|
| Greg's diary |
| February 2019 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Friday, 1 February 2019 | Dereel | Images for 1 February 2019 |
| Top of page | ||
| next day | ||
| last day |
|
Mobile phones: die! die! die!
|
Topic: technology, opinion | Link here |
After my phone died on Wednesday, the obvious thing was to wake up my old Samsung GT-I9100T, which I did yesterday. It was half charged at the time, but it quickly discharged during the day. Put it on overnight charge. Today it was dead as a doornail: no voltage at the terminals of the battery. Why? Is that a result of buying cheap aftermarket batteries?
What do I do? The phone is old, and a new battery will cost me at least $12. Is it worth it? Went looking for something on eBay, with results that are worth their own article: eureka paniced.
Spent some time attending to that, and later Petra Gietz came along. I had bought a battery for her phone from the same supplier, so I asked her how the battery was coming along. Fine, but she wasn't using it any more: she had been given a new phone for Christmas.
Could I borrow it? Sure. She even brought it along in the afternoon. But the battery was discharged: only 2.7 V, when it should have been 3.7 V. OK, put it on charge. But it became apparent that this battery was also dead. After several hours there was no voltage at all.
Three dead phones in 2 days! Is there something wrong with my charger? I don't think so; the Nokia still seems to have charge, and the others died independently of the charger. At least one moral of the story is not to leave phones off charge for extended periods of time (more than a day or two until proof of the contrary).
|
Panic!
|
Topic: technology, opinion | Link here |
While looking for a backup phone, my mouse froze. Dammit, time for new batteries? I've had too many issues with batteries lately. But the keyboard was non-responsive too, and a quick check from teevee showed that eureka was dead in the water.
OK, I have been meaning to move it to the benchtop for a while now. One of the hindrances in finally upgrading it was that I had to grovel on the floor. So I took it out from there, in the process removing much of the dust that it had accumulated (another reason not to leave it on the floor):
|
|||||||||||||||||||||||
Finally placed it to the left of my monitors, where it's relatively out of the way:
|
|||||||||||||||||||||||
Reboot was the usual long-drawn-out process, not made any easier by the fact that my keyboard no longer wanted to work at boot time (why? I can't recall that from earlier occasions), so I had to connect a second keyboard. And I finally need to fix the currdev setting:
In view of the USB issues I had been having, I decided to power cycle the office UPS. While continuing with other things, found that it was making a lot of noise: the RCD in the garage had tripped, presumably when I turned it on again.
Why doesn't /home do a background fsck? Decided to disable fsck and mount it read-only, so that I could complete the fsck after coming up again. But that didn't work: it still complained about the dirty file system, and when I mounted it manually, too many things failed, including writing the DHCP lease information. All that I achieved was to delay the recovery time until I had done the obligatory two passes through fsck (only 18 minutes each, for some reason).
The system went down at 8:24, and I wasn't back up until 11:07. While I was at it, connected up the “new” /home disk, which has been waiting to take over that function for over a year now. But for some reason it wasn't detected by the hardware. Cable problems?
And what was the cause?
Yes, the time is correct; this was the first attempt, before I ran the fsck on /home.
And what was the cause? This time I had no difficulty reading the dump:
In UFS. I should be able to see something there, notably the stack frame for namei():
OK, yet another indication that there's something serious wrong with /home. This was one of the reasons why I wanted to move stuff to the new disk as quickly as possible. But that still looks like it will have to wait for another day.
In passing, it's worth noting that the dump summary gives different frame numbers from the gdb bt command. The namei() frame was 17 in the summary, but 19 in the backtrace. It's the latter that I needed.
|
Toll: Couriers at their best
|
Topic: general, technology, opinion | Link here |
The crash couldn't have come at a worse time. eureka is the main gateway for our home network, including the telephone. And I had given the number to Toll to call before picking up my phone.
OK, call them up and tell them to call Yvonne's phone (the only one that still worked). Where's the number? They're too polite to write it anywhere on the documentation they provided, so I needed to look it up on the web.
But that requires a computer! OK, we have computers, but they connect via eureka. I found a way around that that is worth another article. Finally found the number, 13 15 31, at https://www.mytoll.com/web/guest/contact-us, which requires you to know which button to click, and called that. Endless wait loop, made no easier by the fact that they kept advertising stuff in a manner that sounded as if the call was finally being answered. And when I finally got an answer, they had to connect me to a different service, with another long wait.
Finally I was connected to Ken and explained that I wanted to change a contact phone number for an existing booking. Which number would you like? Manifest ID? Sender Account (number)? ConnNote ID? I gave him all three.
But he wanted more. The new phone number. OK, that's reasonable—once. But he asked for it no less than four times. My name? Only twice. The address? Three times. Dammit, this information is all in the manifest. Sorry, I need this information to access the manifest. Finally, as I could barely take this any more, he asked me “Is this pickup already booked?”.
I've had some stupid “consultants” in my time, but I think this beats everything. Asked to speak to his supervisor. Sorry, he's on the phone (repeatedly). Asked him to leave a message to call me back. OK, will do. And of course he didn't. All I have with any certainty is the number 008992552 and a charge of $2.76 for a 23 minutes mobile phone call, about what it would cost to call Germany for 2½ hours.
Still, the system came back up almost as soon as I had got rid of the idiot, and the phone with it. So whether or not he managed to update the information, I should have received a phone call, or at least voice mail if the connection failed. I didn't. And they didn't show up.
„Toll“ is a German slang word meaning “great”. Somehow it doesn't apply to my experience today, on multiple levels.
|
Bending the network configuration
|
Topic: technology, opinion | Link here |
So I had to access the web while recovering eureka. I can do that, I suppose. Just connect dischord directly to the NTD, which I have decided to mount upside-down so that I can get at the connections:
|
|||||||||||||||||||||||
That ultimately worked. I had to change the network configuration, of course, to use DHCP. And while doing that I got one of the most amazing messages from the system:
|
|||||||||||||||||||||||
Unknown publisher? That's part of the base system! I'll never understand Microsoft, just like I will never understand why there's a per-interface DNS setting. And of course I had to find the network settings page for firefox, now called “Options” and hidden on one of the meaningless icons at top right. But I made it.
Moving back was more difficult. I set the correct address for dischord.lemis.com (192.109.197.169), but it didn't want to know, and gave me a DHCP address anyway. I had to reboot the bloody thing before it accepted what I had told it.
|
Why no background fsck?
|
Topic: technology | Link here |
I have /home on eureka set for background fsck, but it never does it. Why not? Spent some time looking for criteria, which was surprisingly hard. Found this in the FreeBSD handbook:
OK, I fulfil those requirements, maybe. What does "not have been marked as needing a foreground check" mean? Took a bit of a look into the code, but maybe the problem will go away when I replace the /home disk.
|
More garden stuff
|
Topic: gardening | Link here |
Mick the gardener along this morning. As a result of my computer problems, I didn't have much time to discuss things with him, so I asked him to do some weeding.
He did that very effectively; an hour later I came out and saw the difference in the “D” bed (the shape). Here before and after:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, he had removed almost everything, including the yellow-flowering ground cover I had been cultivating. Agreed, there were a lot of weeds in between, and now, as he said, the bulbs that I had planted underneath could finally grow:
|
|||||||||||||||||||||||
I didn't plant no steenking bulbs! Those look like a different kind of weed to me, but I'll leave them to be sure.
He also left a couple of other plants: an Acacia melanoxylon (I think), and also a Betula pendula:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I suppose we should keep the birch and transplant it in the autumn.
Mike didn't finish his work. He complained of severe chest pain and knocked off early. Sent him off with an admonition to go straight to the emergency room of the Ballarat Base Hospital. Hopefully we'll see him again in 2 weeks.
| Saturday, 2 February 2019 | Dereel | Images for 2 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Mobile phone: Die!
|
Topic: technology, opinion | Link here |
So after the failure of no less than three mobile phones in the last few days, I'm left without a phone. I had been planning to buy a second-hand one as a backup, but then I saw a special offer from ALDI today: a refurbished iPhone 6S for $469. Yvonne was going there anyway, so she could pick one up.
I've had fun with iPhones a couple of years ago, and I quickly decided that it wasn't for me. In addition, that's more than I wanted to pay. But maybe I should give it another chance, and anyway, if I don't like it, I can return it within 2 months for a full refund. That will give me plenty of time to bridge the repair of my real mobile phone. And after all, Apple is closer to BSD than Android, which is based on Linux.
It turned out that the special was on Wednesday, but they still had one in stock. I'm used to modern devices coming with little documentation, but this one is clearly very modern: apart from a warranty card, there was no documentation whatsoever. There wasn't even a pin for ejecting the SIM card carrier, and if I hadn't been here before, I wouldn't have known what to do.
OK, fire it up.
"Hello. Press HOME to open".
Where's HOME? No indication. No documentation. Dead in the water. Well, I suspect that many people would be. That must be the button at the bottom. But why no documentation?
It came up with a time display (no seconds, of course, we're modern) that made no sense to me: at 14:00, it showed 7:00, 7 hours behind us. That would be UTC+4. Why that? The minutes were accurate. It took me a while to discover that it was a 12 hour clock, and the system was too polite to show am or pm. It was really set to 19:00 yesterday, the time in Cupertino. Not for the first time, I'm baffled that these things are so inaccurate. And why California? Is that where they were refurbished?
OK, how do I set it up? In times gone by we
would RTFM, but there was no FM. I could
have downloaded one, but I'm clever, right? I should be able to work figure this
out for myself.
First I was given a “Quick Start”:
|
|||||||||||||||||||||||
“If you want, you can also set up this iPhone manually”. Well, it's not that I want; I don't get the choice unless this is a second device. OK, go through the usual pain of entering data on a minuscule touch screen. Next it wanted personal data:
|
|||||||||||||||||||||||
Not just personal data, but they decide what services I can get. Damn you!
Next they wanted an email address. OK, give them the old one:
|
|||||||||||||||||||||||
I thought I deleted that old ID! OK, go through and get them to reset the password, which I hadn't kept:
|
|||||||||||||||||||||||
That seemed to work—I even guessed the 6 digit PIN correctly (no, it's nothing trivial or that would occur to anybody else). Done! Finish setting up, refuse to give them my credit card details (“defer until later”), and on to insert the SIM card. That worked, but not without surprises:
|
|||||||||||||||||||||||
Why Telstra? My provider is ALDIMobile. Yes, they use the Telstra network, but this is the first time that I have seen a phone show the network name rather than the provider.
During setup, by accident I enabled Siri, a voice recognition system. Yes, compared to other systems, it did recognize my voice, but didn't give me anything useful. And I couldn't find a way to access the menus any more. It seems that the duration and number of clicks on the Home button might have helped, but I didn't know that. Instead I asked Siri to go away. “Why would you want to do that?”. I couldn't find a way to get Siri to disable itself. It took me something like 10 minutes to get rid of the bloody thing.
OK, finally. Let's configure the thing to make it marginally less painful. First set the time to something sane. At some point the phone discovered that it was in Australia, and gave me a time 12 hours ago, again without am/pm indication. How do you configure the display? As this thread states, you don't. No seconds: they're out of date, and you can have them on a retro-style analogue display, but not on the digital display. About the only thing you can do is set a 24 hour clock. Given the restrictions they have set up everywhere, I suppose that's something.
How about some apps? As I said on IRC:
I have three things I'd like a phone to do, apart from making phone calls:
The general opinion was “no can do”:
Enjoying your walled garden?
You are mistaking an iphone for a general purpose computer.
And, indeed, why not? The machine has processor power and storage that would run rings round any of the machines on which Unix, BSD and the Internet were developed. Why is it trying to restrict me?
Looking at the three points: I deferred the first as a vain hope for something that was standard 30 years ago, but has now gone the way of the dodo. What about exporting the file system hierarchy, something like the WiFi File Transfer that I use on Android. No, the consensus was that I wouldn't find that either. In fact, I don't think I've ever seen what an iOS directory hierarchy looks like.
OK, let's assume that my approach is wrong. How do you get a file from an iPhone to a computer? The first suggestion I found on the web was
Scream! More searching brought up this page, which suggests:
Talk about interoperability (an old, worn-out magic word, it seems)! Four of these suggestions involve an intermediate server, probably at the other end of the world. Only the first looks like it might be local. I know Easeus, and I use their backup utility on dischord after Microsoft's own backup failed without obvious cause. But why do I need a specific operating system to access files? From a technological standpoint we're back in the 1960s.
OK, next thing. I want a GPS logger, something like the Mendhak GPS logger that I use on Android. Searching the AppStore showed a couple of apps that could potentially do the job, though of course they were too polite to give any details of what the thing does, and possibly it's just intended to keep track of your phone. Tried to download it:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Not OK at all. Why is it disabled? Why wasn't I told earlier? What do I do to fix it?
This might be related to a similar issue I had two years ago, where I suspect that they required my credit card number. OK, finish that setup. What do I find?
|
|||||||||||||||||||||||
That's my information from 2 years ago. Surely I had deleted it? OK, update the information:
|
|||||||||||||||||||||||
Now it tells me? But not why. And what's a code, and how do you redeem it?
OK, Apple, that's enough. Not only do I hate you, the experience made me feel physically sick. They remind me of confidence tricksters trying to get my money. They appear to be particularly successful. I'll use this thing as a phone only and return it when my phone has been repaired.
One last thing was to set the display background, again by jumping through multiple hoops:
|
|||||||||||||||||||||||
Once again, though, I'm baffled. I know lots of computer savvy people, many of whom have been active in the field long before these “smart phones” came along. Why do they put up with them? Why am I the odd man out?
|
Chicken cooking times revisited
|
Topic: food and drink, opinion | Link here |
Yvonne prepared a whole chicken for dinner today, and it was up to me to cook it:
|
|||||||||||||||||||||||
In the past I worked on the basis of 170° in the oven and 50 minutes per kg to get the breast meat to 82°; later I tried 180° and 80°, which might have been on the cool side. Today I tried it again with a 2 kg chicken and the temperature sensor a little further inside, and it was at 82° in about 80 minutes! And that's still too hot for the breast. 80° it is from now on.
|
|||||||||||||||||||||||
| Sunday, 3 February 2019 | Dereel | Images for 3 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Getting rid of Apple
|
Topic: technology | Link here |
So, I've decided: life is too short for iPhones. Let's get rid of this horrible Apple ID. How do I do that?
With extreme difficulty, it seems. Went searching and found this page. While I was doing this, I got a message on the phone:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Where's Darlinghurst? Almost exactly 1000 km from here. Why is somebody trying to log in to my account from there?
Why, is somebody trying to log in to my account from there? I decided that it was Apple's attempt at locating me. After all, the thing has a GPS receiver in it (I think; I haven't been able to find a way to access it), and I had been outside where it could have located me precisely. But no, Darlinghurst is good enough; maybe it has something to do with the net topology. Certainly “Weather”, which I had explicitly allowed to access my location, also thinks I'm in Sydney. Still, this inaccuracy seems to be modern.
Started on the Apple ID deletion process before I realized that it contained no less than 54 steps! And even then it could take up to a week before the data finally went away.
The first step was to remove my devices, which was fairly straightforward. But one of the pages mentioned, https://appleid.apple.com/account/manage, also allowed me to upgrade my credit card details. Since this was with a real computer it wasn't nearly as painful, so I tried it. No, same thing. Why don't they have a clear error recovery documentation?
Tried to continue, but the more I read, the more my eyes went funny. Surely there must be an easier way. Found this page from the relatively authoritative sounding https://www.macworld.co.uk, where I read the reassuring information:
How to get Apple to delete your account
The answer seems to be that you currently can't get Apple to delete your accounts (please let us know in the comments if you know otherwise).
What you can do, however, is edit your personal information to make it a little less personal. Once you've removed your Apple ID from all your devices you can change your name, address and other personal details.
While I was doing this, discovered this new article: “Did Apple really ban Facebook and Googles Apps?”. Yes, it seems. They had reasons that were marginally valid, but as the article says:
This event is a reminder that Apple has control over its mobile operating system and the code that can run on it. Apple not only curates the apps allowed in the App Store but can remove and revoke access to those apps when necessary. Apple does this when malware is discovered in an app that slipped through, for example.
My feeling of physical revulsion is gone, but the more I learn, the more I understand that it's not for me.
Later I wanted to check something, and was asked to enter my Apple ID password. Wait, I just deleted the device? But I entered the password anyway (Revuls1on), and it told me that it was sending a code to my phone (two-factor authentication).
But it didn't. After a minute, OK, please resend. Then the phone rang, and I entered the code. Wrong, fool. Tried again. Nope. Hung up and the phone rang again with a new code, this time the correct one. A classical race condition made easier by the very slow response time of their authentication methods.
OK, that's it. Hopefully my phone will be repaired soon. And if I ever need another phone, it will be Android. Not that I like Android, but I can live with it. Life's too short for marketeers telling me how to live my life.
|
Tomatoes!
|
Topic: food and drink, gardening | Link here |
The tomatoes in the garden are gradually producing more fruit:
|
|||||||||||||||||||||||
The funny thing is that this variety of fresh tomatoes looks so different. Almost all of them are from a single plant. And today we ate the twin tomato:
|
|||||||||||||||||||||||
| Monday, 4 February 2019 | Dereel | Images for 4 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Returning the Nokia 3, try 2
|
Topic: general, technology | Link here |
Friday's attempts at returning the Nokia 3 phone were both extremely frustrating and completely ineffective. But Saturday's experience with the iPhone 6S makes it clear that I need it back as soon as possible. OK, brave Toll (a company whose obvious URL http://www.toll.com/ belongs to a domain squatter) on 13 15 31 and try again.
This time things were different. The phone was answered almost immediately, by unintelligible, I think with a Pilipino accent. She asked me for the usual stuff: ConnNote ID, account number and my name. Then she put put me through.to another unintelligible, whose name on repeated request sounded like “Michaela”. She wanted the same information all over again, and was so hard to understand that after repeating an unintelligible question three times, just gave up. She kept referring to a partial pickup. No, I wanted a complete pickup, only one item. After a few minutes of increasing frustration, I decided (because she was too polite to tell me) that she meant “parcel”, not “partial”. And she wanted a booking reference number. No, not the ConnNote ID. Not the account number. Not the tracking number. Not the shipment number.
Gave up and called QSL. They only seem to have the one phone person, Neha, but she was efficient and told me (this round 12:00) that the item would be picked up between 13:00 and 17:00 this afternoon. Who would believe that?
But how about that: at 14:00 a bloke with a northern English accent and an unmarked van with South Australian registration came and picked it up, after we had locked the dogs away: he had recently been bitten badly by a dog. Only later did I realize that I had absolutely no confirmation that he had anything to do with Toll.
But he did, and the parcel was really delivered to Bankstown, NSW at 9:11 on the following day. Not surprising in itself, but given my experience to date it's somehow out of keeping.
|
New ThinkCentre
|
Topic: technology, opinion | Link here |
The ThinkCentre that I had ordered last week arrived in Napoleons this morning. That's a replacement for Yvonne's defective one, and specially ordered to be the same as dischord.
I needed my old test box back to recover eureka's /home file system, so off to pick it up.
It's completely different! A different way of opening it, a different layout inside, and a different disk carrier:
|
|||||||||||||||||||||||
According to the label, it's a MT - M 7033, but it was sold as an M91p. dischord is marked at MT - M 3132. Like dischord it has an Intel i5-2400 processor—that's why I bought it—but the layout is not as “intuitive” as I thought. In particular, getting the disk out of the carrier proved to be quite a problem, nothing like as easy as with the old carriers, as this page shows.
In passing, the model numbers are somewhat evocative. 7033 sounds like IBM 7030, and 3132 reminds me of the IBM 3032 that I was using 40 years ago.
OK, it comes with Microsoft “Windows” 10, which could be useful. Does it work? Yes, once you let the thing fire up. Both up and down seem to be much slower than with older versions of Microsoft.
Next, check with the disk from tiwi, the FreeBSD system in my test box. No problems at all. It even loaded the correct kernel modules for the Radeon display card (the first I have had) and configured X correctly!
OK, later to replace lagoon. And lagoon, with effectively the same software, didn't detect the card correctly. X came up in archaic 1280x1024 mode. A bit of cursing, put the nVidia card in instead, and it was up and running.
Next back to tiwi. The Radeon card doesn't fit physically because it has a low profile bracket, so I replaced the old, noisy nVidia card: this motherboard is so old that it doesn't have on-board graphics. Then removed the new /home disk from eureka, put it in and booted:
How did that happen? Was that also the reason that eureka couldn't find it? No, no such message there, and the device itself wasn't located. Still, no need to worry about that, just repartition.
And then the eternal disk copy, limited not by the network but the disks themselves:
I estimate that it'll be finished tomorrow evening. Once again I'm left comparing with the IBM 3330 disks I was using 40 years ago. The two walls together had less disk space than tiwi has RAM, and the first 3200 MB were copied in about a minute.
|
Mulch!
|
Topic: gardening | Link here |
On the way to Napoleons I ran into a tree trimming operation in Grassy Gully Road. Just what I need: mulch! Arranged for them to bring their harvest along when they were done, for the somewhat high price of $50:
|
|||||||||||||||||||||||
That'll keep Mick busy for a while.
|
Migraine again
|
Topic: health | Link here |
While watching TV this evening, had difficulties with my vision. Damn! I've been worried about my eyes for some time. But after a few minutes it became clear that it was just a mild migraine, and it passed in about an hour. That's the first time I can recall having a migraine in the last 15 years.
| Tuesday, 5 February 2019 | Dereel | Images for 5 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
RCD trip
|
Topic: general | Link here |
Woke up in the middle of the night and turned the bedside light on. Nothing. Damn! Another power failure? No, the air conditioner was still on. UPS? Looked outside. No, computers still running.
Must be the globe. It's strange how what was once the most obvious reason is so far down the list nowadays. But when I got up, I discovered that the RCD had tripped, taking all the fridges with it.
What causes that? I'm pretty sure that there's nothing wrong with any of the devices. Could it have been the result of another power surge? High time to get solar electricity installed.
|
Photos of paintings
|
Topic: photography, animals, opinion | Link here |
Yvonne frequently asks me to take photos of her paintings. It's like pulling teeth: the surface of the canvas is (deliberately) uneven, and the paint is partially reflective, so it's almost impossible to get a photo without reflections.
|
|
|||||||||
Today I thought I had it, but I managed to get flare from the flash units in the photos, this time in the right eye of the horse:
|
|||||||||||||||||||||||
What next? I should try with umbrellas, but my experience is that that didn't help much either.
| Wednesday, 6 February 2019 | Dereel | Images for 6 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Meat proportions
|
Topic: food and drink, opinion | Link here |
Tried a “MARION'S KITCHEN” Phat Thai mix for breakfast today. I don't really understand Phat Thai well, and I'm hoping that various mixes might help me get a feeling for it.
This one was surprising: for 4 portions, only 200 g of (dried rice) noodles, supplied, but they wanted 400 g of meat as well! That sounds completely out of proportion. Last time I had 130 g of rice sticks an 80 g of meat, and I didn't notice any problems with those proportions.
In the end I added 100 g of (different) rice noodles and put in only 300 g meat. In retrospect, I think that's wrong: the sauce is laid out for 200 g of noodles, and the extra noodles diluted the sauce more than it could take. But the proportion 1:1 for (dried) noodles and meat seems right, though I could maybe use less meat and put in another egg.
In that connection, I'm reminded of this graphic from Statista:
It's interesting to note the relative proportions. Is this indicative of demographic changes (notably the increase of meat consumption in China)? Certainly the fact that goat has overtaken beef (“cattle”) doesn't come from the Western world.
|
Moving forward on solar electricity
|
Topic: general, opinion | Link here |
I've been thinking about solar electricity for years now, and my second to main reason for delay was because I was waiting for the government to finally come up with some good financial support for the idea. But clearly the government (not the one I voted for) doesn't see any reason to provide any support, although electricity prices are one of the key issues with the average voter. Solution: new coal-fired electricity plants.
Still, I've been trying to contact Simon Reid of BREAZE for nearly a year now, both by phone and email. Clearly they're not interested, though they do still seem to be active. Time to look for alternatives, of which, of course, there are many. Ended up filling out a form on two sites, one of which promised me three quotes, and also calling up another, speaking to Josh of Re-Energy, who will come along on Friday. At least we're getting somewhere.
|
Garden in times of drought
|
Topic: gardening | Link here |
It seems that last month was the hottest January on record in Australia. But we didn't see much of that here; I think it was hotter 60 years ago, when we called 40° (OK, 104° F) a “cool change”.
What has been clear, though, is the lack of rain. Last January we had 33.9 mm of rain in January. This year it was 0.6 mm. And the garden is showing it:
|
|||||||||||||||||||||||
I still have hopes that this plant (a Hebe) will survive. There are some new leaves:
|
|||||||||||||||||||||||
But clearly it needs urgent attention. The problem proved to be yet another blocked dripper, and it wasn't helped by yet more clogged water filters. I only cleaned the filter 6 days ago, but it was already clogged again. Presumably the ground water is also receding, and that will probably take a long time to fill up again.
And the rainwater tanks are ⅔ empty. Or so I thought. On checking, the tank I was using was ⅔ empty, but I had turned the tap off the other tank when the water pump failed before Christmas. Given the heavy rain the following week, it was completely full, so we still have ⅔ full rather than ⅔ empty. And that should be sufficient to carry us over this dry spell.
As if to confirm that hope, we did have rain today. Only 2.2 mm, but that's still more than 3 times the whole of January. And we don't need to look far to see where the rain went: this month (6 days), Townsville has had 830 odd mm of rain, more than we get in a year.
|
The joys of “smart” homes
|
Topic: technology, general, opinion | Link here |
There's a lot of talk about “smart homes” recently. In principle it's something that should be close to my heart. But everything I read about them screams “NO!”.
In principle, I have something like that already, though not the way the marketeers conceived: my garden sprinkler system is run by a network connected controller, using a program started by a cron job. It works well.
But that's not what people think of when you say “smart home”. What they think of is voice controlled systems in the house (what, you have to tell it what to do?) and controllers accessible by mobile phone app when you're not there, closely related to the stupidity of an Internet connected fridge.
Sure, there's room for that kind of thing as well, for people who are happy with it. But where's the interconnection? How do I control the whole thing from anywhere I want? How do I automate it? The whole thing is supposed to be about automation, but what I see requires people to do things on each occasion. That's not automation, though it may pass for it in the Microsoft space.
And then I read this article, which sums things up: the whole operation of a number of devices requires an Internet connection to the vendor's servers! To quote: “Companies Can Shut Down and Render Their Products Useless”.
Sorry, people, that's not my kind of technology.
| Thursday, 7 February 2019 | Dereel | Images for 7 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Can't access photo
|
Topic: technology, opinion | Link here |
Mail from Daniel Nebdal this morning, telling me that he was getting an HTTP 403 error accessing some of yesterday's photos, like this one:
|
|||||||||||||||||||||||
What's unusual about that one? Daniel suspected the name, https://lemis.nyc3.digitaloceanspaces.com/grog/Photos/20190206/small/Skjómi-1.jpeg. There's a ó in there, something that US Americans don't use. Could it be that? There's an easy way to find out:
=== grog@eureka (/dev/pts/6) ~/Photos/20190205 223 -> ln Skjómi-1.jpeg Skjomi-1.jpeg
=== grog@eureka (/dev/pts/6) ~/Photos/20190205 224 -> make sync
In passing, it's nice to see how this software recognizes duplicates and does the right thing. So now we have two different names for the same image, one with Skjómi, the other with Skjomi. And how about that, Skjomi works, and Skjómi doesn't.
Why? Clearly a bug in the image representation. But why does it return 403 (“forbidden”)? That seems to be its standard response for a file that doesn't exist:
=== grog@eureka (/dev/pts/6) /var/tmp 244 -> fetch https://lemis.nyc3.digitaloceanspaces.com/grog/Photos/20190206/small/Skjomii-1.jpeg
Maybe that's because it tries to access the directory, which I have prohibited.
|
Microsoft network problems
|
Topic: technology, photography, opinion | Link here |
Ashampoo Photo Optimizer is one of the least reliable programs I know, but they're working on it—to make it less reliable, it seems. Today I had some issues with it, so I tried the standard Microsoft method of restarting it. But no, it told me that my (saved) license key was invalid. Please enter the correct key (displayed).
But that didn't work because it hung, not for the first time. OK, next step: reboot the machine (dischord). And when it came up it told me that it couldn't connect to the Internet. In fact, it couldn't even resolve non-local names.
What's wrong there? In to look at this completely non-intuitive network menu system. It had recently added its fifth “network” without my involvement. Could it be that I had some strange profile quirk?
Much messing around and cursing. Finally I discovered that it had lost its (per-interface) default gateway address. Put that back again, and suddenly DNS lookups (local) worked again, and I could access the rest of the Internet.
So why did Ashampoo fail? Presumably “incorrect license key” is its way of saying “Can't contact key verification server”. And “can't resolve name” is Microsoft's way of saying “can't find a path to the resolved name”. How I love accurate error messages.
|
Skjómi painting, yet again
|
Topic: photography, opinion | Link here |
I did some more playing around with Yvonne's painting of Skjómi yesterday and today. I'm still not satisfied, but I've given up. Here's the final version, taken with flash and umbrellas. The reflections are clearly visible, but the lower microcontrast makes them less painful:
|
|||||||||||||||||||||||
|
More solar power insights
|
Topic: Stones Road house, technology, opinion | Link here |
Call from Thomas Kucera of Effective Electrical today, about the solar power installation. He had a completely different approach: first look at my daily power usage for the past year and then make a suggestion based on that.
How does he do that? Download the information from Powercor. All you need is a meter number, and you can get the usage data, apparently for the last 2 years, itemized in 30 minute increments, and in CSV format:
He has programs that import it into Microsoft “Excel”, and he doesn't have any further information. But it's clear that the first item on each line is a record type, and 300 is the information he's looking for. But it's also clear that there's more information there than what he's looking for. I should go looking for the data format.
| Friday, 8 February 2019 | Dereel | Images for 8 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More Hibiscus flowers
|
Topic: photography, gardening, opinion | Link here |
Another Hibiscus rosa-sinensis flower opened today, another opportunity for a time-lapse series. One of the problems with them is that they start opening with little warning early in the morning. By the time I got to see it happening, it had already been opening for a while.
The other issue is guessing how big it would be. This time I made sure to leave enough space around the flower so that I would still have it in frame when it was open:
|
|||||||||||||||||||||||
It didn't help. The flower was not just bigger, it also sagged:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Olympus has some new software that promises time-lapse functionality. I must find time to look at it.
|
More solar power
|
Topic: Stones Road house, opinion | Link here |
Call from Luke McKenzie of the strangely named Cola Solar this morning, giving me some suggestions on prices round the $20,000, and promising a formal quote later in the afternoon. It seems that there are a number of discounts available: $2,462 from the Federal Government, already factored into the price, and about the same from the State government, which I'll have to chase myself.
Later Josh from Re-Energy came along as planned. It seems that we hadn't prepared as well as we could have: he is only involved in “conventional” installations, not including “hybrid” (systems including both grid and battery installations). For that, Sanjay, his boss, is responsible. He'll call me next week.
Still, it wasn't a complete waste of time. He left me with a number of things to think about, including why he wanted to put panels on the east side of the house (actually pointing about 99°) rather than on the more obvious north (9°) side. There's also the question of which panels: they offer panels from Trinasolar (305 W, $230-240, easier to understand as $0.78 per W) with a 10 year warranty, or from LG (325 W, $280, $0.86/W) with 25 year warranty. For the 5.7 kW that he sucked out of the air, that would be a difference of $464 ($4,446 against $4,910). Is the warranty worth it? The price difference amounts to two panels. How many would be likely to fail? He wasn't able to give me useful information about the reliability of the things. Something more to investigate. But he was able to offer installation within a month, which seems completely acceptable.
Also the formal, if somewhat vague, quote from Cola Solar: $20,470, with a rebate of $2,225 from the State government, for a total of $18,245. He has put 20 panels on the roof: 15 on the north side, 5 on the west. They're Risen panels, 285 W, for—coincidentally—5.7 kW total. Also a Tesla Powerwall with a storage capacity of 13.5 kWh. Warranties are interesting: the Powerwall only has a 10 year warranty, and the panels have a 12 year product warranty and a 25 year performance warranty, whatever that implies. But how do they connect to our household electrical circuits? There are a whole lot of questions circulating in my head that I need to put down on paper, including priorities (charge batteries first, use excess energy for heavy consumers; how much feedback to the grid?). I need to write things down.
The other issue is the time frame: June. That's a lot longer than from Re-energy, but it seems that the sticking point is the Powerwall, which Josh didn't consider. Luke says that he might be able to find another one more quickly; clearly this isn't the mass business that the standard installations are.
| Saturday, 9 February 2019 | Dereel | Images for 9 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
System upgrade, try 2
|
Topic: technology, opinion | Link here |
Tuesday's repartitioning of the new /home disk for eureka proved to have been a little hasty. I took the parameters from an older setup, giving myself 20 GB for the root file system (including /usr, and in particular /usr/local) and 10 GB for swap. In addition, I used effectively default settings for newfs on the new /home.
The result was that the /home file system took up considerably more space than the original, several gigabytes more. Maybe part of the issue was the ridiculous number of inodes. Apart from that, the root file system was smaller than the current contents of eureka's root file system (30 GB), and I had already noticed that, despite have 32 GB of memory, eureka frequently comes close to 10 GB of swap usage.
OK, time to update my partition HOWTO page, in particular to include more comments (like “don't try to newfs a disk device”). Repartitioned the disk and tried again, in the process copying the root file system from eureso (my next generation eureka) from VM to the root file system.
Booting the new system was surprisingly simple, and so I started bringing the system up to date and copying /home again. But still no go. After a while I had:
=== root@eureso (/dev/pts/1) /eureka/home 6 -> df -i /mnt
=== grog@eureka (/dev/pts/14) /var/tmp 85 -> df -i /home
I had managed to create a file system with only a little over 2,100,000 inodes, and the current /home file system already uses nearly 5 times that amount. That's what comes of copying newfs parameters without thinking (in this case, from teevee's /spool file system, designed for large video files). Back again and recreated the file system, this time, for the fun of it, enabling soft updates journaling. And off again.
Somehow the speed didn't seem to be as high, and round 18:00 it seemed to have ground to a halt. iostat showed that it was performing about 90 small I/Os a second, but the df output showed no change in size.
Journaling? To be fair, copying files is about the worst case scenario for journaling. OK, it's a switch, so I can turn it off until the disk has been copied. Simply umount the file system, tuna the fish and continue.
Simply? Well, no, umount took longer than expected. Normally it's done in a second or so after flushing buffers. In fact, umount took much longer than expected. After several minutes I checked. top showed:
Those are the 90 I/Os a second continuing.
OK, leave it go. umount took very much longer than expected. By late evening it was still going (if that's the word). Left it over night.
|
Power usage data formats
|
Topic: technology, Stones Road house | Link here |
Mail from Daniel Nebdal today:
That'll be something to follow up another day.
|
Jane Gregor visits
|
Topic: food and drink | Link here |
Saturday evening meal almost as usual today—Wiener Schnitzel—except that we also had a visitor, Jane Gregor, who is talking to Chris about horses.
| Sunday, 10 February 2019 | Dereel | Images for 10 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
umount completed
|
Topic: technology | Link here |
In to the office this morning to see how my umount was getting on. Still not done, after 16 hours. OK, there's no reason to believe it will, and in any case, if the file system is corrupted, it would be quicker to start again than to wait any longer. So Big Red Button time.
The system came back with few issues:
|
|||||||||||||||||||||||
The first two file systems recovered instantaneously. The third (/dev/ada0p5, the new /home file system) took a little over a minute. And when it was done, a large number of files were missing. They were all tiny files, so there was a lot of metadata, and if I understand this line correctly, the journal was just about full:
|
|||||||||||||||||||||||
Still, that's about what you'd expect under these circumstances. OK, turn journaling off:
=== root@eureso (/dev/pts/3) ~ 2 -> tunefs -j disable /dev/ada0p5
I hadn't really expected it to work with the file system mounted, but those messages are somewhat contradictory. Did it turn off journaling or didn't it? It didn't. I had to umount the file system, after which I got only the messages I wanted
=== root@eureso (/dev/pts/3) ~ 3 -> umount /mnt
=== root@eureso (/dev/pts/3) ~ 4 -> tunefs -j disable /dev/ada0p5
=== root@eureso (/dev/pts/3) ~ 5 -> mount /mnt
=== root@eureso (/dev/pts/3) ~ 6 -> l /mnt/.sujournal
=== root@eureso (/dev/pts/3) ~ 7 ->
The message “remove .sujournal” was an imperative, not an indicative. For the 32 MB it takes up, I decided to leave it there. Spent the rest of the day copying files, noting the amazing number of small files, typically svn “pristine” files, of which I clearly have far too many.
Apart from that, watched Kirk McKusick's presentation on soft updates journaling, which was quite informative. At the time (2010) they were using a 16 MB journal, and Kirk claimed that 1 MB would normally be enough. The fact that the default is now 32 MB suggests that they've changed their mind on that one, and maybe there still be daemons in that area.
|
Interruptible power supply
|
Topic: technology, general, opinion | Link here |
While looking at my soft updates journaling issues, looked at my uptime information. lagoon and teevee had both rebooted about half an hour before. Why? It was round 7:30, when we were not yet up, and nothing else had failed.
My best bet is my main UPS. It is very flaky when the bore pump is running (currently between 4:00 and 10:00), and I continually hear the second UPS (in series) beep and the lights flicker. My guess is that it disconnected a little longer for once and took two computers out.
That's ridiculous. This http://www.eaton.com.au UPS is really a very bad choice. I was told before I bought it that it would run off a generator; it doesn't. And it complains about other power as well, such as the bore pump. The result is that, far from avoiding power interruptions, it causes them! And that a well-known brand and by far the most expensive UPS I have ever bought.
Not much I can do about it now except to replace it with solar power. Hopefully the UPS will carry on to run without another failure like this one.
|
How much power from a solar panel?
|
Topic: Stones Road house, opinion | Link here |
Discussion with Peter Jeremy on solar panels. They have a power rating, of course, but what's it based on? Insolation, of course, but how much? Looking at my brochures, it seems that they can be rated at either 1 kW/m² or 800 W/m². How much do we get here on a sunny day? Five years ago I noted 1.3 kW/m², but not where I got the figure from. I'm sure I had some comparative figures even earlier on in a rebuttal to a Wendy McClelland claim. Still, 1.3 kW/m² sounds reasonable.
More to the point, though: how much sun do we get in a day? This page includes this overview:
|
|
|||||||||
|
|
|||||||||
That needs rethinking to be useful, but assuming the base data are good, it's a reasonable start. The suggestion is that there are 3.65 hours of “sunshine” (in other words 3.65 kW/m²) per day in the yearly average, with a ratio of (coincidentally) 3.64:1 between summer and winter.
| Monday, 11 February 2019 | Dereel → Gisborne → Dereel | Images for 11 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Navigation in the time of GPS
|
Topic: technology, general | Link here |
I've been using map-based GPS navigation for nearly 9 years. It's been an interesting experience, on the whole positive. But that's only a fraction of my navigational experience. Over 50 years ago my father and I navigated from Madras (now Chennai) to Tavistock, something that Google Maps still can't do. Clearly we didn't use GPS; we had a mixture of second-rate maps and aviation charts given to us by a pilot whom we knew. And somehow we didn't get lost very often; I can only recall one problem, at the very beginning,
GPS navigation is clearly much easier. The trouble is, what if we want to go somewhere that doesn't fit into the expected scheme of things? Today Yvonne wanted to go to Gisborne to try out a riding vest. We didn't want to drive 250 km just for that vest. Let's take the dogs, visit a national park and get some photos. And that's a situation that shows the down side. Have an address? Type it in to the system and it'll take there, possibly with some strangenesses: it took us on to the freeway for two segments, then off the freeway again for one segment, then back on again. Why? Presumably the simplistic “shortest” setting. But that's necessary because the “fastest” profile is just plain incorrect. And I can't display the route because the granularity is too low, something that I need to investigate.
First problem: which national park? That's relatively easy to find. I settled on the non-intuitively spelt Lerderderg State Park, almost directly on the way. Next, what is there to see there? Vehicle based sightseeing looks like a good choice. But no, that's just bus tours. The same goes for Car / Limousine tours. In the end, I took a look at the Google Map, which shows this:
|
|||||||||||||||||||||||
Clearly not its target demographics. OK, what about the parks web site? There are a total of 7 PDF files there, including one map:
|
|||||||||||||||||||||||
Not overly informative, in particular not easy to correlate with the surroundings, but at least I have some information. OK, let's print it out.
|
|||||||||||||||||||||||
What's that? It's the same map! Why does it look so different? It's not monochrome: there are colours here too, just different ones. I've never seen that in a PDF document. Still, it seemed to be the best I could find, so I decided on two places: the entrance near Bullengarook at the North-West (“Walker's Entrance” is a description, not a name), and the entrance at the south.
Off on our way and arrived at the first place as planned. Not much to see:
|
||||||||||||||||||||||||
Walked around for a while, but it wasn't very interesting, and we have similar terrain just round the corner in Enfield State Park. So we went on to Gisborne and took a look at the vest:
|
|||||||||||||||||||||||
That was interesting for another reason: it's in Gisborne, but for some reason it's in an area only accessible from the other side. And to my puzzlement, the road ran north/south, not east/west. Why? Looking at the map, it's clear that my first estimate was correct. But there's no way to tell that with the help of the GPS navigator: it doesn't show compass directions, and the angles were so unusual they confused my innate sense of direction. But compass directions are an old, worn-out approach to navigation, it seems.
And then there's an accuracy. The Exif data show a location photo of 4.6 km from the real location. Why? I haven't had these issues before, and the route looks correct. Is there some issue with the software that converted the details? Looking at the map, there's clearly something wrong:
|
|||||||||||||||||||||||
That lower straight line is just plain wrong, as the maps show. I returned the way I came. Strangely, it's very close to the direct line that I made later:
|
|||||||||||||||||||||||
Is this something to do with the fact that I merged several track logs? It would make sense if the first track log, which finished at the extreme east of this image, was followed by the second track log that didn't start tracking until at the extreme west in this image. But why? It should have started within a few hundred metres of the return journey. Still more to examine.
On back to Lerderderg forest, where we found an unwelcome sign:
|
|||||||||||||||||||||||
|
||||||||||||||||||||||||
No dogs allowed (on the post to the right)! Now why couldn't they have make that clearer on the web site? 10 km for nothing!
But that was still not everything. When I started writing this up, I discovered that the only page marked as “map” was in fact one part of the real state park. The other was to the west, though it's difficult to find that without further help:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The first part was in the a document titled ”Lerderderg State Park Visitor Guide”. That wouldn't be so bad if they didn't also have a document called just “Pyrete Range Map”. The whole thing was so confusing that I didn't even realize that the second location was not where I thought it was, but in the other part.
All in all, a complete waste of time. The one-sidedness of the maps was so bad that I couldn't even find my way around. Surely this can be improved.
|
Piccola's Doppengängerin
|
Topic: animals | Link here |
When we got home, took advantage of the fact that we had the dogs in the car, and drove up to Westons Road and took them for a walk there. We haven't really been there since our disastrous walk nearly 5 years ago, when we had only just got Leonid. For some reason, Yvonne stopped just before we got to the point where we got stuck, though we were walking:
|
|||||||||||||||||||||||
Things have changed a little in that time; the people to the north now have a cat:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
No, it's not Piccola (though we first confirmed that she was locked in the house, so she couldn't have been). And in fact she (he) is obviously younger, not yet fully grown. But also clearly a chocolate point Siamese, and not one of these rat-faced ones. I wonder if it's related to Piccola.
| Tuesday, 12 February 2019 | Dereel → Ballarat → Dereel | Images for 12 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another bloody migraine!
|
Topic: health | Link here |
Up this morning and in the course of the morning developed another migraine. That's the second in a little over a week. Why the change? I've been without migraine for years.
Could it be because I talked to Paul Smith about migraine last week? Hardly. But the reason we talked about it might be relevant: until recently, doctors recommended 100 mg aspirin per day for people in my age, as a prophylaxis against cardiovascular disease. Now they have changed their minds, and only people in danger should still receive it. That doesn't include me, so I stopped.
What do I do with remaining aspirin? Paul said that it's not even very good for pain: paracetamol is better. The only place where aspirin can be useful is for migraine.
So: no aspirin, migraine again. That fits. Is it the reason, or just a coincidence? And if it's the reason, is it withdrawal symptoms or liable to stay that way? To be observed. But articles like this one (British, and too close to “take two aspirins and call me in the morning” for my liking) suggest that there might be some connection.
|
Photographing wildflowers
|
Topic: gardening, photography | Link here |
There are a number of plants round here, somewhere between creepers and bushes, that are currently in flower. They have pretty deep blue flowers:
|
|||||||||||||||||||||||
Really? Nothing to be seen there. You need to come much closer:
|
|||||||||||||||||||||||
And even then it's hard to see. Still closer?
|
|
|||||||||
The funny thing is that they're easy to see from the distance of the first photo. Why is it so difficult to see in a photo?
|
Sleep-related issues
|
Topic: health | Link here |
I've had some leg pain in the early mornings recently, after about 6 hours' sleep. It's not going away, so off today to see Heather Dalman, the physiotherapist. It was a late appointment (17:00), so first off to the rather inappropriately named Forty Winks, where I wasn't knocked off my feet by salespeople. Finally got talking to Lyndon, who took me to a test bed to analyse my sleep requirements from lying on my back (not my normal sleeping position) without a pillow while the bed and associated computer machinery performed some unspecified functions. Then I was to lie on one side only for a minute or so.
The results? “Green”, which apparently meant “medium to firm mattress required”. He didn't give me the printout, and since only my Christian name was on the sheet, it's not clear what good it is to him. He also tried me with a couple of surprisingly thick pillows. At the end of it, I was left wondering how much science and how much hocus-pocus went into the evaluation.
Heather had a completely different approach. She wasn't interested in my sleep issues; it was on of three glutei, probably medius, but maybe minimus. A bit of massage, a bit of pep talk (“no hopping or skipping, jumping is OK”) and a surprisingly simple but difficult exercise: stand upright with your feet together, one a little in front of the other, and your eyes closed. Stay like that for about a minute. It sounds straightforward, but keeping absolutely still is (at least for me) surprisingly difficult, and Heather suggested that it might be even more difficult for some people: “Stand in a corner with your back to the corner. You could fall to either side or to the back, but not forward”. I don't see that I'm in any danger of falling—yet.
|
New Indian grocery
|
Topic: food and drink, opinion | Link here |
I've been looking for kidney beans to make frijoles de olla. That shouldn't be an issue, should it? No, it shouldn't be, but it is. Even white beans are sold here as “international specialties”, clearly ignoring the British staple baked beans, and last time Yvonne had to go searching for kidney beans, only to find some very strange looking ones.
OK, there's a new Indian grocery in Dana Street. Yvonne was there a while back to buy some garam masala, and if there's one place where you could expect to find dried legumes, it's an Indian grocery. So off to take a look.
What is it about Indian groceries that makes them look so tatty? It's a world away from the Korean Grocery that I visited in Bridge Mall ten years ago. And I've seen many like it: more a hole in the wall (in this case behind the Curry Star Restaurant) than a real shop.
But they have the stuff that I want, and it's not even as expensive as in the supermarkets, also a difference from the Korean shop 10 years ago. In addition, they're in sane quantities (1 kg instead of 375 g bags). So I got my beans, and I know where to go for other stuff when I need it. In addition I discovered that there are two different kinds of kidney beans, the dark ones (the ones I know) and lighter ones, the kind that Yvonne found last time. They also had two different sizes of white beans: “Great Northern” and “Lima”; the Limas were definitely too big for my liking.
| Wednesday, 13 February 2019 | Dereel | Images for 13 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Yet another bloody power failure
|
Topic: general | Link here |
Another short power failure at 4:01 this morning. I was going to say “hopefully this will soon be a thing of the past”, but in fact it won't; we'll still need Powercor for ovens, air conditioning and stuff.
|
Organizing TV series
|
Topic: photography, technology, opinion | Link here |
We download a large number of TV series. I have the space, so I keep the episodes when they're done, especially since in many cases they come out of sequence.
But how? TV and other content suppliers are terrible when it comes to consistency. It has taken me a while to come up with a consistent plan. Here with the example of Großstadtrevier (“The daily lives of police officers in the St. Pauli district of Hamburg”), and thus distinct from Notruf Hafenkante (“The stories of police officers and doctors working near the Port of Hamburg”; everybody knows, of course, that the port is centred around St. Pauli).
Why a German series? Because it's legal. I also get series from the Australian SBS TV, but they're usually shorter, and they have a strange choice of themes.
„Großstatdrevier“ has been going for 32 years now, but we only discovered it recently. Episodes are trickling in for no less than 10 seasons. How do I store them? How do we watch them?
After a number of false starts, I've decided on the following procedure:
Download the files with MediathekView, which gives me sets of files:
Get rid of those horrible spaces in the file names! For that I use the bash function despace, which replaces spaces and some other irritating characters with underscores.
Identify the season. The second one („Trinkgeld“) is clear: season 23, episode 12. But the first? IMDB has most episodes, though sometimes it's a week or two behind. But that's stored in 32 separate web pages. Somehow I need to get them together. Rather to my surprise, just concatenating them “works”, though it's a real mess and really uses lots of browser CPU time:
That's the current situation; I'll improve on it, and the latest version is here.
In this case, it proves that Großstadtrevier---Heile-Welt-20190213-220000.mp4 is season 27, episode 9. So I rename it with yet another small function, prepend, which also strips the (now unnecessary) title:
=== grog@teevee (/dev/pts/4) /src/Series/Großstadtrevier 52 -> l
=== grog@teevee (/dev/pts/4) /src/Series/Großstadtrevier 53 -> prepend -s Großstadtrevier--- 27-09- Großstadtrevier---Heile-Welt-20190213-220000.*
=== grog@teevee (/dev/pts/4) /src/Series/Großstadtrevier 55 -> prepend -s Großstadtrevier-- 23-12- Großstadtrevier--S23_E12----Trinkgeld-20190211-210000.*
=== grog@teevee (/dev/pts/4) /src/Series/Großstadtrevier 56 -> l
It's worth noting that the names are very inconsistent. That's the fault of the TV company (ARD in this case). Part of this step is to make things more consistent, so that they at least sort correctly in an ls listing. Thus the doubled episode numbers for 23-12.
Store these episodes in the hierarchy /spool/Series/Großstadtrevier/23 and /spool/Series/Großstadtrevier/27.
Link the images in this directory to the same name in another matching directory hierarchy, /spool/Already. This directory is intended to keep the files after we've seen them. For this I use the script already:
=== grog@teevee (/dev/pts/4) /spool/Series/Großstadtrevier/23 64 -> l 23-12-S23_E12----Trinkgeld-20190211-210000.*
=== grog@teevee (/dev/pts/4) /spool/Series/Großstadtrevier/23 65 -> already 23-12-S23_E12----Trinkgeld-20190211-210000.*
=== grog@teevee (/dev/pts/4) /spool/Series/Großstadtrevier/23 66 -> l 23-12-S23_E12----Trinkgeld-20190211-210000.*
=== grog@teevee (/dev/pts/4) /spool/Series/Großstadtrevier/23 67 -> l /spool/Already/Series/Großstadtrevier/23/23-12-S23_E12----Trinkgeld-20190211-210000.*
The difference in the two lists? The link count. When I finish watching this episode, I can just delete it. The copy remains in /spool/Already/Series/Großstadtrevier/23.
|
Goodbye, David
|
Topic: general | Link here |
Decades ago we bought a very poor copy of Michelangelo's “David” statue, for which we never found a good place, though ten years ago we tried some photos:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
But all that is no more. For a while he's been standing in front of the front door. Today I heard a noise and came out just in time to see the rod intended for the flower pots fall over, knock on a spade which then knock David forward on his face:
|
|||||||||||||||||||||||
Clearly an ancient ruin. How do we display it to best advantage?
|
Video stabilization with YouTube
|
Topic: photography, opinion | Link here |
On Monday I took a video of Yvonne trying on the riding vest, not helped by her nor Michelle getting in the way:
I somehow had to move while taking the video, and for that I thought it wasn't too bad. But there's video software available out there to stabilize videos, and one of the recommendations was to use YouTube's own stabilization.
OK, upload it. Damn, these files are big! This 134 second clip was 876 MB in size!
=== grog@eureka (/dev/pts/18) ~/Photos/20190211 187 -> l orig/*V
=== grog@eureka (/dev/pts/18) ~/Photos/20190211 188 -> mpid orig/*V
Came back and lost the page. Where's my video? I couldn't find it. OK, upload again.
The video is a duplicate of a video you already uploaded.
OK, fool, which? Ah, too polite to assume I didn't know that. It hadn't been published, so it wasn't in my “channel”. Finally I found it in some obscure corner (Video Manager, I think):
|
|||||||||||||||||||||||
OK, now to stabilize it. I've seen that before, somewhere in the Video Manager, but I couldn't find it. OK, look for a tutorial. How To Stabilize Shaky Video sounds right. And how about that, it showed me exactly what to do:
|
|||||||||||||||||||||||
Only that has absolutely nothing to do with what I saw:
|
|||||||||||||||||||||||
Elsewhere I had seen error messages telling me that certain functions weren't available to me (not leet enough, it seems; more likely didn't pay enough). Sorry, YouTube, if I have to pay money for the functionality, I'll buy real software.
|
Leonid unwell?
|
Topic: animals, opinion | Link here |
Leonid was a little strange today. He didn't go outside in the morning to urinate, and it wasn't until early afternoon, when we went for our daily walk, that he finally relieved himself, copiously. Why?
Then in the evening he didn't want to go outside to eat. Was he sick? There was no obvious sign. Yvonne contacted Pene, but she didn't have anything useful to say. While cooking food I had a bit of fresh minced veal left over, so I offered it to him. Yes, that was OK. Later Yvonne took him outside, and he ate again.
What's wrong with him? His actions gave us the feeling that he was afraid to go outside. In that connection, I noted that he came back inside as soon as we had finished our walk. Normally he stays outside for quite some time.
We've had a lot of kangaroos close to the house recently. Can it be that his experience two months ago has left him feeling insecure? To be observed.
|
MARION'S KITCHEN San Choy Bow
|
Topic: food and drink, opinion | Link here |
A couple of weeks ago Yvonne bought a number of spice mixes on special at ALDI, including four different kinds from MARION'S KITCHEN, which I have tried before. They're rather loud, and they tend to be expensive, but this special was round half price. So last week I tried their Phat Thai, and today we tried the San Choy Bow, a dish I had never heard of before. Neither, it seems, has Wikipedia.
Again far too much meat (500 g for 4 people), but this time I left it as it was. There's also a quantity of water chestnuts, which I extended because we had them, and Yvonne likes them.
The result? Boring. It's one of these fancy dishes where the important thing is serving the food in lettuce leaves. But the sauce tasted like pretty average soya sauce mix, and the balance just wasn't there. There's no reason to buy this stuff for a bit of sauce mix.
| Thursday, 14 February 2019 | Dereel | Images for 14 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Protected bacon
|
Topic: food and drink, opinion | Link here |
Yvonne brought home a new kind of bacon from Woolworths yesterday. It's called “Traditional Smoked British Style Back Bacon”, with the clarification “Australian Made The British Way”:
|
|
||||||||
That sounded good: “bacon” in Australia is very different from bacon in the UK, where they typically use pork belly (“streaky bacon”). But this was a different beast, much leaner, and very well protected. To get at it at all I needed scissors and a knife, and even then I had my difficulties:
|
|||||||||||||||||||||||
The result was a completely destroyed package, so I needed a new bag to put it in.
And the result? Boring. Given that this is in 200 g packages (normally we buy by the kilogram), I'd expect that it's also expensive. That's the last time.
|
More phone spam
|
Topic: general, technology, opinion | Link here |
It seems that some non-obvious configuration change that I did to my phone setup has resulted in me finally getting emails with voice mail from my main phone number, which is on permanent auto-answer to discourage spammers. In my mail today:
I don't recognize any of the numbers, but I still need to listen to them. They're all the same! An auto-spam caller which conveniently leaves out the identification of who they are. Clearly the faked phone numbers don't help.
|
Another bloody migraine!
|
Topic: health, opinion | Link here |
Another migraine today. Clearly something has changed, and I need to do something about it. It started after I stopped taking my daily low-dose aspirin, so an obvious solution is to start taking aspirin again. As they say, for any complicated problem there's a solution that's obvious, simple, elegant... and wrong.
Tried to get an appointment with Paul Smith, but he's booked out until this time next week, so I have an appointment with Dr. Geoff/Jeff tomorrow at noon.
|
Leonid's progress
|
Topic: animals, opinion | Link here |
Leonid still doesn't want to stay outside the house alone, but is happy enough when we're there with him. So is he afraid of kangaroos? Today we went walking to the „Große Linde“ (really a conifer) and saw a mob of kangaroos. No, Leo was all for chasing them. That's not the answer.
And then in the evening he had trouble eating. Yvonne tells me that he took a bite of chicken frame, then yelped and spat it out. Throat obstruction? Seems unlikely over such a period. We gave him some softer food, which he ate; potentially his throat is just still sore. More observation required.
|
Toll: we can't be bothered
|
Topic: general, opinion | Link here |
On the way back from walking the dogs, found a notification in the letter box:
|
||||||||||||||||||||||
That must be my phone. Damn! Did we miss him while walking the dogs? No, the time was clear: 9:50, when I was either in the office or preparing breakfast. So why didn't he ring? Couldn't be bothered? Dammit, this is a priority package!
Called up the number on the card, 1300 865 547, and was asked “Please enter your Commnote number or press * if you don't have one”. What's a Commnote number? Why do they continually use terms that don't appear on any document?
Connected to Jamie, who apologized for the non-delivery and told me that I had called the wrong number. But it's the number on the card! No, this is a Toll Priority delivery, different number (131 531). The driver must have left me the wrong card.
She connected me to wait music that was finally ended by a barely intelligible Steve, who wanted to know about a chnbdx, which proved to resolve to “check box”. OK, which check box? The one at top right. There's no box there. OK, is this a new item or one to be delivered? Sometimes I wonder what these people use for brains. In any case, he arranged to have it delivered tomorrow between 8:00 and 17:00 (I think; I really missed a lot of what he was saying), gave me the reference number 07446616 and the suggestion that I put a note on the front door to advise him that we're at home—about the only good thing he said. Of course, it went on the letter box:
|
|||||||||||||||||||||||
Hopefully that will get the desired results.
What a company! As the Germans would say, ironically, „Toll!“
| Friday, 15 February 2019 | Dereel | Images for 15 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More strange voice mail
|
Topic: general, opinion | Link here |
After yesterday's phone spam attack, I received another voice mail message today:
But this one was completely different. It was my sister Bev, calling (as I later discovered) from India to tell me that our mother had died at home in Bendigo yesterday (which proved to have been two days ago).
Damn! I hadn't seen her for quite some time, and I had been paving the way to hell with good intentions. But like with my uncle Max last year, events overtook me. When I later got in contact with my cousin Brendan (Herbert), he told me that she had apparently died, unexpectedly, in her armchair and had been discovered by the chemist who had come by to deliver some medication. The good news is that she went peacefully, and at nearly 95 years old she had had a good life span and also the consolation of not having to go into a nursing home, which she had resisted strongly. She had been going downhill for the last six months. But somehow it's sad that she didn't quite make her 95th birthday (on 12 April).
The whole situation was unexpected enough that her body has been sent to the coroner in Melbourne for investigation, and we don't know when we will see her again.
Other details: she wanted to be buried in Bendigo, and Bev, who will be back home in the UK tomorrow, will try to be there too. There's also a house to clean out, and Bev suggested (thanks, Bev!) that she and I should clean it out together. I wonder what old, worthless treasures we'll find there; at least some of my early documents that she had held on to as souvenirs or something.
|
Garden work again
|
Topic: gardening | Link here |
Mick the gardener along today to do mainly mowing and weeding. He's still not his old self—he may never be—and gave up after four hours, having at least considerably tidied the place up.
|
Doctor again
|
Topic: health, opinion | Link here |
As a result of my migraines over the last 10 days, into town today to see Dr. Geoff, whose full name proves to be Geoffrey Chen. Is it migraine? Let him diagnose.
In the end he decided that I should start taking aspirin again and see how that works out.
|
New statues?
|
Topic: general | Link here |
While in town, dropped into the garden statue place in Wallis St, where indeed they had David statues on sale that made it uninteresting to repair the old one. These ones are twice the height, and have much more detailed features:
|
|||||||||||||||||||||||
The real question is, do we need one? We never really found a good place for the old one. But I'll have to discuss that with Yvonne.
|
Phone repaired
|
Topic: general, technology, opinion | Link here |
Back home, and finally my phone had come back, no thanks to Toll, who had caused no fewer than 5 phone calls and lots of frustration dealing with near-humans, with two attempts at each end. The phone was delivered by the same cynophobe who picked it up. Was he so afraid of our dogs that he didn't dare to go to the front door? Then he should look for a different job. And somehow the state of the package seemed to reflect their lack of care:
|
|||||||||||||||||||||||
Of course the first question was: what had been wrong with the phone? The slip stated:
Replaced lower pcba. Performed software upgrade and FQC functional test. Phone was finally QA tested to manufacturer specifications.
All that looks perfectly understandable. Hardware failure. But there's a fly in the ointment there: the phone was returned with Android 8.0, though the latest version (which I had installed on the old incarnation) is 8.1:
|
|||||||||||||||||||||||
On the positive side, the phone was still not locked to any network. Maybe this happens automatically if you have a repair, or maybe it is related to some part of the phone that wasn't replaced.
OK, install that, wondering in the process if Google doesn't offer a way of reinstalling the apps previously installed. This should be particularly important if you've paid money for them (something that I have so far resisted). Is the IMEI number the same? Later check: yes, it is. But as it was, after waiting the eternity for the OS upgrade to trickle in and to discover that no further updates were waiting, I just installed the ones I wanted again, giving me the advantage of getting rid of stupid, broken apps like VicEmergency's incident app. And then the phone told me ha ha, only joking, there is another update to load.
How I love Android! If there's one good thing about the iPhone 6, it was to show me the positive sides of Android. But “love” is most certainly not the most appropriate term.
|
More worries about Leonid
|
Topic: animals, opinion | Link here |
Leonid seems normal enough most of the time. But he still doesn't want to be outside by himself. He's fine when we're with him, but if we put him outside by himself, he tries to get back inside as quickly as possible.
We had wondered about this over the last couple of days. Fear of kangaroos? No, he disproved that hypothesis yesterday. But Petra Gietz was here today, and she suggested that it might be magpies. They're quite aggressive, and some have settled in the tree across the road from our house entrance. It seems quite possible that that's the cause.
But that's not the only problem. Feeding him this evening was a problem because it's normally outdoors. So I took him into the garage, where he took a 4 cm cube of beef, apparently swallowed it whole, howled and stopped eating. We couldn't even persuade him to eat soft dog food.
Clearly that's not normal, and it has nothing to do with the magpies. There must be some issue with his throat, but we couldn't see it. We need to get him to a vet tomorrow.
| Saturday, 16 February 2019 | Dereel → Cape Clear → Dereel → Ballarat → Dereel | Images for 16 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Leo, day 3
|
Topic: animals | Link here |
Clearly we had to do something about Leonid's eating pains, so off to Cape Clear before breakfast to show him to Pene Kirk. She had already said that she wasn't equipped to examine him properly, but she took a quick look and saw a swelling round his larynx. Tumour? Maybe thyroid problems? I should get him looked at in a vet practice.
But which one? I had been less than positively impressed last time we had him at a vet, this time last year. What I perceive as a lack of experience resulted in us getting a bill of nearly $1000, not to mention the worry it caused. In the end we decided to go back to Ballarat Veterinary Practice, where we haven't been for over 5 years. They, too, are expensive, but we had no issues with the quality of the care they gave.
Yvonne was able to get an appointment so quickly that we barely had time for breakfast. Then off, spending the time waiting talking to Brendan and Robert Herbert about the funeral arrangements.
Finally got to talk to Richard Lawrence, who took a look at Leo and decided that the swelling was only on one side of the throat, near the epiglottis, and suspected a grass seed lodged in his throat. Most times, he says, it will detach by itself, though the possible requirement of surgical removal remains. He gave him a painkiller and a course of antibiotics (Juroclav 500 mg, one twice a day for a week). Yvonne was visibly relieved; she had been reckoning with a malignant tumour. Hopefully he's right, but he made a very good impression. I wonder what a junior vet would have done.
In the evening Leo should have eaten. He took one look at the food and turned away. Dammit, what does he want? After a couple of attempts, took some food and put it in his mouth. He chewed, dropped some but swallowed most of it. How about the food now? No, thanks. Tried again, same thing.
Yvonne took over, and after another attempt he finally got the message and finished his food. So apart from the pain, there must have been a significant psychological factor involved.
|
Garden flowers in late summer
|
Topic: gardening | Link here |
Middle of the month, time for the monthly garden flower photos. It's been a long, dry summer, and many of the plants are suffering as a result. I've been following the Hebes, which suffered badly from the combination of heat, drought and problems with the sprinklers. The sprinklers are still not working as well as they should. The result are some very sick looking plants, and I wonder if they will survive:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Others, though, are definitely surviving:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
I hope that they'll improve as the weather cools down.
The trees in the area are also suffering. Our oak is still alive, and so is the ornamental cherry, but I can't say that they look good:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
On the other hand, the Paulownia kawakamii seems to have benefited from the fertilizer, making me wonder whether I shouldn't spread a lot more:
|
||||||||||||||||||||||||
Another genus that seems to be benefiting from fertilizer are the tomatoes. These ones are on the north side of the house, and they're promising to bear more fruit than we can eat for the next few months:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
On the other hand, the ones round the water tanks are anything but happy:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
They're much smaller, and the second one has no fruit at all. They're getting enough water; time for more fertilizer.
The gladioli finished flowering a month or two ago—most of them. For some reason, probably related to irrigation, we have two new shoots:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
And one of the Hibiscus syriacus is also flowering:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The other, more sheltered one, is growing better, but so far there has been no evidence of buds.
The roses are also doing better than it's easy to demonstrate:
|
|||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
And the corner succulent bed is gradually looking the way we would expect:
|
|||||||||||||||||||||||
|
Arranging the funeral
|
Topic: general | Link here |
When my father died, my sister Bev completely took over the funeral and other arrangements, and I wasn't involved at all. With my mother it's different: we now have four people involved. Apart from Bev and myself they're our cousins Brendan and Robert Herbert, who live in Bendigo and thus had frequent contact with my mother. As a result of our discussion on the phone today, and also a mail exchange with Bev, we have:
What has occurred to me since then:
|
MARION'S KITCHEN massaman curry
|
Topic: food and drink, opinion | Link here |
Another dish from MARION'S KITCHEN this evening, a Thai “Massaman curry”. Once again ridiculous quantities of meat, and otherwise not very convincing. The other packages go back.
That has been four different dishes from MARION'S KITCHEN. The curry laksa was OK, the Phat Thai acceptable, but the San Choy Bow and the Massaman curry are really not worth the trouble.
| Sunday, 17 February 2019 | Dereel | Images for 17 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Coroner and funeral arrangements
|
Topic: general, language, opinion | Link here |
Rather to my surprise, got a call from Tracy at the Coroners Court of Victoria today, a Sunday. She didn't ask for any form of identification, but had some important information:
Another question we're going to have to address: who is going to pay for the funeral? Various discussions suggest that her estate may not be released for quite some time.
|
Another wildflower
|
Topic: gardening | Link here |
While walking the dogs, saw this relatively forgettable flower:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
The only interesting thing is that it is a bush, while most flowers round here are very small and grow directly on a stalk. The flower looks like a Goodenia, but the leaves of the bush look more like a Hibiscus.
|
Leonid's progress
|
Topic: animals | Link here |
Leonid is much better after his medication, and ate a lot of food this morning. He's also much livelier. The most surprising thing, though, is that he is no longer afraid of going outside. Certainly it wasn't magpies: we went straight past some returning from the walk, and he didn't pay any attention to them.
So somehow his throat pain made him not want to stay outside. Why?
| Monday, 18 February 2019 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
My mother: cause of death
|
Topic: health, general | Link here |
Another call from the Coroners Court of Victoria today, this time Sharee, with the results of the investigations of my mother's body. Nothing unusual: she suffered from coronary heart disease, hypertension and hypercholesterolemia. Nothing unusual there in a 94-year-old, except the question as to whether she had been undergoing treatment. No specific cause of death, though a heart attack seems the most likely, but also nothing suspicious, so she can return to Bendigo as soon as Brendan has found a funeral company.
And when will the funeral be? Bev can't make it here until the beginning of next month, so it looks like it'll be nearly 2 weeks yet.
|
More BigPond pain?
|
Topic: technology, general, opinion | Link here |
Both Brendan and Robert Herbert have email accounts with BigPond. That's a real pain: I've had great trouble communicating with BigPond subscribers, and it seems not to have improved. The issue is their anti-spam measures, which seem to exclude security (digital signatures trigger their mechanism, which appears to simply move the message to /dev/null). I turned off all security today, but that doesn't seem to have been enough. What a pain these people are!
| Tuesday, 19 February 2019 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Ring tones again
|
Topic: multimedia, technology, opinion | Link here |
A few months ago I set a custom ring tone on my mobile phone. All well and good, except that it wasn't loud enough. I thought that mencoder might do it, but failed to find out how. Callum Gibson suggested that ffmpeg could do it.
Now that I have effectively a virgin phone again, time to try it out. Started with a search ffmpeg set volume, which gave me lots of good looking hits, notably this one, which also has the advantage of being nicely formatted. There I read:
If you want to normalize the (perceived) loudness of the file, use the loudnorm filter, which implements the EBU R128 algorithm:
ffmpeg -i input.wav -filter:a loudnorm output.wavThis is recommended for most applications, as it will lead to a more uniform loudness level compared to simple peak-based normalization. However, it is recommended to run the normalization with two passes, extracting the measured values from the first run, then using the values in a second run with linear normalization enabled. See the loudnorm filter documentation for more.
To automate the normalization processes with ffmpeg without having to manually perform two passes, and run normalization on multiple files (including video), you can also use the ffmpeg-normalize Python program via pip install ffmpeg-normalize.
OK, what does that mean? How do I do a two-pass normalization? Followed the link, which gave me nothing useful, just a reference. About the only suggestion was the Python script, which seems a bit over the top.
OK, try the example:
=== grog@eureka (/dev/pts/28) ~/Downloads 2 -> ffmpeg -i Andante-ringtone.mp3 -filter:0 loudnorm Andante-loud.mp3
Ugh. How I hate coloured output! And why can't it create MP3 output? Much searching around, during which I found various suggestions, none of which worked. OK, dammit, let's output in WAV format.
=== grog@eureka (/dev/pts/28) ~/Downloads 4 -> ffmpeg -i Andante-ringtone.mp3 -filter:0 loudnorm Andante-loud.wav
Dammit, what does work? Messed around for a while and found that I could set the volume with things like
=== grog@eureka (/dev/pts/28) ~/Downloads 5 -> ffmpeg -i Andante-ringtone.mp3 -filter:a "volume=3" Andante-loud.wav
=== grog@eureka (/dev/pts/28) ~/Downloads 6 -> l Andante-loud.wav
OK, we can get rid of that. And finally I got some output!
Next, how do I convert it to MP3? Do I need to? Upload it to the phone and see what the phone things of a WAV file. And the next question: where do I put it? Android storage hierarchies completely confuse me, and I forgot to note such details last time. OK, /storage/emulated/0/Music seems an obvious place, at least for this phone. Upload there, select Sound/Phone ringtone, and sure enough, it found it and was able to set it. One point to Android, none to ffmpeg, though arguably this is an issue with the FreeBSD port.
|
Statin test?
|
Topic: health, opinion | Link here |
2½ years ago I went to see Dr. Paul Smith for a second opinion about a prescription of Rosuvastatin from my new doctor. I wasn't happy about it, and Paul agreed. Ultimately I decided to change to him as my new doctor.
All the more surprise when I recently received a letter from Monash University asking me to participate in a clinical trial in the use of statins for the prevention of dementia. But it seems that in the intervening time people have decided that the side-effects of statins have been overestimated.
OK, always interested in an experiment. Called them up and was subjected to a verbal questionnaire. Do I have diabetes? Yes. OK, sorry, can't include you. I thought about it: after all, Paul suggested me, and it's really borderline. But looking at it from their perspective, I'm the kind of issue they don't want to know about. Still, there could be other studies...
| Wednesday, 20 February 2019 | Dereel | Images for 20 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Gymnastics in the Hall
|
Topic: health | Link here |
Down to the Dereel Hall this morning to start a gymnastics course. Not my favourite pastime, even when I was young, and nothing has changed. But both Yvonne and Dr. Paul Smith thought it would be a good idea.
Was it? I don't see any advantage. There wasn't really anything that I couldn't have done equally well at home. But I'll carry on for a while. It does have the one advantage that I will really do the exercises.
|
The demise of μFT, part X
|
Topic: photography, opinion | Link here |
More doom and gloom about the future of the Micro Four Thirds system today. Reports are that Olympus have actually announced the end-of-life of the Olympus PEN-F, without a replacement.
Is that wrong? It's clear that the camera industry as a whole is shrinking, and there's just no longer any justification for the plethora of models that Olympus has. Currently there are 6 models, including the PEN-F: the E-PL9, the PEN-F, E-M10, E-M5, E-M1 and the E-M1X. That contrasts with the three or four models that they had 10 years ago in the DSLR days. It makes sense to reduce the number of models.
But which? This page shows a comparison of the appearance of the cameras, starting with the E-PM2, which Yvonne has: one of my concerns is what to replace it with when it wears out. There are really only three different kinds:
|
|||||||||||||||||||||||
I never found the PEN-F to be interesting, and maybe I wasn't alone. Which should remain? Some discussion on the (closed) Facebook group. Maybe they should reduce the range to the E-M1, the E-M5 or E-M10 (there's really not that much to choose) and the E-PL series. I still can't accept the E-M1X.
As if that wasn't, there was also a video by an Olympus aficionado decrying the FUD surrounding the rumours, but also coming to the (sad) conclusion that it is going to go away, and that in a matter of months.
I'm not so sure. Yes, apart from this abomination of an E-M1X, there have not been any new cameras for a while. But they're still bringing out interesting new lenses: the Panasonic Leica Vario-Summilux 10-25 mm f/1.7, the M.Zuiko 12-200 mm f/3.5-5.6 and a 150-400 mm lens with integrated teleconverter (making it 150-500 mm, longer than anything else I have seen). That doesn't look like boring lenses or the actions of companies that intend to stop production in the next few months. On the other hand, until they have some real improvement on the current cameras, why should they bring out something barely different? I still see the E-M1X as a way to explore new features.
|
Facebook breakage and insults
|
Topic: technology, general, photography, opinion | Link here |
I tried to post the URL of the photos above on the (closed) Facebook group. I failed: Facebook mangled the URL to a point that it couldn't find anything at all. Spent quite some time trying to work around the breakage, watching things get worse and worse:
|
|
|||||||||
|
|
|||||||||
Does it not like the “Bloody Facebook!”? No, a transient popup told me something about them sending an email. And to a certain extent, that's what happened:
Date: Tue, 19 Feb 2019 20:05:06 -0800
From: Facebook <security@facebookmail.com>
To: Greg Lehey <groggyhimself@lemis.com>
Subject: We sent you a message about your removed post.
X-Mailer: ZuckMail [version 1.00]
Reply-to: noreply <noreply@facebookmail.com>
========================================
Go to your Support Inbox
https://www.facebook.com/n/?support%2F&item_id=102173489157...
========================================
Hi Greg,
Hi Greg,
We sent you a message about your removed post.
- The Facebook Help Team
WHAT YOU CAN DO Accept DecisionNo one else can see your post and you'll no longer be able to request a review Request ReviewWe'll take another look at this post
Thanks,
The Facebook Team
Now doesn't that make a lot of sense? They didn't even say what the issue was. For that I had to go to another place. Why didn't they just give me a link?
Finally found my way there and discovered:
|
|
|||||||||
Spam? Idiots! OK, what can I do? Grin and bear it or ask them to review, with no further feedback. It doesn't seem to be possible to contact a human being there. How do you report bugs like this one? Facebook, I hated you before. Now it's with a vengeance.
|
More solar energy feedback
|
Topic: general, opinion | Link here |
Talk with Tomas Kucera of EffectiveElectrical today. It seems that Powercor have only approved 2.8 kW feedin to their network instead of the normal 5 kW. What can I do about it? Nothing. It seems that the courts are on the side of the companies, not the people. About the only good thing is that the feedin tariff is now $0.11 per kWh instead of the $0.05 that I had been fearing. But at least Tomas did his research first, which is positive. Hopefully he'll be able to give me a good price.
| Thursday, 21 February 2019 | Dereel | Images for 21 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Another power failure
|
Topic: general | Link here |
Another power failure today while we were cooking breakfast. And I didn't even notice until I tried to heat something in the microwave oven.
When did it happen? Both the microwave oven and the conventional oven are too polite to display anything until I tell them, so they were both showing “power on”. But when I went to look later, the second microwave oven in the pantry was showing 4:15 or so. What does that mean? What does it reset to? I've decided that it must reset to 12:00 (a likely value) and then count in 12 hour notation, so the power must have been reset 4 hours, 15 minutes previously. But I had thought that it reset to 1:00 (another obvious value). But that would give me a time later than when I discovered the issue, so I'll have to go for 10:10 and the oven resetting to 12:00. I'll have to keep an eye on it next time.
|
What does “ISO” mean?
|
Topic: photography, opinion | Link here |
Tony Northrup has brought out another video: “ISO is totally FAKE. Seriously”. I've already found fault with his videos; last time I accused him of deliberate falsification. This time I'm not sure: maybe he just doesn't understand the issues.
His argument is that “ISO” (he means the arithmetic sensitivity measurement according to ISO 12232:2019, or maybe the older ISO 12232:2006) is not consistent between makers, and that it doesn't mean anything anyway because it can be compensated for in post-processing. He gives examples with the Olympus OM-D E-M1X, which, he says, is out by 1 EV from the stated ISO, as demonstrated by photo from a different camera. And yes, indeed, the two images appear offset by 1 EV or so.
But what is ISO? He seems to think that it has its roots in the ASA system, which is only part of the truth. ASA is one of the most recent of a plethora of sensitivity systems and goes back only to 1943, over 60 years after Leon Warnerke brought out his first system.
He also seems to think that a typical rating has always been 100/21° ISO, when in fact for much of my youth that was a relatively high sensitivity. My father took ciné films with Kodachrome I, with a sensitivity of only 12 ASA, or (coincidentally) 12° DIN, and I normally used films with sensitivities between 25 ASA/15 DIN and 125 ASA/22 DIN.
But how do you measure photographic sensitivity? I have a book somewhere, but I can't find it. And there's not very much information about it on the web. Determining film speed states:
Film speed is found from a plot of optical density vs. log of exposure for the film, known as the D–log H curve or Hurter–Driffield curve. There typically are five regions in the curve: the base + fog, the toe, the linear region, the shoulder, and the overexposed region. For black-and-white negative film, the "speed point" m is the point on the curve where density exceeds the base + fog density by 0.1 when the negative is developed so that a point n where the log of exposure is 1.3 units greater than the exposure at point m has a density 0.8 greater than the density at point m.
It also shows this graph:
Which system? The graph refers to ISO 6:1993. And these slides state that for DIN 6-1974, the sensitivity is defined in the same way:
|
|||||||||||||||||||||||
That's completely the other way round from the way you'd do it nowadays, where you go from the highlights. And it means that the sensitivity depends on the dynamic range of the sensor. Increase the dynamic range by 1 EV and you're also increasing the sensitivity by 1 EV.
Is that the way ISO 12232:2019 still defines it? I can't find out. I would have to pay CHF 118 just to look at the standard. The best I can do is this preview, which shows a tastefully truncated table of contents:
Another detail from the preview is interesting:
this document specifies an exposure index (EI) that corresponds to the focal plane exposure of a typical mid-tone.
This page also contains interesting detail. I haven't finished digesting it yet, but it basically confirms what I have written above, and goes into detail about the saturation-based and noise-based calculations. What I can say so far is: there's no single way to look at things.
|
Tony Northrup's understanding of image capture
|
Topic: photography, opinion | Link here |
But Tony Northrup wasn't done with saying that some makers lie about camera sensitivity. As the name (“ISO is totally FAKE. Seriously”) suggests, he says that the whole concept is wrong. Got a problem with an image exposure? Fix it in post-processing.
That's an overly simplistic view of how it works. What we have is:
How does this conversion take place? Tony doesn't go into detail, maybe because he doesn't understand it. What he does say is that you can take these numbers after storage and manipulate them. For a correctly exposed image, the highlights should be round 4095. For an image that's underexposed by 2 EV, they'll be round 1023. No worries, multiply by 4 and Bob's your uncle. He even shows an image that he corrected like this, and it looks OK on his video.
OK, I can do this too, in the process checking what he claims about sensitivity. Here's a boring photo of the back of my office, taken with my Olympus OM-D E-M1 Mark II with automatic exposure (P setting):
|
|||||||||||||||||||||||
I was going to follow this with a manual exposure using the settings shown by my exposure meter, but they were the same. Whatever issue might exist with the camera, it wasn't the exposure metering. And does this image look underexposed? Not to me. Yes, there are issues in the shadows—that's why I chose this motive—but make the highlights any higher and they'll be burnt out.
OK, now let's try with 6400/39° ISO, followed by the same image (run the cursor over an image to compare it with its neighbour):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, there's a minor difference in brightness there. Is it maybe related to the dynamic range? That would almost make sense. But it's certainly nothing like the 1 EV difference that Tony Northrup claims.
Next, let's look at his claim that you can compensate for exposure in postprocessing. Here's the same image again taken with the same exposure as the 6400/39° one, followed by the same image corrected in post-processing, and finally the correctly exposed 6400/39° one:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Yes, the results are surprisingly good. But what's that colour cast? My take is that that's a non-linearity in the sensor curves between the different colours. The problem is that multiplying all values by 32 means that the low-order pixels have values 32, 64, 96, 128 etc., and nothing in between. And that is the difference between setting the correct ISO sensitivity when taking the photo, and faking it after the event. Like converting to JPEG, you're losing detail with underexposure.
|
Kangaroo jawbone
|
Topic: animals | Link here |
While walking the dogs today, found this:
|
||||||||||||||||||||||||
It's the lower jaw of a kangaroo, which Nikolai found on the side of the road and then decided that it was uninteresting. Leonid got it a little later and also decided that it was uninteresting, but not before breaking it: the join on the sides is quite fragile.
| Friday, 22 February 2019 | Dereel | Images for 22 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Contact details in Malaysia
|
Topic: general | Link here |
It's been 4 days since I sent a mail message to Tengku Razaleigh, and so far I haven't had much feedback. So how do I contact people? And should I contact the Sultan of Kelantan or not? Maybe the Malaysian High Commission in Canberra could give me advice about correct protocol. Called 02 6120 0300 shortly before noon and was waiting for an answer before it occurred to me that that was probably the worst time of the week to call, so put it off until Monday.
|
More system update stuff
|
Topic: technology, opinion | Link here |
As a result of other recent issues, I've lost track of updating eureka. Back to work today, wondering how to migrate as painlessly as possible. And then I thought of x2x, a program to join X servers. I used it for years in Wantadilla, but stopped 12 years ago when I moved to Dereel in the assumption that four displays are enough, and I can do that with a single system.
But I have a fifth monitor, and in fact the display card in eureso can drive two displays—I think. Why not connect the left-most monitor on eureka (eureka:0.0) to eureso as eureso:0.1, and connect eureso:0.1 to eureka:0.1 via x2x. That way I can carry on using the old eureka where I have issues with the new one.
To my surprise, x2x is still installed on both machines. But I've forgotten how to use it. While playing around, discovered that the options on the two versions are different, and in particular the one on eureka won't allow me to go up or down (“north” and “south” in x2x-speak). I'm sure that that wasn't always the case.
And where have all the mice gone? I couldn't find one for eureso. While messing around and trying to run X on eureso without a mouse, discovered that it's not a good idea to run x2x on both sides of a join: it seemed to hang both servers (maybe just x2x looping), and I had to shoot them down before I could continue.
In the end got it working, but the X configuration on eureso is messed up: only one display works, and that at 1280x1024. Should I fix it, or just move on to the planned update without the safety net?
| Saturday, 23 February 2019 | Dereel | Images for 23 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Soft hardware issues?
|
Topic: technology, opinion | Link here |
Coming into the office this morning, found an unexpected message from my overnight backups:
Date: Sat, 23 Feb 2019 04:00:00 +1100 (AEDT)
From: Cron Daemon <root@teevee.lemis.com>
To: root@lemis.com
Subject: Cron <root@teevee> /home/local/bin/cleanup
/bin/sh: /home/local/bin/cleanup: Input/output error
Huh? This was clearly from teevee. And sure enough, I couldn't access /home/local, which is NFS mounted from eureka.
More problems with /eureka/home, that disk that I've been promising to replace for over a year now? No, it was OK. OK, remount the file system. Nope. I/O error.
Log file time. The backup of lagoon completed successfully at 21:06, so there was a 7 hour window. Nothing much there except:
The “error” messages really need improving. The take-home message is “Logical unit is in process of becoming ready”, in other words spinning up, and it's normal. No idea what the “could not remount” message means; I suppose I should check on that. And yes, all these messages were at 21:00 on the previous evening. But there wasn't anything else until I tried the remount, which just gave me:
And teevee's log files showed nothing.
What's wrong there? Not all file systems on eureka were affected; /Photos and /dump were still mounted.
In the end, I SIGHUPped mountd and tried again. That worked. But what caused the problem? Too late to look now, I suppose. There are probably good reasons to take core dumps in cases like these.
|
Android battery problems?
|
Topic: technology, general, opinion | Link here |
I am somewhat paranoid about keeping batteries charged, and I try to ensure that my devices with non-replaceable rechargeable batteries are on charge at all time. But since getting my phone back from repair last week, it doesn't really seem to have been charging properly:
|
|
|||||||||
|
|
|||||||||
That was after being powered off and connected to the charger overnight. Clearly something's wrong. Has the problem that required the repair also damaged the battery? The thought of the pain of sending it off again was unbearable.
OK, what if the charger or cable is defective? Tried another one, and how about that, it worked! Was the first one too weak? Maybe, but not according to the specs: both chargers and cables looked identical, though 0.25 A doesn't seem overly powerful. How much current does it need? Ah, they're too polite to assume that I don't know that, so they don't mention it in the instruction sheet. Clearly I'll have to do some measurements.
|
Downloading web audio
|
Topic: general, multimedia, opinion | Link here |
The majority of my multimedia content now comes from the web, and in general broadcasters are very forthcoming with their content. So when Melanie Bahlo told me of an interview that she did on ABC radio last week and asked me to download it, it sounded like a no-brainer.
Maybe a brain is missing somewhere, but I failed. mpv failed because it didn't have a download link. youtube-dl, which normally works on all sorts of web pages, drew a blank. OK, take a look at the web page, which brought up a few interesting links: https://abcradiomodhls-vh.akamaihd.net/i/local_melbourne/audio/aba-2019-02-18.m4a/master.m3u8 proved to be only 3 lines long:
OK, that looks promising. Take a look at https://abcradiomodhls-vh.akamaihd.net/i/local_melbourne/audio/aba-2019-02-18.m4a/index_0_a.m3u8. That was much longer:
There were hundreds of these URLs. That looks like a whole lot of individual 10 second clips, and indeed that's what it is. I could write a script to collect the whole mess together, but isn't there an easier way?
| Sunday, 24 February 2019 | Dereel | Images for 24 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Nail clipping
|
Topic: animals, opinion | Link here |
In the afternoon, as usual, I grabbed my camera and prepared for walking the dogs. But Yvonne was busy clipping their claws:
|
|||||||||||||||||||||||
That in itself was nothing unusual, but look at Nikolai (on the right). In the past he was terrified of having his claws clipped. And here he is showing interest. And sure enough, Yvonne was able to clip his claws with no trouble:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
All because he saw Leonid getting treats for each clipped claw?
|
Downloading ABC audio
|
Topic: multimedia, technology, opinion | Link here |
So how do I download yesterday's audio? Surely there must be a solution out there. And sure enough, there is: Chrome Audio Capture, clearly intended for Google Chrome. Installed that and found that yes, it would work, but could only capture 20 minutes' worth of audio due to “Chrome memory constraints”. The programme was 2 hours long. OK, tried it anyway. It proved that Melanie's part of the programme was only about 11:43 long, so that was no problem.
But there were mentions of the issue earlier on, where the host (clearly particularly tongue-tied) promised a “genecist”. How about that, I can override the 20 minutes, and sure enough, it ran for the full 2 hours (in real time!). But it hung after about 80 minutes.
Surely I can do better than that. Look at that https://abcradiomodhls-vh.akamaihd.net/i/local_melbourne/audio/aba-2019-02-18.m4a/index_0_a.m3u8 file that I downloaded. With only minor tweaking I was able to download the components:
=== grog@teevee (/dev/pts/8) /spool/Download 103 -> grep ^h index_0_a.m3u8 | sed 's:^:youtube-dl :'| sh
And how about that, it downloaded all 690 segments in only half an hour (due to the immense overhead of a separate download for every 10 seconds):
Can I just concatenate them? Not that simply, due to the out-of-sequence file names, but:
=== grog@teevee (/dev/pts/8) /spool/Download 106 -> cat segment?_0* segment??_0* segment???_0* > Bahlo
And how about that, it worked!
What format? For some reason none of my programs was really happy about telling me. TS suggests a transport stream, but which? Looking at the file itself was not overly helpful beyond suggesting that it was created by ffmpeg:
=== grog@teevee (/dev/pts/1) /spool/Download/Melanie 5 -> less segment1_0_a-segment1_0_a.ts
Still, what's ffmpeg for?
=== grog@teevee (/dev/pts/4) /spool/Download 5 -> ffmpeg -i Bahlo Bahlo.mp3
But a conversion to Ogg format worked.
|
Gene testing on twins
|
Topic: health, opinion | Link here |
The real topic of Melanie Professor Melanie Bahlo PhD's interview was a
relatively interesting one: two twins sent
off DNA samples to a number of testing labs,
and not one of them noted that they were twins. I didn't get all the details, just
Melanie's answer, that though the results would have been the same for all labs, the
algorithms that they use are not, and this can come up with wildly differing results.
Somehow that's not very satisfying. These tests test for about 600,000 SNPs. How many would differ? That depends on whether they're monozygotic (“identical”) or dizygotic (“fraternal”). But the level of discussion on this programme was a little too low to even mention this distinction. I'll have to discuss it with Melanie next time I see her.
|
MP3 formats and ffmpeg
|
Topic: multimedia, technology, opinion | Link here |
Why won't ffmpeg process MP3? Surely it can, but the default FreeBSD port is too polite to include software for formats of dubious license status, such as MP3. OK, build the port manually. No way to select MP3, but there's something about “nonfree” formats. OK, configure that and build, watching confirmation of MP3 input and output support fly by.
Install and... still no MP3 support. What went wrong there? OK, put it off for another time.
|
Ports hell
|
Topic: technology, opinion | Link here |
On to watch TV.
=== grog@teevee (/dev/pts/2) /spool/Images 25 -> mpv Water_Diviner_The-2016-04-23-2131
Ugh! Not only did my ffmpeg not deliver the desired results with MP3, it delivered some very undesired results. What do I do now? Upgrade mpv? Or remove ffmpeg?
=== root@teevee (/dev/pts/11) /home/grog 1 -> pkg delete ffmpeg
Those are the programs I use most! Damn that! So I did a make deinstall in the port directory instead. Then:
=== root@teevee (/dev/pts/11) /usr/ports/multimedia/ffmpeg 4 -> pkg install ffmpeg
And it was really only these two. Well, that's not so bad. I don't use palemoon, so there was only one package to reinstall. Did that, then:
=== root@teevee (/dev/pts/11) /usr/ports/multimedia/ffmpeg 5 -> pkg install firefox-esr
Will it never end? OK, Hugin and enblend aren't important on this system—I only installed them for test purposes. But why two enchants and two webkits? Still, things are broken enough as it is, so continue...
=== grog@teevee (/dev/pts/2) /spool/Images 14 -> firefox
=== grog@teevee (/dev/pts/0) /spool/Docco/Zeitreise 93 -> mpv Zeitreise---Die-Welt-im-Jahr-500-zeitreise-die-welt-im-jahre-500-100.mp4
=== grog@teevee (/dev/pts/0) /spool/Docco/Zeitreise 95 -> mplayer
=== grog@teevee (/dev/pts/0) /spool/Docco/Zeitreise 96 -> pkg search libdl
What's that? Clearly a library needed by at least three other libraries. But where is it? It's not in the ports. The locate database was older than the damage done, so it should have shown where the library was. And the only place was in the source tree! Tried copying it from there:
=== root@teevee (/dev/pts/11) /usr/ports/multimedia/ffmpeg 11 -> cp -p /home/obj/eureka/home/src/FreeBSD/svn/stable/11/lib/libdl/libdl.so.1 /usr/local/lib
But that didn't help much:
=== grog@teevee (/dev/pts/2) /spool/Images 26 -> chrome
=== grog@teevee (/dev/pts/2) /spool/Images 27 -> firefox
No browsers! The good news was that both mpv and mplayer worked. I was able to get firefox to run across the net (chromium is too polite for such things), so at least I could hobble. I'm reminded of the xkcd cartoon:
I should consider investigating snapshots before I do this again.
| Monday, 25 February 2019 | Dereel | Images for 25 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Upgrading teevee
|
Topic: technology, opinion | Link here |
After yesterday's catastrophe with teevee, the only logical way out was a complete system update: copy the entire root file system to the secondary one (which I call /destdir) and upgrade to the current status.
Before:
=== root@teevee (/dev/pts/3) /home/grog 13 -> df
After:
=== root@teevee (/dev/pts/3) /home/grog 13 -> df
This involves changing /boot/loader.conf on the first file system:
Why disk0 instead of disk1? I don't know either. It must be some BIOS/UEFI setup issue, but on some systems, like eureka, the system disk is disk1, and on others, like teevee, it proved to be disk0.
That went surprisingly smoothly modulo pkgNG's insistence on removing Emacs and BIND, clearly part of some conspiracy. But I was able to reinstall Emacs, and I don't need BIND on this system anyway—I could have reinstalled it if I did. In particular, my feared issues with the nvidia driver didn't eventuate.
|
The Emacs elimination conspiracy
|
Topic: technology, opinion | Link here |
I've already noted that pkgNG seems insistent on removing Emacs on every update. But that's trivial; I can reinstall it.
But maybe it's part of a larger conspiracy? I've been using Emacs and clones, along with other program such as bash, for nearly 40 years now, and it's ingrained in my cerebellum. Compared to other programs (including vi), it allows me to type without moving the balls of my hand from the desktop. But I'm beyond editor wars.
After about 12 years, this new-fangled thing called mosaic, a “web browser” came along. It was a bit clunky, but it allowed me to enter text and edit it minimally with Emacs-like key bindings.
But then Microsoft came along, and newer browsers came along. Fifteen years ago I installed firefox. It was hate at first sight, not only because it took away my beloved Emacs bindings. I found a workaround for that, though: create a file ~/gtkrc-2.0 and add the following entry:
gtk-key-theme-name = "Emacs"
As I noted at the time, though:
Not a big deal in the end; what surprises me is that nobody seems to care any more. Is this a sign of growing illiteracy?
With 15 years' hindsight, I think the answer is: yes. Now people don't even use keyboards any more. Who needs editing keys?
Time went on. Nine years ago, after a hardware-related crash, I installed Ubuntu on Yvonne's computer, and discovered that the gtkrc trick no longer worked. But that was just on Ubuntu, one of the reasons we didn't stick with it.
Still, they're working on it. 2½ years ago it stopped working again. And once again I found a workaround, firemacs, which actually worked quite well.
For a year or so. Then firefox decided to break the interface that firemacs uses, without replacement. I was forced to use an old version, firefox-esr, that still supports the interface that firemacs uses.
But after today's upgrade (to firefox-esr-60.5.1,1), that, too, was gone. And when I went looking, I discovered that I could no longer download it. Why? Where did it go? Why did it go?
I suppose I could probably install the version that I have on eureka, but I can't find it. Where is it located? It's still running there (and providing an excellent reason not to upgrade), but I can't find it.
More to the point, though: why do people break things like this? Yes, most people seem to use mobile phones with keyboard substitutes, producing documents that would have failed any elementary school English class, but that's no reason to try to make it more difficult for people who care. My current instantiation of firefox uses 4 GB of virtual memory, and it hasn't warmed up yet. My first Emacs-like editor, MINCE, ran in 48 kB of memory on a Z-80. The keymap is of trivial size. Why can't they put one in this overgrown behemoth that is firefox and let people configure their own keymaps?
For completeness' sake, yes, there are Emacs-like add-ons for firefox:
But it seems like, Firefox is allowing websites to suppress the bindings, which is super annoying when your bindings only work on a few sites :(
|
Dying Hebes
|
Topic: gardening, opinion | Link here |
Despite all my attempts, my hebes are not doing at all well:
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
|
||||||||||||||||||||||||
Part of it is the extreme heat and drought that we've had—so far this year we've had 12.3 mm, compared to 43.5 mm last year. And part is that my bore water filter keeps clogging up, maybe because there's not much water in 50 m depth either. I think we'll have to plant something else there, something more suited to Australian conditions.
| Tuesday, 26 February 2019 | Dereel → Ballarat → Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Emacs partially reinstated
|
Topic: technology, opinion | Link here |
More snooping around the web today looking for a way to reenable Emacs key bindings for web browsers. Is there maybe something for Google Chrome? Yes! And after looking carefully, it seems that I didn't look carefully yesterday.
The old, worn-out magic word was:
In days gone by, you put that in ~/gtkrc-2.0, and all was well. But that naming is clearly too Unix-like, so GTK 3 made it more tastefully Microsoft-like. I had seen that in this page, but not noticed the detail differences. They changed the name of the file from ~/gtkrc-2.0 to ~/.config/gtk-3.0/settings.ini and added these funny bracketed headings:
And how about that, now both firefox and chrome have limited Emacs key bindings. They're nothing like as complete as the firemacs support, unfortunately. In particular, c-s opens a “file save” window instead of initiating a search (how do you do that now?) and things like m-c and m-l don't work. c-y does something similar to Emacs, but not what I expected: it inserts something unrelated to the last deletion. Still. it's a partial pain relief.
|
More physiotherapy
|
Topic: health, opinion | Link here |
Into Ballarat today to see Heather Dalman, who gave me less than optimistic prognoses for what she's now calling hip bursitis, which, according to this page, could be trochanteric bursitis or iliopsoas bursitis, almost certainly the former. It looks like I'm going to have to learn to sleep on my back, something that I have not managed in the past.
|
Kitchen trolley
|
Topic: general, opinion | Link here |
We've had a couple of kitchen trolleys from ALDI, neither which lasted very well. The second was of better quality, but last month it lost a wheel:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
That wasn't just the poor quality of the components, though that helped: the wheel had got some dog hair caught in it and jammed:
|
|
|||||||||
I couldn't get it out, and since the axle is riveted in place, I can't dismantle it. So I did what I usually do in such cases: nothing.
And then I had an insight. One of the reasons that we bought these inferior ALDI things in the first place is because it's so difficult to find something. In fact, 28 years ago Yvonne bought one in Homberg an der Ohm. The wheels were tarnished, and I was a bit upset that she had bought it. In the end we removed the wheels and used it as a table, and it was still next to my armchair in the lounge room.
OK, how about some new wheels? While in town today, dropped into Bunnings and spent a good 20 minutes looking at possible wheels for both trolleys. Also bought some bolts to use instead of the rivets in the old wheel should I choose to repair it.
Back home screwed some of the wheels on the older trolley (the one from Homberg) with remarkably little difficulty and in much less time that it had taken me to choose them:
|
|||||||||||||||||||||||
It's stable and rolls well, but so far it still seems too low to me. We'll see how we get used to it.
| Wednesday, 27 February 2019 | Dereel | Images for 27 February 2019 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More gymnastics
|
Topic: health, opinion | Link here |
Off to the Dereel Hall again today for another round of gymnastics. I've never been keen on gymnastics, and things haven't changed. Is it doing me any good? I don't know. I didn't feel overly exercised at the end, and it's not clear that 45 minutes a week will make much difference one way or another.
One surprise: this is supposed to be for old folks, and I certainly qualify. But for some reason we had to sign in not only with our names, but also with our dates of birth and address. From that I read that I'm the second oldest (range of birth years from 1947 to 1969).
|
Next solar electricity quote
|
Topic: Stones Road house, general, opinion | Link here |
Finally a quote from Tomas Kucera of EffectiveElectrical today, in fact no less than four quotes. They certainly need a look at in some detail. What I see is:
So where do I go from here? Think it through more carefully, for one thing. Are the numbers that Tomas chosen chosen wisely? Much of the power consumption in the night could be heating or cooling, something that I don't always need solar power for. On the other hand, I want to get this sorted out.
|
Moving the fridge
|
Topic: general, opinion | Link here |
We've had the old fridge in the library for nearly a month, but it needs to go, especially as we're expecting overnight visitors in the near future. I had asked Chris Bahlo to help, but it occurred to me that we could try moving it by ourselves first, so that Chris didn't have as much to do.
Surprise, surprise! We got it back out into the hallway with no further problem, simply by turning it around the other way. And we got it out the front door by ourselves too, something that I didn't think we could do at all. It's about 4 cm less deep than the new fridge, which makes the difference. So Chris didn't need to come at all. A good thing, too: she had forgotten about the issue.
|
Bratwurst sous vide
|
Topic: food and drink, opinion | Link here |
Bratwurst for dinner tonight. That's surprisingly complicated: it needs nearly 20 minutes in the toaster oven, and cleaning the trays is a nightmare. How about an alternative? Sous-vide?. Have people tried this?
A web search brought me to this page, complete with tasteful visual overload suppression:
|
|
|||||||||
But once I got it to divulge its secrets, they were quite interesting.
And then there was another page with surprisingly similar content. When some of the wording repeated, it became clear that this was by the same person, so not that interesting after all.
Still, the ideas seemed to make sense: cook sous-vide for about an hour at temperatures ranging between 60° and 71.1° (bloody Fahrenheit!), and then sear. OK, worth a try, so I chose 65°. At the end they actually looked almost eatable as they were:
|
|||||||||||||||||||||||
And after searing they looked pretty much normal:
|
|||||||||||||||||||||||
Yvonne liked them and found them juicier than normal. I still found them a bit firm, and I'm wondering whether a lower temperature wouldn't be a better choice. Maybe 62° next time?
| Thursday, 28 February 2019 | Dereel | Images for 28 February 2019 |
| Top of page | ||
| previous day |
|
The horrors of revision control
|
Topic: technology, opinion | Link here |
Once upon a time revision control was simple: you used
a Revision Control
System. Of course, that was too easy to use simplistic: it didn't scale well
to multiple users or dependencies between files. So we moved
to CVS.
But there was something wrong with that, too, and so Subversion appeared, which we're still using in the FreeBSD project, and also Mercurial and a number of others that I have mercifully forgotten about.
In the meantime Larry McVoy had convinced Linus Torvalds to finally use BitKeeper—until he came up with some licensing issues that upset the Linux community. So Linus, ever one to name software after himself, wrote his own and called it Git.
For reasons I don't understand (because I haven't investigated them), many people in the FreeBSD community also started using git, though all that I know have complained about the complexity. But now somebody has come out with a man page generator that looks remarkably genuine.
|
Breaking Emacs, continued
|
Topic: technology | Link here |
Tried to start an Emacs on teevee today. Nothing happened. While I was pondering the issue, the window finally came up, accompanied by a message on the xterm:
What does that mean? But admire that modern yellow on white display.
|
Configuring X, 30 years on
|
Topic: technology, opinion | Link here |
It's been nearly 30 years since I started to use X. In the early days, setting it up wasn't easy, and I wrote one of the earliest documents trying to help users of free operating systems minimize the pain, summarized in a chapter in The Complete FreeBSD.
In the course of the decades, things have changed, mainly for the better. Trivial X configurations don't even need a config file; the server can work it out for itself. But now there are other problems. In the early days the issues were with the mode lines; images ended up all over the place. Since the advent of digital displays, that's all gone, but now we have issues with the display cards and drivers. I have a dual display nvidia card on eureso, but X didn't recognize the card correctly, and I ended up with a single monitor 1280x1024 display.
How do you fix that? Hard to say. Somewhere I read that “X -configure” is an old, worn-out magic word, and sure enough, it crashed on me—but only after creating a completely boring configuration file using the vesa driver.
OK, what do we have on teevee? Only one display, of course, but clearly the big difference is the driver name. Change it to nvidia and... it just works. Both monitors joined together as a single 3840x1080 display. I can live with that.
|
Ports pain
|
Topic: technology, opinion | Link here |
Hugin is coming out with its annual update, once again with a different naming scheme:
That's sloppy, of course: the first URL isn't the download URL at all. Point a browser at it and it offers to download https://excellmedia.dl.sourceforge.net/hugin-2019.0beta2.tar.bz. But you have to go there to find out. And the name is only marginally similar to the second URL, and both deviate significantly from the URLs for normal downloads, like https://cytranet.dl.sourceforge.net/project/hugin/hugin/hugin-2018.0/hugin-2018.0.0.tar.bz2 for the current version.
But that's not the way we write it. There are multiple places that you can get sourceforge sources from; there's also the alternative https://excellmedia.dl.sourceforge.net/project/hugin/hugin/hugin-2018.0/hugin-2018.0.0.tar.bz2 and many more. So the Ports Collection has a description format:
PORTNAME= hugin
PORTVERSION= 2018.0.0
MASTER_SITES= SF/${PORTNAME}/${PORTNAME}/${PORTNAME}-${PORTVERSION:R}
How do you change this to work for the betas? Every time I try, I end up with my eyes going funny. Count those hugins. What ending to the name? Where do you write beta? It's just too complicated.
Today I did what I've done before: download the file and link it to a name that the Ports Collection expects (in this case, hugin-2019.0.0.tar.bz2 instead of hugin-2019.0beta2.tar.bz; note the different ending. This also works around a second issue: the Ports Collection expects the top level directory name to match the package name, so if I had bent the Makefile to point to hugin-2019.0beta2.tar.bz, it would expect to find the extracted sources in work/hugin-2019.0beta2. But they're really in work/hugin-2019.0.0. Yes, there's a twisty little Makefile variable to handle that, but after looking through the documentation I couldn't find the correct way to do it. Since this is just for me, this is just plain easier.
Of course, some things could do with improvements. My first attempt failed like this:
(Many lines building ninja)
===> Checking if ninja is already installed
===> Registering installation for ninja-1.8.2_1,2 as automatic
Installing ninja-1.8.2_1,2...
===> hugin-2019.0.0 depends on executable: ninja - found
===> Returning to build of hugin-2019.0.0
===> hugin-2019.0.0 depends on executable: update-desktop-database - found
===> hugin-2019.0.0 depends on executable: msgfmt - found
===> hugin-2019.0.0 depends on package: pkgconf>=1.3.0_1 - found
===> hugin-2019.0.0 depends on executable: update-mime-database - found
===> hugin-2019.0.0 depends on file: /usr/local/bin/python2.7 - found
===> hugin-2019.0.0 depends on executable: gtk-update-icon-cache - found
Errors with dependencies.
*** Error code 1
What went wrong there? It checked for a dependency ninja, found that it wasn't installed, installed it, and continued. Check for more dependencies, finds them, and then fails for no obvious reason. It wasn't until the second attempt that I understood:
===> hugin-2019.0.0 depends on executable: swig2.0 - not found
Error a dependency refers to a non existing origin: /eureka/home/src/FreeBSD/svn/ports/devel/swig20 in BUILD_DEPENDS
===> hugin-2019.0.0 depends on file: /usr/local/bin/cmake - found
===> hugin-2019.0.0 depends on executable: ninja - found
===> hugin-2019.0.0 depends on executable: update-desktop-database - found
===> hugin-2019.0.0 depends on executable: msgfmt - found
===> hugin-2019.0.0 depends on package: pkgconf>=1.3.0_1 - found
===> hugin-2019.0.0 depends on executable: update-mime-database - found
===> hugin-2019.0.0 depends on file: /usr/local/bin/python2.7 - found
===> hugin-2019.0.0 depends on executable: gtk-update-icon-cache - found
Errors with dependencies.
*** Error code 1
The error related to swig2.0 (savour that two-digit appendage, now 3.0). It didn't find it, didn't fail, but generated copious output building ninja and blowing away the error message. How could we do that better?
Apart from that, to my surprise, it Just Built! Finally they have attended to this bug that caused in-tree builds to fail. Just for FreeBSD? I don't want to look.
Of course, we're not done yet. I still need to rebuild the pkg-plist. Somehow I don't have the stamina I once did.
|
Bird photos: the pain
|
Topic: photography, animals | Link here |
We've had large swarms of Sulphur-crested cockatoos flying around the area lately. While walking the dogs, a number landed in the gum trees in front of me:
|
|||||||||||||||||||||||
Where are there cockatoos? Nowhere. Well, they're there, but barely visible. It wasn't until I processed them that I discovered that they were Corellas, which look like cockies with a hangover:
|
|
|||||||||
Of course, my photos serve only one good purpose, as a bad example.
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 |
