. It came from Facebook with the URL https://scontent.fmel2-1.fna.fbcdn.net/v/t1.0-9/26239345_10155170281073314_1580132511722798607_n.jpg?oh=c700068a1734b3fa523852f1f16b7e99&oe=5AF41032.
| Friday, 19 January 2018 | Dereel | Images for 19 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
The difference between life and death
|
Topic: animals, opinion | Link here |
Decades ago Yvonne used to ride without a helmet. That terrified me, and I worked on her to ensure that she always used a helmet. That paid off. This is what the helmet looked like after her accident on Sunday:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
As it was, she didn't even have concussion. Without the helmet, there's a good chance that she would have been dead.
|
NBN: the experts
|
Topic: technology, opinion | Link here |
Yesterday I added a link to the National Broadband Network web site in the outage article. I check all links, of course, but I wasn't prepared for this one:
|
|||||||||||||||||||||||
|
|
|||||||||
What an advertisement! And what grammar!
|
Dead pump
|
Topic: general | Link here |
Call from Jarrod at UPI today: the pump that I took in on Monday has has extensive water damage, and a repair isn't worthwhile. It seems that pumps aren't waterproof, and in particular the electrical side isn't protected against water. I wish they had told me that earlier.
|
Power fail! Fail! Fail! Fail!
|
Topic: general, opinion | Link here |
Today was the hottest day since I installed the weather station in this location: 43.5° maximum. Thank God for air conditioners! But round 14:45, the power failed. Out to start the emergency generator, which mercifully worked, and to hope that the failure wouldn't be long. The Powercor outage page was its usual useless self, showing the whole of post code 3352 blacked out, although the postal district has nothing to do with the distribution topology. The cause was “under investigation” for the duration of the outage, but the location kept changing: first Magpie, then Mount Clear, then the Buninyong-Mount Mercer Road, and finally (the way they put it):
As so often, all users south of Buninyong were without power, and they put in an estimated restoration time of 17:00, a brave estimate given that they didn't know what or where the failure was. Within 30 minutes, this had slipped to 18:00.
In the middle of all this, the generator stopped. I had half expected that: I hadn't filled the tank. Filled the tank, but nothing that I could do would restart the generator. And then I saw:
|
|||||||||||||||||||||||
What's that thing hanging down? On closer examination, it proved to be the spark plug lead. Where's the plug? From a normal perspective, you can't see it:
|
|||||||||||||||||||||||
But climbing on the ground, it's (barely) visible above the rocker cover:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Why did it come off? It didn't seem to fit very well, and it wasn't the time to investigate. Put the thing back on again, and it started immediately.
The temperature in my office climbed, and by 16:00 it had reached 27.3°. And then the power came back, at 16:03! Not that bad after all.
It stayed on for nearly an hour, until 16:57, by which time the office temperature had only begun to drop. Then, while I was watching the news on TV, it failed again and didn't come back until 17:26. No change on the Powercor web site, apart from the location. This time I didn't reboot all the non-essential computers.
A good decision. At 18:29, while I was preparing quite a complicated evening meal, requiring the deep fryer, and with office temperature still at 26.4°, it failed again! Out to restart the generator, which had lost its spark plug lead again. Back in, and almost immediately the power came back, at 18:32.
On with dinner preparations. And then it failed again, at 19:22. More vehement cursing, but this time I had barely got back from starting the generator when it came back, at 19:24. And, oh wonder, it stayed on.
Should I complain about this? The last two outages suggest an electrician taking the power down briefly to do some repair work. The estimated restoration time had gradually increased to 19:30, and they actually kept to that. The alternative would have been to have had the power off all the time. At least this way we were able to run the air conditioner. The real issue is that there is no way to know what is going on.
Well, you could go to the fake news, in this case the Ballarat Courier, which knew that it was due to politics and the failure of a power station. And potentially that would explain the second outage: it seems that the station really did fail, round 16:00, as their update at 15:40 predicted. But why should it only affect people who had already had what appeared to be a localized power outage? I can blame Powercor for not explaining the cause of the wandering power failure, but I think that the Courier editors are extrapolating their expectations with little concrete evidence.
| Saturday, 20 January 2018 | Dereel | Images for 20 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
nVidia bug cornered
|
Topic: multimedia, technology | Link here |
On today with my attempts to stop the nvidia driver from reconnecting to the TV one or two times a second. One of the suggestions I had seen (where? lost the URL) was to revert to an older version of the driver. OK, I can try that. Currently FreeBSD has three versions in the Ports Collection:
=== root@teevee (/dev/pts/1) ~ 13 -> pkg search nvidia-driver
I had nvidia-driver-384.98 (the latest) installed. OK, remove, install nvidia-driver-304-304.137, reboot, and see how that worked. Not good:
But there's still nvidia-driver-340-340.104. OK, install that, reboot, and it worked! Before I had continual:
The number in brackets is a time in seconds (since boot?). The messages were gone. Instead I had other messages that I recall from years ago:
That's almost certainly correct, but it was only once. Or was it? It occurred again at time 2173.262, 13725.539, 16071.715 and 20527.052, rather less than once an hour, but at seemingly random intervals. Why? It would be nice to know, but I can live with that. It's clear that there's something wrong in the XXX TV's behaviour. But it's equally clear that the newer driver can't handle it as well, and there are clearly many devices that show this behaviour.
|
Rebooting teevee: finger trouble
|
Topic: technology, opinion | Link here |
The first time I rebooted teevee, I saw, half a second too late: it was eureka, my “always up” machine! AAARGH! But there was nothing to be done with it. On my way to the office to watch the reboot, I thought of David Newall, who, years ago, once unnecessarily rebooted www.auug.org.au, in the days where the site still meant something. I ensured that wouldn't happen again by renaming the shutdown command to something that wouldn't be chosen by accident:
=== root@eureka (/dev/pts/5) ~ 4 -> mv /sbin/shutdown /sbin/shutdown-if-you-must
As if that wasn't bad enough, the system wouldn't come up again! Can't find /kernel!
After a bit of messing around, it seems that the BIOSUEFI setup had chosen one of
the external USB disks to boot from.
Disconnect them and all was well.
|
The ghosts that I called...
|
Topic: general, opinion | Link here |
While I was playing around with the X drivers on teevee, I got a call from a spammer (hint: you answer, and get no response in 5 seconds). But the number was valid, and he called back almost immediately. Not a spammer: David Newall, with whom I haven't spoken in over 10 years—just after I was thinking of him. He was in Horsham, on his way to Melbourne, and he had had trouble with Bluetooth (doesn't everybody?). Looks like he'll be here tomorrow.
| Sunday, 21 January 2018 | Dereel | Images for 21 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Server down again!
|
Topic: technology, opinion | Link here |
Into the office this morning and found a surprising lack of mail. Where did it get stuck? Connected to the external server and found:
=== root@www (/dev/pts/0) /home/grog 2 -> date; uptime
So the machine had rebooted yesterday evening. Why? /var/log/messages only showed:
I intended to get rid of those qpopper messages, but they're useful to show that the system is up. What I see from this log file extract is:
No explanation from RootBSD as to the cause of the problem, or even that it occurred. That's sad. But at least I got round to fixing the configuration files. The issue was a second address range (alias) on the main interface. At the shell level, you can configure the interface with:
=== root@www (/dev/pts/0) ~ 28 -> ifconfig xn0 208.86.226.86 netmask 0xfffffffc
=== root@www (/dev/pts/0) ~ 28 -> ifconfig xn0 192.109.197.81 netmask 0xffffff80 alias
Why a netmask of 0xffffff80 for the alias? Ah, that's magic. The other half of the /24 (starting at 192.109.197.128) is local at home, and it's part of various experiments I've been doing.
But that's not the way you do it in /etc/rc.conf. A typical entry there is
Most people don't use aliases, and it's not easy to find the syntax. But for every problem there's a solution that is simple, obvious, elegant and... wrong, like what I did the first time round:
That's silly, of course, and it should have kicked me as soon as I entered it. The second line overwrites the first. Logan at RootBSD “fixed” it to:
That looks better, doesn't it? Well, no, he changed the net masks too, and in particular the second one is wrong. But it didn't work. Why not? More investigation, during which I discovered that it's not worth looking in the handbook. The clue is in /etc/defaults/rc.conf, which contains a number of commented-out entries. The correct entry for the alias is:
Note that alias0, and not alias. It'll take a reboot to confirm, and hopefully I'll not have one of them again in the near future.
|
David Newall visits
|
Topic: general, technology, photography, multimedia | Link here |
David Newall and friend Julianne along to visit this afternoon. It's been over 10 years since I last saw him. Much active discussion about all things technical, including the new TV. David came up with a good way to report bugs about products with no clearly defined bug reporting mechanism: Facebook or Twitter. That sounds interesting. I suppose I should try that.
They were also surprisingly interested in my photographic equipment, particularly the field monitor, which I discussed in some detail on 5 December and 10 December:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Reporting nvidia bugs
|
Topic: technology, multimedia, opinion | Link here |
I have a regression issue with the nvidia graphics card driver: the current version 384 loops identifying the TV at frequent intervals. I've found a workaround (revert to version 340), and I'd like to submit a bug report, but I can't find out how. I searched the Nvidia web site without success, but Google pointed me at this page. But that's only for people developing software based on the drivers! It states:
For end user issues this is not the correct forum. There is http://forums.geforce.com and http://www.nvidia.com/page/support.html.
OK, I checked those pages. As the name suggests, the first is a general forum site. No obvious forum for bug reports. GeForce Drivers Board seems the closest, but that's all sorts of general help, not bug reports. The second link is the place for downloads and end user help. After a bit of further searching I found this page, which states:
You can create a post detailing any new issues you encounter in the forums or you can submit a bug report privately via chat at our support center. Our product and support teams are ready to assist.
Somehow that doesn't sound right. Before submitting a bug, you should do your best to check whether the problem isn't known already. And I don't want to “chat”, which implies maintaining a session until the bug is fixed. Where's the bug list?
OK, if they don't have their reports online, at least this one will be (next article):
|
nVidia driver bug report
|
Topic: technology, multimedia | Link here |
After searching in vain for an online bug reporting system for the nvidia X driver, I've decided to report it here. nvidia, feel free to contact me and explain where the bug database is really located.
Title: nvidia Unix driver loops identifying display
Description: The current nvidia Unix driver (version 384.98) identifies certain flat panel displays correctly, but continues to log identification information at the rate of up to one or two times a second. This causes performance degradation.
Severity: Moderate. Driver works, but performance is degraded and motion is jerky.
Operating system: FreeBSD, though others have reported the same problem with Linux. There is no obvious reason to assume an OS sensitivity.
Display card: NVIDIA GPU GeForce GT 710 (GK208). Some of your online information asks for a display card BIOS revision, but I can't see any for this card.
Driver version: NVIDIA GLX Module 384.98 Thu Oct 26 09:39:17 PDT 2017
Display device: Multiple televisions connected by HDMI. See description below.
Description: I have observed the problem with four different television sets. Three are of of related construction. They are marketed by ALDI in Australia under the name Bauhn. The oldest (about 5 years old) identifies itself in the EDID as “SKYDATA S.P.A. BAUHN TV”. The other two are marketed as Bauhn ATV65UHD-1217 and Bauhn ATV75UHD-1117. I assume that 1117 and 1217 relate to the MMYY date of introduction of the model. They identify themselves as “XXX TV”. Here's the EDID for the second model; the other is very similar:
The fourth model is marketed as “TCL 60"(152cm) UHD LED LCD Smart TV” and identifies as “Tech Concepts TCL SMART TV”. The behaviour with this device is somewhat different from the Bauhn devices.
The driver configures the displays correctly, and the display is correct. In the case of the Bauhn devices, at not completely regular intervals, less than a second apart, the display stutters, which is only visible with moving images. The log files show:
In the case of the TCL device, the messages are far less frequent, round once an hour, but also at very irregular intervals, between 50 ms and 2 hours. See link to log files below.
This problem does not occur with other displays that I connected, so it is clearly a display sensitivity, possibly related to TV displays.
This problem has been reported before, though the difficulty in describing it makes references hard to find. What I have found is openSUSE with GTX 970 and driver version 375.66. Display identifies as Philips PHL 223G5. On this occasion it seems that the problem went away when a different HDMI cable was used. I tried this on my installation with three different HDMI cables, none of which made any difference.
Elsewhere I found a reference, which unfortunately I could not find again, where a similar problem occurred with a Linux version. The submitter solved the problem by reverting to an older version of the nvidia driver. I tried this with version NVIDIA GLX Module 340.104 Thu Sep 14 15:41:27 PDT 2017, and the problem no longer occurs, though some reconnects still occur, roughly once an hour:
The xorg.conf file used in the latest version is at /grog/Day/20180121/xorg.conf. An X log file showing the problem is at /grog/Day/20180121/Xorg.1.log, and part of a verbose log is at /grog/Day/20180121/Xorg.0.log-logverbose6. A log file from the TCL display is at /grog/Day/20180121/Xorg.0.log.last-tcl.
| Monday, 22 January 2018 | Dereel | Images for 22 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Reporting the Nvidia bug
|
Topic: technology, multimedia, opinion | Link here |
Yesterday I got round to entering a bug report on the Nvidia X driver. Normally that's the difficult part. All I need to do is submit it.
But how? I've been through this before, most recently yesterday. They don't seem to have a web-facing bug tracking system. OK, the least ridiculous way to submit it appears to be by “chat” (thus giving an idea how ridiculous the other options are). Where's the chat? It's mentioned here, along with a big, green CHAT NOW button that takes me to a site custhelp.com, which proves to be run by Oracle. OK, worth a try.
The site is like a typical chat site, almost. First I had to enter a name, a subject line (“nvidia Unix driver loops identifying display”) and an email address. Then I got:
|
|||||||||||||||||||||||
OK, the first way to help me is to tell me how to enter a response. Simple: it's hidden below the too-small and unresizable window. After scrolling down and a bit of discussion, I had:
|
|||||||||||||||||||||||
It takes time and experience to digest a bug report. You need to read it all, think about the implications, and check if anything is missing. Chat is the wrong way to do that, and Kenny fell foul of it. Here's the complete transcript:
He asked no fewer than six different questions that had already been answered in the report. On one occasion he acknowledged that and asked anyway. My responses were almost all cut and paste.
I'll be interested to see where things go from here.
| Tuesday, 23 January 2018 | Dereel | Images for 23 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Power fail
|
Topic: general | Link here |
Another power failure this morning at 7:40. I didn't find out until later, so it can't have been too long.
|
Nvidia responds
|
Topic: technology | Link here |
Not one, but four messages from Nvidia today. The first two confirmed receipt of the report and that it had been sent to the Linux Support group (why, when they call the driver a Unix driver?). The third told me:
You can submit a linux bug by emailing linux-bugs@nvidia.comThere is more info here: http://us.download.nvidia.com/XFree86/Linux-x86_64/384.111/README/addtlresources.html
It seems difficult to get it across that this is a FreeBSD installation. But yes, with some searching, I found the corresponding page for FreeBSD, which told me:
If you believe that you have found a bug or have a problem that you need assistance with and cannot find the solution elsewhere, or if you have found inaccuracies in this document, send email to <freebsd-gfx-bugs@nvidia.com>
That makes perfect sense (except that it's not exactly a s/linux/freebsd/ on linux-bugs@nvidia.com). Why didn't I see it? It's in chapter 27 of a total of 38 pages on the driver, and the web site doesn't find anything useful.
The message didn't make it clear if they wanted me to submit the report again, and my question to this effect remained unanswered. The message also contained:
We will assume your issue has been resolved if we do not hear from you within 72 hours.
What did they want to hear? They didn't say. But my reply generated a fourth mail message, giving me details of how to access the bug report, which, it seems, can only be accessed by people connected to the case. I suppose that's one way to ensure that bugs get reported multiple times.
|
No faith in American leadership
|
Topic: politics, opinion | Link here |
Another sign of the times from Statista:
|
Valvoline web site pain
|
Topic: general, technology, opinion | Link here |
Our Holden Commodore has low power steering fluid level—in fact, I could barely find any at all. Problems, or normal leakage over several years? Hard to say, but clearly the first thing to do is to refill and see if it stays that way.
Chris Bahlo can sing a song about that with her Toyota Landcruiser, which is old enough to have voted in the last three or four federal elections. She carries fluid with her at all times. But is it the same fluid? She sent me a photo indicating that it was part number 8631 LB8631, though maybe there was a different indication on the other side. Off to the Valvoline web site to look. I couldn't make head or tail of the site, and in the end I called the number that was also written on the label. Quick check: no, not suitable.
So why didn't I find that? The consultant told me to go to some field on the home page. But it wasn't there. Ah, did I enter http://valvoline.com.au or http://valvoline.com? Guilty as charged. Went to http://valvoline.com.au and got the message “the owner of valvoline.com.au has configured their web site improperly”. Fought my way through that and was presented with a login screen.
Tried again on a less fussy browser and was automatically redirected to https://www.valvoline.com/en-australia, which showed me what appears to be exactly the same home page as the standard US American version. Finally I found what I was looking for. As I asked, why do I need to call up their help line to find this? I asked him to report the problem, and he said yes, and sounded like no.
And then I read what it said:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Automatic transmission fluid!
It seems that:
And Valvoline is all three!
Well. Discussed the matter on IRC, and Juha Kupiainen told me that yes, indeed, you really use automatic transmission fluid in Commodores. OK, but in that case they should make it clear, rather than just repeating their blurb for automatic transmissions.
| Wednesday, 24 January 2018 | Dereel → Ballarat → Dereel | Images for 24 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
New taps
|
Topic: general | Link here |
Marc van Gils from Cape Clear and a helper (son?) along today to change the silly (mixer) taps in the kitchen and bathroom. It wasn't as simple as they thought, and the whole thing took them two hours. The hole in the sink needed enlarging, and it seems that he did it with tin snips rather than the promised punch. He also managed to tighten it up so much that I can barely swivel the lower part. Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
For some reason, probably at least because of the strange shape of the wash basin, the other tap took them a very long time. He had never seen a standard European plug mechanism before, and I had to show him how to put it together, not once, but three times. And though I had suggested that he keep the original hoses, he put the new ones on the tap—and then found that the connection at the other end was wrong, so he had to do it all over again.
Finally it was done, and I discovered that he hadn't bothered to clean the surface of the wash basin before mounting the tap, and that the new tap and drain didn't quite cover the holes in the basin:
Image title: Tap 4 Dimensions: 5185 x 3888, 2794 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Wednesday, 24 January 2018, thumbnails All images taken on Wednesday, 24 January 2018, small Diary entry for Wednesday, 24 January 2018 Complete exposure details
Image title: Taps 3 Dimensions: 5184 x 3888, 4827 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Thursday, 25 January 2018, thumbnails All images taken on Thursday, 25 January 2018, small Diary entry for Thursday, 25 January 2018 Complete exposure details
Image title: Taps 4 Dimensions: 5185 x 3888, 5149 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Thursday, 25 January 2018, thumbnails All images taken on Thursday, 25 January 2018, small Diary entry for Thursday, 25 January 2018 Complete exposure details
Image title: Drain Dimensions: 4348 x 3880, 2313 kB Make a single page with this image Hide this image Make this image a thumbnail Make thumbnails of all images on this page Make this image small again Display small version of all images on this page All images taken on Friday, 26 January 2018, thumbnails All images taken on Friday, 26 January 2018, small Diary entry for Friday, 26 January 2018 Complete exposure details
The little blue mark at the top of the drain is part of the protective foil that they didn't bother to remove. Still, we can live with that. Both are much better than what came with the house. The funny thing is how small they appear, though of course they're normal, while the ones they replace are ridiculously large.
They were also too polite to disturb me when they spilt some water, so they wiped up some of it with my bath towel, and when that was soaking, left the rest, including (as I discovered later toilet paper and medication packages) inside the cabinet, to drip dry:
|
|||||||||||||||||||||||
Finally, they filled in the hole that Sasha made years ago. That didn't take long, helped by the fact that they didn't bother to tidy up:
|
|||||||||||||||||||||||
|
Shopping again
|
Topic: general, photography, opinion | Link here |
Off to town in the afternoon for shopping and other things. The steering of the car was the most important thing, but before going to Ballarat Automative it occurred to me that if the fluid was only borderline, it would make more sense to refill the reservoir first and ask questions later.
First to Woolworths for normal shopping, including looking for power steering fluid. I didn't find that, but I did have an opportunity to observe the do-it-yourself checkout more carefully. Last week I had been confused by the payment terminal not wanting to do anything obvious. This week I looked more carefully.
Scanning the items works relatively well, but when you're done there's an arrow pointing to the right to the payment terminal:
|
|||||||||||||||||||||||
Clearly that's an indication to continue on that terminal. But no. You must tap that arrow. Then you get another window telling me that I have the wrong brand of mobile phone:
|
|||||||||||||||||||||||
And then I have to select the means of payment, something that you normally do on the payment terminal. Only then are you allowed to actually use the terminal, indicated by the change from advertising to instructions on that terminal:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Somehow that's less than obvious, both the intention and also the rationale. In passing, while processing the photos, I found this:
|
|
|||||||||
It's a lot easier just to look over your shoulder and wave at the guard
assistant.
And the information on the screen seems to have been translated from English into some code that must warm the hearts of budding boy cryptographers:
|
|
|||||||||
After leaving the shop, I was wondering whether I'm not becoming too old and intolerant. And then I found an old woman complaining about the way some people park their trolleys:
|
|||||||||||||||||||||||
Yes, silly, and it extended behind her car. “What's the world coming to?”. I had found it rather indicative of the stupidity of some people, but I held it out of the way while she drove her car out:
|
|||||||||||||||||||||||
|
Medication instructions
|
Topic: health, opinion | Link here |
Last week I discovered that Yvonne had been prescribed two different opioid medications, and that the instructions were anything but clear. Gradually it's time to replenish other medications, notably paracetamol. That's not as dangerous as oxycodone—maybe. On the one hand, it seems that opioid abuse kills more people than heroin. In Yvonne's circumstances, the biggest danger is addiction. But she has been taking up to 8 tablets of paracetamol a day, with intervals as short as 4 hours. The tablets weren't the ones prescribed to her: she had them already, so all she had was verbal instructions.
The tablets have the positively stupid name “Osteomol”, apparently to suggest that they have something to do with osteoarthritis. In fact, they're straight paracetamol, but with 665 mg instead of the Australian normal 500 mg. They're made by Pharmacor, a company with what appears to be a deliberately non-functional web site.
Paracetamol is one of those underrated drugs. It's generally agreed that the maximum daily dose is 4 g, and it seems that even this amount causes significant issues with the liver when taken over a long period of time. Yvonne had been taking up to 5⅓ g, risking acute liver failure, a fate worse than addiction. Why?
Lack of instructions, and unclear instructions. The chemist told her that the tablets were the same as “Osteomol”, and that she should take up to 8 tablets per day—she says. But she was hardly in the best of form at the time. And since there were no repeats, we didn't even get the prescription back. While I was in town, I dropped in to UFS and asked. After about 10 minutes, they dragged out the prescription. No, the prescription was for 500 mg tablets.
The more I see this, the more it's clear that the current information isn't sufficient. Currently the Therapeutic Goods Administration is running a consultation on the opioid epidemic. They have suggested 8 courses of action (which they term options). Option 5 might be what I'm thinking of: “Review of label warnings and revision to Consumer Medicines Information”. But somehow that's the wrong way round. The labels on “Osteomol” do contain almost all the required information, but it's diluted with other things (“risk of serious liver damage” way below “don't use if the package is open”). If cigarette packages can carry graphic depictions of the health effects, why can't medications contain more obvious warnings? And why can't patients receive documents tailored to their condition, without all the boilerplate?
|
Fixing the steering
|
Topic: general | Link here |
Finally found some transmission power steering fluid, ATF DX-3, made by Valvoline,
at Dahlsens Gays. The
salesperson told me that it was the wrong stuff: after all, it had “transmission fluid”
written on it, and nothing else. I explained that I had been there before, and
that Commodores used it in the
steering. He still didn't believe me, though he told me that he drove a Commodore too.
OK, that's simple enough: call Valvoline and ask them. Yes, the right stuff. Out and put some in. How much? Hard to tell. Holden keep that information secret, and the dipstick is almost impossible to read. But it seemed to work, and the steering was lighter again, though not light enough. After picking up the food at Davis, I checked:
|
|||||||||||||||||||||||
That's more than just topping up. Off to Ballarat Automotive and got there just before closing time. Paul took a look and said “there's often an issue with a loose hose here”, pointing to pretty much where the leak was. He crawled underneath and said “It's burst! I've never seen anything like that”. He jacked the car up, and I confirmed:
|
|||||||||||||||||||||||
Later I got better photos of the hose. It looks as if something has cut it in three places about 1 cm apart (the third, on the right, is barely noticeable):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I had discovered the leak (first photo) at 16:42. The second one was taken at 16:59, and by 17:10 I was on my way again with the fault repaired. Total price: $30, including some fluid. I had paid nearly $12 for the fluid alone. You can't complain about that.
|
Duck eggs
|
Topic: food and drink, opinion | Link here |
The final task was to pick up some duck eggs for Yvonne, down Ferrers Road. What number? First on the left after Arnolds Road. I can't put that in the GPS. What number? 179.
Got there, drove straight past it. Not surprising. It doesn't exist. It's really number 169, as indeed was written on the egg box I had to return. And I know that house: we visited it with a view to purchase in August 2012. Since then it has changed hands twice. Met MJ Wisbey, the duck egg woman, who had baked a duck egg cake for Yvonne as a “get well soon” gesture. That was nice of her.
|
Photos: mobile phone or real camera?
|
Topic: photography, opinion | Link here |
As planned, I'm now carrying my Olympus E-PM1 with the Panasonic Lumix G 20 mm f/1.7, so today I was able to compare the performance with last week's photos taken with the Samsung Galaxy S II. Here the same display of beans, last week with the Samsung, then this week with the Olympus:
|
||||||||||||||||||||||||
|
|||||||||||||||||||||||
And here some crops of the price tags:
|
|
|
||||||||||
|
|
|||||||||
Yes, the results are better, but not as much as I had expected.
| Thursday, 25 January 2018 | Dereel | Images for 25 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Housework
|
Topic: general, opinion | Link here |
One of the agreements we came to when we moved out of Wantadilla was that I would do less work, and that Yvonne would take care of the household. Clearly that's not possible now in her current state, and I've been doing most of it myself. It's a new experience. I'm sure that when I get used to it, it won't be as much work.
In particular, I'm amazed how often we need to run the dish washer, about 1.6 times a day. When we got our first dish washer in the mid-1970s, I was talking to Ferry Lang, our Big Investor, about it. He said “don't worry, you won't need to run it more than once or maybe twice a week”. How times change.
| Friday, 26 January 2018 | Dereel | Images for 26 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Austraya Day
|
Topic: general, opinion | Link here |
230 years ago today the First Fleet entered Port Jackson, forming the first permanent colony of Europeans on the Australian continent. As a result, it is now celebrated as Australia Day.
Well, by most people. Many of the former inhabitants of the continent, once called Aborigines, then (apparently more politely) Aboriginals, and now Aboriginal and Torres Strait Islanders, claim that it was the day of disaster, when the white people took over their continent. Yes, that happened, but I can't help feeling that the complaints are rather artificial. Did the original inhabitants celebrate anniversaries? Did all of the indigenous nations suffer because of the arrival? Other places celebrate disasters (“Remember the Alamo”). I'm left with the feeling that this is a rather contrived disaster.
From our point of view, today was a non-event. Business as usual. About the best thing about it is that Yvonne is gradually feeling better.
|
Power fail? fail? fail?
|
Topic: general | Link here |
No less than three power failures today between 16:11 and 16:23. Or were there? They were sufficient to take out my printer, which isn't on UPS, but not enough for the kitchen equipment with the clocks on it. And I heard a couple of suspicious beeps at other times too, so maybe they were further ones.
How serious is that? If the printer had been printing, it would have been an issue. My main concern is that they might be indications of worse to come, but today it didn't.
|
Rating world leaders
|
Topic: politics, opinion | Link here |
I have been under no illusion that Donald Trump is making the USA irrelevant again. I know nobody outside the USA who thinks he is doing anything but damage. And even those few in the USA who disagree with me have no coherent arguments.
In the last day or two Gallup have released a report that roughly confirms my opinions: there are still a number of people in the USA who think Trump is doing a good job, but outside the USA the opinion of US leadership has dropped significantly—in some cases, such as Australia, to an all-time low, where only 19% approve of the US leadership, and 65% disapprove. World-wide, more people (43%) disapprove of the US leadership than of the Russian (36%) or Chinese (30%) leaderships. Interestingly, in Russia and China, the impression has increased. I'd be interested to understand the background for that.
|
Focus stacking again
|
Topic: photography, gardening | Link here |
It's been some time since I've done any photography with focus stacking. The main reason has been that I'm trying to understand how to set things up. What aperture? What focus step increment? How many shots? I still have almost no feeling for that.
Today I found that the flame-red eucalypt on the corner of Stones Road and Progress Road wasn't dead after all. Last spring a eucalyptus tree had been knocked down by a storm, but it seems not to have been the one I admired:
|
|||||||||||||||||||||||
How about some photos of the flowers? They're about 2 cm across, but out in the wind and shade the results aren't spectacular:
|
|||||||||||||||||||||||
OK, take a sprig home and try some properly focus stacked shots, this time with the ring flash: I'm a little concerned about the life of the flash tubes in the studio flash if I take dozens of shots at a time. And just setting that up kept me busy for the rest of the afternoon. I'll have to process them tomorrow.
| Saturday, 27 January 2018 | Dereel | Images for 27 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More focus stacking
|
Topic: photography, gardening, opinion | Link here |
Got round to stacking my eucalypt flower photos, and on the whole I'm happy with the results. Here first in the wild, and then the focus stacked versions:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
My concerns about the number of shots were justified: I took 40 of the last image, but I only needed 6, numbered from 6 to 12. Somehow there must be a way to determine the right number more accurately.
The other thing that surprised me was the difference in the results from the PMax and DMap algorithms. In this case, DMap produced far inferior results to PMax (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
Cosy winter dinner
|
Topic: food and drink, opinion | Link here |
Raclette for dinner tonight, befitting for a cool winter evening, and most certainly not the 35° that we had today. In the process, I was able to marvel at the mess that they make of sliced ham and cheese (“Swiss”, more specifically Maasdam, I'm told) at the supermarkets. Yvonne pays attention to that, so it's not usually as bad. I saw it when I bought it, but I thought that that was normal. You'd think that the foodstuffs had personally insulted the people who cut them.
| Sunday, 28 January 2018 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More network pain
|
Topic: technology, opinion | Link here |
Why does it take such a long time to download files from German TV stations? In general I can't get more than 500 kB/s, and though it's intended to be viewed online, the speed can't keep up.
Today, though, things were much worse. Downloading the Tagesthemen ran at about 30 kB/s, and firefox gleefully told me that it would take 6 hours. Where's the bottleneck?
Looking at my network statistics showed a familiar sight:
That shows 4 minutes in a row with at least one system not reachable, and with ping times to the National Broadband Network POI of up to 236 ms—a time that I normally need to ping www.lemis.com in the USA. OK, time to reply to that last message regarding ticket. To my surprise, I got a response almost immediately—on a Sunday afternoon on a long weekend:
To get this raised to NBN in regards to the packet loss issues that are occurring, I would suggest running the following tests:We wish to perform some tests from 2 computers directly connected to your NBNco NTD. Please ensure your Computers have a firewall active at the very least as well as antivirus software and are plugged directly into the NBNco NTD uni D port that your router was connected into via an ethernet cable. Please ensure that nothing else is using any bandwidth on your devices ie;Speed tests, Youtube, file sharing programs such as BitTorrent/UTorrent or any files downloading. These can affect your results and NBN may reject a ticket due to this.
First, open the 'Command Prompt' by Opening the 'Start' Menu, and typing 'cmd' into the search box at the bottom, and clicking on 'cmd' in the search results list.
Once you are in the command prompt, please copy and paste or type the following:
ping 202.142.142.142 -n 100
ping 180.150.17.170 -n 100
ping 8.8.8.8 -n 100
ping www.google.com -n 100
tracert 202.142.142.142
tracert 180.150.17.170
tracert 8.8.8.8
tracert www.google.com
How Microsoft-centric! The second program is called traceroute everywhere except with Microsoft. And it wasn't until I copied and pasted the first ping command that I realized that it's all wrong: -n specifies no reverse lookup, and the count parameter is -c. And they have to be before the address:
And the second machine? I suppose that makes some sense—but only in the case where you're not using a router: the problem could be this side of the NTD. So I went to quite some difficulty to connect the old tiwi to the NTD, and... nothing. I had set the IP address manually, and I could see traffic with tcpdump, but the DHCP server didn't respond to any of my requests. Cached MAC address? That would be my guess.
But why all this anyway? Firstly, it's effectively what I've been doing (except that I chose my addresses more carefully), and secondly only one ping is of interest, and for that system traceroute is trivial. Still, I did it, and how about that, I got packet loss rates of between 5% and 10%. Sent that off, and this time didn't get an immediate reply.
OK, maybe it's time to reconsider my network stats page. It has essentially grown over four different link media, from ADSL, Satellite IP, Mobile broadband to NBN. But the issues have changed: previously I had never had this much packet loss. OK, modify the scripts to do multiple pings and calculate the packet loss.
Previously my file /grog/net/nbnlinkstats looked like this:
To this I added another column, number 4:
Now the columns are:
And so I was able to create what was ultimately a much easier graph to read. Here before and after:
A couple of things are strange here. Firstly, the packet loss rates look very high. With that much packet loss I wouldn't expect any useful traffic at all. And second, unlike my problems 10 days ago, the ping time to the POI wasn't significantly elevated. But it was dropping packets anyway.
What can this mean? Can it be that whatever is dropping the packets is preferentially dropping ICMP? One of the things that struck me then was that the time of day did not correspond with my expectations of heavy traffic. But now, looking at the weekly graph, things appear different:
That's still the old-style graph, but it appears to show a clear correlation between status and time of day, being worst at the times when demand is highest. That doesn't make things easier. That could be Aussie Broadband's problem too.
|
Power fail!
|
Topic: general | Link here |
I had just gone off to sleep when Yvonne woke me again, at about 22:55. The power had failed. Dammit, I'm not going out to the shed to start up the generator again. Off to shut down eureka before the UPS failed, in the process tripping over both my renamed shutdown and NFS hangs. In the end I had to use halt.
Then turned off lagoon, Yvonne's machine. And then the power came back! All that effort for nothing!
| Monday, 29 January 2018 | Dereel → Ballarat → Dereel | Images for 29 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Rebooting
|
Topic: technology, opinion | Link here |
First thing this morning was to reboot eureka, which had been down for one of the longest periods ever. Things came back pretty quickly, sort of: connecting to the National Broadband Network failed first time round, and I had to stop and restart dhclient. This has nothing to do with yesterday's caching issues: this was the same interface and thus the same MAC address.
And then officephone.lemis.com, my VoIP ATA, couldn't connect. Examination showed it sending registration packets to the SIP server, but getting no reply.
Dammit, why do these things happen when I have other things to do? Spent a lot of time trying to contact the server. Yes, no problems. Then it occurred to me: natd, the NAT daemon. Wasn't there an issue there where it would silently fail if the interface wasn't configured? Yes, I had no network access from lagoon either. It only worked on eureka because I was using the external interface address and thus not performing NAT. Stopped and restarted natd, and all was well. And yes, I had noted this in the past, though in that case the configuration had changed, and I hadn't considered it a bug. I should take a look at that code.
Just by coincidence, received mail from Warner Losh about the reboot program. Not an area I know well.
|
More network monitoring
|
Topic: technology | Link here |
Yesterday's network monitoring experiments took a while to show results: I needed to gather the statistics. This morning things looked better, but it became clear that there was still significant packet drop, and that 5% steps just didn't cut it. So I did a bit of tuning—ping -c 100 instead of ping -c 20—and came up with some graphs that looked a lot more reliable:
But why so much packet loss (scale on the right)? It goes up to 10%! That should cripple the link, but in fact it wasn't that bad. Are the numbers right? At least one issue is that I'm looking at ICMP packet loss, and what I really want is TCP packet loss. How do I measure that? Real-life measurements are best. Does the FreeBSD network stack have knobs (sysctl) that measure retransmits and things? Off to look for documentation, but didn't find anything.
Feedback from Aussie Broadband, addressing some of my issues with yesterday's “tests”, but giving me the impression that they had not taken note of the results. Also a separate reply from somebody else, confirming that the DHCP server cached the MAC address, and wouldn't respond to anything else (until when? He didn't say).
|
Ballarat again
|
Topic: general, multimedia, gardening, food and drink | Link here |
Off to Ballarat in the afternoon for a number of chores. The first stop, to have my hair cut, was a non-starter: the place was packed, and I'll be in town several times again this week, so I'll do it then—maybe.
To ALDI to return the TV I bought last month, but discovered that nobody was prepared to help me get the TV out of the car and into the shop: against OH&S regulations, it seems. That sounds highly dubious, and if true, a clear indication that the insurance isn't appropriate. Somehow managed to get the thing out of the car and into the shop without any obvious damage or lost parts, but it left me with a rather negative impression.
At UPI the news wasn't good. I knew that I needed a new pump, but the cheapest I could get was round $800. In addition, what I really need is a DC pump that starts without the current surges associated with AC pumps, so that I can run it off the photovoltaic solar electricity system that I really must get installed soon. No help. But Chris Bahlo has one! And she bought it from UPI. After returning home, over to take a look at Chris' pump. Not DC after all, in fact very similar to mine. The only difference was that it has a “pressure cell” instead of the flow controller on mine:
|
||||||||||||||||||||||||
Somehow that's not very satisfactory. And I'm concerned about the prices. I need to look around, but since I have the other pump on loan from them, I can't wait too long.
Then to pick up a steamer that I had found on Freecycle. Why? I have two already, a Chinese one out of bamboo, designed to sit on top of a saucepan of boiling water, and a French one out of enamelware that incorporates the saucepan. But this one is electric and will hopefully be able to moderate the steam better. It's also transparent, so you can see what's going on:
|
|||||||||||||||||||||||
While I was there, I noticed a number of shrubs that interested me:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
What are they? I've seen the second one, in fact, just last week when picking up the duck eggs, and I had brought a sample with me. This time I took real cuttings, and I'll see if I can propagate them.
| Tuesday, 30 January 2018 | Dereel | Images for 30 January 2018 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More network monitoring issues
|
Topic: technology, opinion | Link here |
Into the office today to find a strange looking graph of the network link stats:
What went wrong there? It started at midnight, the time when my log files wrap round, so it seems reasonable to think that I broke something in my updates. But no, the log files looked normal. It took me quite some while to realize that this is the way it should look: connectivity 5 at all times, and packet loss 0. It just looked so strange.
So what changed? It almost looks as if somebody fixed something at midnight (a common time for the National Broadband Network to do such work). It also seemed as if the link latency had become less. But as I watched during the day, I saw a number of dropped packets, if not as bad as previously. Still more observation needed.
|
Huevos a la flamenca: otra vez
|
Topic: food and drink, opinion | Link here |
Last month I made some fake huevos a la flamenca, but I have serious doubts as to the authenticity. I've been doing some more research since then, and in the end I came up with two recipes that looked good (in other words, the photos were attractive). Printed them out and tried to interpret them in a direction that suited me.
How different they are! This one, which uses mixed metric and archaic measures, wanted 15-25 g of tomato for four people.
This one came from SBS and, with typical SBS accuracy, was called “Huevos a la flamenco”. For the same number of servings it wanted 750 g of tomatoes and “1 tsp” of tomato paste, between 30 and 50 times as much as the other recipe. Interestingly, unlike the (Spanish) basis for last month's recipe, neither recipe wanted any potatoes.
What to do? Use what I have. I ended up with, for two people:
| quantity | ingredient | step | ||
| 120 g | onion | 1 | ||
| 7 g | garlic | 1 | ||
| olive oil | 1 | |||
| 350 g | cooked potatoes | 2 | ||
| 70 g | bacon | 3 | ||
| 80 g | tomatoes | 3 | ||
| 90 g | peas | 3 | ||
| chili (optional) | 3 | |||
| 20 g | capsicum strips | 4 | ||
| 80 g | chorizo slices | 4 | ||
| 2 | eggs | 5 | ||
That's the way I did it this time:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The results were acceptable. But clearly grilling the eggs is suboptimal. The yolks got hard in places, elsewhere the whites were still raw, and elsewhere again they were tough as leather. I think I'll give up grilling and just fry the eggs.
|
Cakes for invalids
|
Topic: general, animals, food and drink | Link here |
Diane and Garry Marriott over this afternoon, Diane bringing a cake for Yvonne, and Garry bringing the terriers. Our dogs ran for them and terrified them, upon which Sasha (the terrier, not our Sasha) took it out on Mishka, and Garry decided to take them home again after all.
|
Smart TV pain
|
Topic: multimedia, opinion | Link here |
While the Marriotts were at our place, we got to discussing smart TVs. As I had noted last month, they also have a new “smart” TV, and they're having their trouble understanding it. So I came back with them and tried to find my way round it. The remote control is interesting: it has a mouse function where mouse movement is simulated by tipping the remote control in the desired direction. I didn't come to terms with it in the time I was there, but potentially that's a good idea.
Apart from that, though, it didn't seem any better than Android TV. It had one of these appalling on-screen toy keyboards, and though we had both the keyboard on the remote control and a USB keyboard that I had brought with me, there were places where the on-screen dialogue required mouse clicks. Garry wanted to sign up for SBS on demand, and it was excruciating. It seems that the signup screen thought so too, and after about 5 minutes it hung.
OK, “smart TV” is a concept that I see either dying (unlikely) or mutating into something useful, but thus very different from what we have now. Garry had an HDMI cable, and Di's laptop actually has an HDMI output, so we were able to connect it with no difficulty, beyond noting that the INPUT button on the remote control displays two completely different menus (both of which have the same function) depending on the context. And how about that, it displayed with no problems, although the laptop has a resolution of 1366×768. Then the only issue was getting anything useful out of Microsoft “Windows” 10, which in this case proved very painful. Google? Yes. eBay? Yes. Youtube? Yes. Web browser? No. Or so it seemed. It took me quite a while to find one, and then it was only Microsoft “Internet Explorer”. But finally we managed it, only to discover that SBS On Demand isn't worth the trouble. Glacially slow, low resolution, and with forced commercials. No thanks, SBS. Still, that's the Marriott's learning experience, and they may well find something more interesting.
But once again it's a learning experience. What I have learnt today: Smart TVs seem to have missed the market. Not only I hate them; so do normal users. The interface is painful beyond measure. The same seems to be true of modern operating systems, if Microsoft “Windows” 10 is anything to go by. Microsoft has always been painful, but they seem to be outdoing themselves.
| Wednesday, 31 January 2018 | Dereel → Ballarat → Dereel | Images for 31 January 2018 |
| Top of page | ||
| previous day |
|
The daily net monitoring
|
Topic: technology | Link here |
One of the interesting things that I saw in my network logs yesterday was that the latency (more accurately, ping time to the next IP address, affectionately called radiation-tower.ausssiebb.net) had dropped. Two weeks ago I had stated that it ranged from 15 to 50 ms, but yesterday it looked like it had dropped to generally under 20 ms. OK, we can track that (blue trace):
Yes, indeed, it's looking a lot better, barely ever over 25 ms. But later it became clear that things deteriorate in the evening, with packet loss rates exceeding 20%:
What else can I add to these graphs? When I have time, I should fix the weekly graph (separate code, still the old format) to match.
|
Into Ballarat again
|
Topic: health, general | Link here |
Off with Yvonne to Ballarat and the Base hospital for a checkup (successful), during which I headed off to the Ballarat Pump Shop to look for alternatives for my pump. Spoke to Jeff, who proved much more knowledgeable than the people at UPI—not my previous assessment—and who made it clear that the pump I had was in fact the best choice for running with solar electricity. Chris Bahlo's pump is a little underdimensioned for our household, the air tank costs about $300 and only helps when drawing small amounts of water, round 10 l or less, so not showers, dishwasher or washing machine. They had the same model that I had, and recommended it, but it was $180 cheaper, and they could fit the controller from the old pump to save a further $70. OK, that convinced me.
|
A new car?
|
Topic: general, opinion | Link here |
My car has been off the road since late December. It has a cooling system leak, and maybe that will be the straw that breaks the camel's back. So on the way home, Yvonne suggested that we look at a few used car sales yards. She pointed me to one in Wiltshire Lane that I had managed to overlook in the past. Not much of interest there until the very last one, a Peugeot 307 that had been reduced from $8000 to $5000. Silly looking headlights, but I suppose that's modern. 2002 model—no newer than my current car—and over 200,000 km, which is more than my current car. Still, worth considering. Lots of nice extras, good leading bay at the back. Took it round the block and discovered that, despite temperature controlled interior, I couldn't turn the heating off, not even when set to 15°. And a few other odds and ends, including a windscreen wiper switch where the indicator should be, and which refused to turn off after I put it on interval.
Bailey (the salesman) was prepared to fix it all, but why did he let me drive it in that condition in the first place? Clearly if I buy this car, I'll need a good checkup. But is it the one for me? Years ago I decided that European cars were too expensive in maintenance. Time for some more research, I suppose.
|
SBS on demand
|
Topic: multimedia, technology, opinion | Link here |
Yesterday I had lots of pain trying to register Garry Marriott for SBS on demand, and noted the slow responses and the poor image quality. The image quality is probably SBS' fault, but what about the rest? Decided to sign up with a Real Computer.
Result: yes, much of the pain is thanks to SBS. It turned out that I already had an account, but I couldn't access it because they wouldn't accept my password, which was most certainly correct last time I used it. Send password reminder? It didn't like the email address (also once correct). OK, sign up again. Yet another of these stupid password check programs that don't tell you what they're checking for:
|
|
|||||||||
It seems that, like so many, a password like sbs stinks won't be accepted because it has neither upper-case letters nor digits. SBS st1nks works fine. But I couldn't change the “DISPLAY NAME”, at least not there. And of course there's the question why do I need an account, clearly asked enough that they provided a link to a page which doesn't answer the question. Still, finally I was done:
|
|||||||||||||||||||||||
OK, that looks better. Click as requested:
|
|
|||||||||
And that was all. I was, however, able to log in and set my profile (mainly to change that silly DISPLAY NAME). And there I saw that I can enter my address:
|
|
|||||||||
Default country USA? This is an Australian company, and I wouldn't be surprised if they tried to limit access to Australian residents!
Tried to access some programmes, but it was so slow that I gave up. What an advertisement!
|
Blue moon! Red moon! No moon!
|
Topic: general, photography, opinion | Link here |
Today was a blue moon, the second full moon this month. It was also a total lunar eclipse, due round midnight.
This is the second time in eight years that this has happened at the beginning of the year, and it implies that it will happen again in March. I commented about the accuracy of the reporting last time, with claims like:
Mr Jacob said this was an unusual year because there would be two blue moons, this month and in March, an occurrence that would not happen for another 19 years.If you've had enough of full moons, February will be without one, a one-in-30 years occurrence.
Should I stay up? What are the chances of the sky being clear, especially the way the weather has been lately? But come 22:00, as we were planning to go to bed, the sky was cloudless, and (of course) the moon shone brightly. This page gave me the details for Ballarat, and I was even able to get a prediction for Dereel, telling me that the partial eclipse would begin at 22:48, in only half an hour, and totality would be achieved by 23:51. OK, that was worth trying. Dragged out the OM-D E-M1 Mark II and the Vario-Elmar 100-400 mm f/4.0-6.3 and mounted it on a tripod.
400 mm isn't really long enough—postprocessing results suggested that 1400 mm is optimal—but I was able to get a few photos of the partial eclipse, including in with high-resolution mode (80 MP):
|
|||||||||||||||||||||||
Cropped versions weren't really sharp enough, though:
|
|||||||||||||||||||||||
Why? The moon is moving, of course. Did that get in the way?
Round 23:45, just before totality, the unthinkable occurred: the sky clouded over. It stayed that way long enough for me to go to bed. I got up at round 00:30, close to the middle of the totality. Still nothing to be seen. When I woke at 2:30, after the end, the sky was clear again. Grrr.
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
| Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |
