|
| Greg's diary |
| July 2013 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Monday, 1 July 2013 | Dereel | |
| Top of page | ||
| next day | ||
| last day |
|
^X responds
|
Topic: language, opinion | Link here |
Not surprisingly, ^X, whose name proves to be Justin Smith, didn't agree with yesterday's article. If I understand correctly, he had expected me to purchase and read all the books to which he partially referred yesterday before coming to any conclusion. Callum Gibson suggested that I at least look at the YouTube reference, which proved to be a debunking of a specific theory of the origins of the Indo-Europeans. Just downloading it was difficult enough: it's 700 MB in size, and in the middle of it my network connection dropped. So I reloaded it on freefall and then copied the rest with rsync:
=== grog@freefall (/dev/pts/20) ~/youtube 5 -> bin/youtube-dl -tw 4jHsy4xeuoQ
16.1 MB/s! That's faster than the NBN, even faster than my home network.
The lecture is very interesting. It's over an hour in length, and by the time I finally got it on site, it was too late to watch it all. But in summary it rejects the theory because it goes only by comparison of vocabulary, not grammar, and doesn't account for borrowings, thus coming up with demonstrably incorrect relationships between languages. The theory then uses these relationships to deduce the origins of the Indo-Europeans. The lecture shows numerous amusingly incorrect claims about the Indo-Europeans; Justin's would fit in well. And it gives a number of examples of Indo-Aryan languages, which it accepts without comment.
This is my big problem with Justin's arguments: he presents quotations without explaining how they should support his views. The two I have seen so far have, not surprisingly, refuted his views, either explicitly or implicitly. I have little doubt that the others will as well.
He also took me to task for his anti-Muslim views. But there we have logs. From 17 March 2013 it appears that maybe his hatred is directed towards the Arabs, not the Muslims, though in other places he spoke against Malay Muslims:
I think the end of that discussion was lost on him.
One concrete thing came out of the discussion, though: he may have some extremist views, but they're not associated with an organized religion. He claims to be an atheist. I still have no idea what motivates him. By comparison Wendy McClelland is an open book.
It's also clear that, despite his engagement, he either didn't read my article, didn't understand it, or chose to ignore things. I recall some reprise of his argument on word order, but I can't find it in the (voluminous) logs. He's also hazy on a number of things that you'd expect him to know: he confuses Malay with Malayalam, and he thinks that the Himalayas are proof that the Indo-Europeans couldn't have made it from Europe to India. He drags in the Turks, who have nothing to do with the matter, and also managed to get South-East Asia involved:
He asked me to not edit a comment he planned to leave on my blog, until others pointed out that the article was in my diary, which has, at the bottom:
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. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details.
But clearly it's unfair not to give him a chance to respond, and so he finally sent in a mail message written in quoted-printable with a Microsoft character set. He explicitly requested me to leave it in this form. It contains more more meandering past things we had already addressed. Interestingly, the time zone was UTC+5:30, which is India, not Singapore, though in subsequent discussion he claims to be in an unspecified San Jose, possibly California.
His references to the Qur'an are, like others, unfathomable. The first (Al-Baqara 65) reads:
And you had already known about those who transgressed among you concerning the sabbath, and We said to them, "Be apes, despised."
I wonder what he's trying to say with that. But who cares? My experiment yesterday was to try to convince him. I failed spectacularly. What I have gained from it is a better understanding of how somebody can cling to an incorrect belief despite overwhelming evidence to the contrary, even to the extent of claiming that that evidence supports his views.
| Tuesday, 2 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Still more language arguments
|
Topic: language, opinion | Link here |
I had really expected the discussion about the origins of Sanskrit to be over yesterday, but no such luck. Justin disagrees violently but somewhat incoherently with what I wrote, and wants me to retract it. I've noted that he's an atheist, but I see no reason to change my opinion.
But there was one thing: he claims that the YouTube lecture supports his claim that Sanskrit is not an Indo-European language. That's easy enough to verify: I sent a message to the authors asking, as neutrally as I could, if that were the case, and got a reply from Asya Pereltsvaig:
Regarding your question about Sanskrit, it's beyond any doubt at this point (and for some two hundred years) that it is an Indo-European language. Whoever your opponent is, he (she?) clearly has an axe to grind, maybe more than one. It is also clear that this person is not interested in reasoned debate or unbiased research. Thus, no matter what you tell him, it would be the same nonsense he would continue to spout. From my own experience, it is best not to waste too much of one's time on such people and their crazy ideas. The one thing I do want to stress from reading your debate is that one has to separate races/genes from languages. The two do not necessarily coincide, but this point is often missed completely by the likes of your opponent.
So basically a complete refutation of Justin's claims, going beyond my own question. Yesterday he had written:
So I put Asya's reply to Justin and got a somewhat unexpected response:
At this point you might doubt Justin's sanity. It seems more than evident that he either doesn't understand what's going on, or that he tries to change his tune to fit the current situation. No wonder he jumps from one topic to another. His presentation is also so vague that it's not clear what he's trying to say. In that last line he was in fact quoting Peter Jeremy from a couple of days ago, but he didn't say so, and he didn't give an explanation for the quote, which, I think, he still doesn't support. But he repeated it later on, giving the impression that he does now believe it. Who cares?
He spent the afternoon throwing invective such as:
Still, after some time he did come up with some references that at first sight appear support his ideas. This article distances Aryans from Europeans, which is valid. But the article handles a matter of national identity, a bad thing to mix with history. India didn't exist 3000 years ago. Neither did Europe, and certainly not Germany, a country first formed 142 years ago. But the author writes:
Since there is no conclusive evidence to support the theory of an Aryan invasion/migration into India, and on the contrary, there is compelling evidence to refute it; and since the theory seriously damages the integrity of the Hindu tradition and its connection to India; we call for a serious reconsideration of this theory, and a revision of all educational material on this issue that includes the most recent and reliable scholarship.
He leaves out the compelling evidence. But although the Hindu tradition really goes go back to the Vedas, suppressing history has never worked well. Advocating its suppression based on potential damage to modern Hinduism is stupid. More to the point, though, the article doesn't mention the term “Indo-European” at all. But it's easy to see how this kind of article could have confused Justin.
On a purely linguistic level, the article is also flawed. It claims that the term arya was never used to identify people. The people of Iran would be horrified. I pointed Justin at it, who typically ignored the reference to the name Aryan and picked on the name Persia. On the other hand, he then came up with this link which, once again, seems to refute all his claims.
He gave me a number of other links, such as this one, none of which make any claims that Sanskrit is not an Indo-European language. This particular article considers the possibility that the original Indo-Europeans all came from India. While that seems unlikely, it's certainly not beyond the bounds of possibility. But like just about all these links, the emphasis is on national identity, seen from a modern perspective. That's the same kind of attitude that in more extreme form gave rise to the Nazi concept of blond-haired, blue-eyed “Aryans”.
To support his assertions, Elst claims that the Vedas were written about 3000 BC, before the first known horse-drawn chariots, dating to about 2000 BC, were found in the Ural mountains. This appears necessary to explain the arguments: consensus is that the Vedas were written somewhere in the time frame 1500 BC to 500 BC. But all of this is tangential to the central theme that Sanskrit is undoubtedly an Indo-European language.
|
BBC said it! It must be right!
|
Topic: language, opinion | Link here |
People like Koenraad Elst are not without their critics, but then there are reputable news service, like the BBC, who published an article about the Mapping the origin of Indo-European article that Lewis and Pereltsvaig debunked in their YouTube lecture. The title “English language 'originated in Turkey'” sets the tone. It goes into more detail of Phylogenetics than language. To be fair to the BBC, they apparently had the agreement of Professor Mark Pagel, who sees the study as being conclusive. It's not clear to what extent he was involved in the study—he has co-authored another study with the main author of the Anatolian paper—but if such people are so easily satisfied, it's understandable that people like Justin Smith can be misled.
|
Baskets for the freezer
|
Topic: general, opinion | Link here |
One of the most stupid things about our new freezer is that it only has two baskets:
I don't know why people do this. It's almost impossible to find things in a fridge like that, and it almost invariably requires removing everything on a shelf. But none of the upright freezers we saw have more than one or two baskets. We accepted the inevitable and decided to buy baskets separately.
Not easy, of course. It would be nice to find wire mesh baskets with adjustable size, but all Yvonne could find were standard polypropylene boxes, which are designed to be stacked and covered and thus have slanting walls and a ridge at the top, which greatly reduces the space available:
|
||||||||||||||||||||
So she called up the manufacturer's representative. Yes, we can buy more of either of the two baskets supplied with the fridge (they're of different size). The bigger one, about 10 cm too shallow for the top shelves, would cost $193 each, and we'd need 4 of them—$772 in total, and that for a fridge that cost $1169! Somehow we'll find a better solution.
|
Learning Python, again
|
Topic: technology, opinion | Link here |
I never finished the computational photography course that I started a couple of months ago. I had started in mid-course, and it became apparent that both my linear algebra and python skills were lacking. I started a Python course a little later, but it was too elementary, so I gave up on that too. Now they're offering a course on linear algebra, coincidentally using Python, so I've enrolled in that.
First issue: it requires python 3.3.2, but the version installed on my machine is 2.7.2. They're not compatible (bad python!), so rather than remove the installed version from eureka, my main machine, I did it on teevee, the TV machine. And to my astonishment there were several hundred dependencies—but make deinstall removed it anyway. Fortunately I was able to reinstall version 2.7.5, which doesn't seem to have any compatibility issues—and then I discovered that the two versions can co-exist.
Spent the afternoon going through the introduction to python, which took me a lot further than the python course had done in two weeks. Once again I'm Just Plain Puzzled by python—it seems so different from other languages. The conditional expressions seem just plain bizarre, but it's easier to look at some of the constructs from a LISP perspective. Comprehensions remind me of re-ordered lambda expressions. And then there are things that don't get explained enough, at least here: which objects are passed by value, which by reference? Numbers belong to the former, sets to the latter. But strings? I haven't got that far, but this look-ahead makes it almost more difficult for me than for a beginner.
| Wednesday, 3 July 2013 | Dereel | Images for 3 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Hindu nationalism and linguistic history
|
Topic: language, opinion | Link here |
I've been investigating Justin Smith's opinions on Sanskrit for some days now, and really it's over and done with. But I'm interested in the topic, and it has opened a lot of new avenues of investigation. One rather surprising one was that Justin today finally gave me an indirect pointer via a review of an article which has since been removed. The second link provides excerpts that not only support his views, but probably formed them. It's not clear why he didn't come up with this earlier. Justin seems to think that the first article agrees, though it's clear that it does not.
The review page is sheer unmitigated nonsense. It starts off by discussing the pain that Hungarians feel because Magyar has not gained entry to the Indo-European language club. It goes on (all original spelling):
Building primitive lexicons that show similar roots for certain common words can hardly be an adequate basis of linguistic classification. Especially if that classification is going to be further used to generate implications about sociological and cultural development. If the commonality between Indian and European langauages extends only to a small pastoral-era oral lexicon, the Indo-European theory of langauges could hardly be called in to justify the “Aryan Invasion” theory let alone infer that the Vedas were written by “Indo-European Aryan” migrants
He has carefully omitted grammar, in particular inflection, here. His real point seems to be:
In fact, one of the unintended (or even intended) consequences of such linguistic speculation is that there has been a needless intellectual division between North Indians and South Indians,
Why should that be? And what relevance does that have? To Shishir Thadani, it seems to be the central matter. He goes on to analyse languages based on writing systems that were imposed later, such as Devanagari and Tamil script:
And although linguists are divided as to which came first, both Sanskrit and Tamil are written in very similar ways. Unlike the European langauges that are written using alphabets (derived from Greek, and branching off from Latin or Cyrillic), all Indian languages are written using syllables made up of (simple or compound) consonant shapes that are modified by the symbols for vowels that connect the consonants.
That's completely irrelevant to the language itself. How can anybody take it seriously? It would imply that nobody could speak if he couldn't first write.
He then goes on to say:
It may also be noted that across India, both Sanskrit and Tamil derived languages use SOV (subject Object Verb) word order as a default. But several Indo-European langauges such as English, French, Portugese and Bulgarian use SVO word order.
At the end the reviewer states that Sh Thadani was ”assisted in his reasearch[sic] by Giti Thadani, who is intimately familiar with several European langauges including German, French and Hungarian (as well as Sanskrit)”. If that's the case, then the statements are deliberately misleading. Even modern Hindi is closer to, say, German than Magyar (“Hungarian”) is, and as I've already mentioned, German frequently mandates verbs at the end of the sentence. Once again I'm left with the impression that this kind of article is deliberately misleading.
Reading about the Out of India theory confirms what I've been thinking all along. The whole business is primarily a misguided attempt to promote Hindu nationalism. I have no objection to the Hindu faith, but I object to extremists wherever they occur.
|
Comprehending python
|
Topic: technology, opinion | Link here |
Most language courses are boringly simplistic. The one I'm going through for Python is not. One of the issues, of course, is the lack of description of the syntax, particularly since it's so baroque. But mainly the issue is that it requires a completely different approach to programming from what I've seen before. I hope it gets easier once I have accepted the basics.
|
US Government Morals
|
Topic: general, opinion | Link here |
Doubtless the recent revelations of Edward Snowden have shocked the world, including most US citizens. But somehow it seems to be only the tip of the iceberg. Where is Snowden? Still in Moscow airport? It seems that yesterday a number of people believed him to be in the private jet of Evo Morales, the President of Bolivia. And a number of European countries refused him passage through their air space.
The reports are varied and garbled. A number of sources report (here from the Washington Post):
Choquehuanca said in a statement that after France and Portugal canceled authorization for the flight, Spain’s government allowed the plane to be refueled in its territory. From there the Falcon plane flew on to Vienna.
But why return from Spain to Vienna? Have people looked at a map? As this map shows, to get from Spain (in fact, as reported elsewhere, the Canary Islands) to Bolivia you don't need to go through Portuguese or French airspace:
|
|
||||||||
It makes sense for them to have planned to refuel in Lisbon, but Fuertaventura would in fact be a better choice. But if they're there, why on earth return to Vienna, especially as other reports claim that Italy also cancelled authorization to fly through their air space. That route is in exactly the wrong direction, and it crosses both French and Italian air space:
|
|
||||||||
Why go there, especially since getting there without crossing Italian or French air space would involve flying south past all of Italy, a significant detour?
My best guess is that the reporters got it all wrong and didn't have enough insight to notice that it couldn't be correct. In all likelihood, the plane never got further from Moscow than Vienna. Spain would have allowed refuelling where Portugal didn't, but the problem was getting there in the first place—the issue with French and Italian air space in the opposite direction. That's confirmed by this report. François Hollande apologized with the explanation that they were unaware that Morales was on board. Believe that if you will, and then ask why it should make any difference.
But how plausible is the story in the first place? It's clear that the reports contradict each other, and France, of all places, is not known for bowing to US pressure. Indeed the second article above mentions that France wants to postpone talks with the US until they explain allegations of US espionage in European diplomatic missions. And even Russia was not prepared to grant asylum to Snowden unless he promised to stop revealing state secrets. It's difficult to believe that such countries would take such an extraordinary step without extreme influence of the USA.
So, back to the USA, that bastion of freedom and law. What has their government done to maintain this status?
They maintain an tropical gulag on Cuba (the words of George W. Bush, though I suspect he was confused about what he meant. Amnesty International wasn't, though their web content seems to die quickly, and I can no longer access it).
They have violated their own laws by spying on their citizens, in violation of their own constitution.
They have apparently spied on allied countries, in violation of international law. This report comes from the Spiegel, in my mind hardly the most trustworthy magazine, but clearly the European heads of state are taking the allegations seriously enough.
They persecute people who bring this sorry state of affairs to light. Yes, I think that Snowden could have done things differently, but the reaction is out of all proportion.
It seems clear to me that the US government must have applied unimaginable force behind the scenes if both Russia and France bowed to it.
On a completely unrelated note, but equally concerning, they continue to violate the territorial integrity of Pakistan with drones. In the news today more civilians were killed in a US attack targeting the Taliban. While I agree that the Taliban are a worthy target, it doesn't justify breaking international law to do so.
What does that have to do with the USA? Well, the people of the US are some of the victims. But clearly the government is not sticking to its principles. It's turning a bastion of truth, honesty and freedom into a global bully. Shame on them.
|
Biff à la Lindström
|
Topic: food and drink | Link here |
Why do the Swedes like meat balls so much? They're so popular that the Wikipedia page on Swedish cuisine shows them as the main representative. Went looking for a recipe today and discovered there's really nothing very special about them. But I did find a modified version, biff à la Lindström, which is certainly different enough to be interesting. I'll probably need a second pass to get it right.
| Thursday, 4 July 2013 | Dereel | Images for 4 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More on Morales story
|
Topic: general, opinion | Link here |
Spent some time today investigating the contradictions in the story of Evo Morales' forced stopover in Vienna. The results are interesting for a number of reasons. Firstly, as I suspected, the plane didn't leave Austrian air space, as this image shows:
But why did they turn back? Yes, they were approaching Italian air space. But according to this report the plane didn't have enough fuel to get to the Canary Islands, some distance further than Lisbon. The plane is a Dassault Falcon 900EX, which has a range of 4,500 nautical miles. From Moscow to Fuerteventura is 2,730 nautical miles, well within the range of the plane. But it seems that over Austria they developed problems with the fuel indicator, as this transcript of the exchange with Schwechat air traffic control shows. Under those circumstances, it's perfectly normal to land at the nearest airport.
And the Austrians (obviously) didn't refuse entry to their air space. So why should they search the plane for Snowden? Did they search the plane for Snowden? Reports there are very vague as well.
All in all, the more I look at this matter, the less plausible it seems. Yes, no question, Morales made an unscheduled landing at Schwechat. And there were some issues with authorization to enter French air space, something that he hardly needed anyway. The Great Circle barely cuts through the south of France:
|
|
||||||||
But it also goes north of Austria, while the real flight path went over Austria. Possibly this path already indicated corrections made because of France's refusal. Portugal, it seems, had refused entry some days earlier for “technical reasons” that could hardly have had anything to do with Snowden, which is why Morales' team had planned to refuel on the Canary Islands.
On the other hand, it's clear that France did initially refuse entry. Why? They don't say. But that doesn't seem to have been the reason for the landing in Schwechat. And Italy? There's no confirmation that they did refuse entry, and the facts suggest that they didn't. The truth may come out, but I'm no longer as convinced of US involvement as I was yesterday. There are silly reports like “trying to kidnap“ Morales, which, it seems, he started. And so far I haven't seen any report linking the affair to Edward Snowden.
On the other hand, the research has brought out some interesting web sites. This report includes a link to a map which, at some point, shows the path of the flight, though I haven't worked out how to control it yet. And this page discusses some of the points I raised yesterday, though better.
And if that isn't enough, this report claims that the Embassy of Ecuador in London—current home of Julian Assange—has been bugged, and that they had discovered it a couple of weeks ago and not reported it. Somehow, though the whole matter horrified me, I'm gradually left wondering how much truth there is in the allegations.
| Friday, 5 July 2013 | Dereel | Images for 5 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
News media accuracy
|
Topic: general, opinion | Link here |
This matter with Evo Morales' unplanned stopover in Wien gets stranger and stranger. Gradually the news media are beginning to ask obvious questions. Strangely, the UK Daily Mail—not a newspaper I would turn to for in-depth coverage—has asked a surprisingly detailed report with a number of pertinent questions. But there are many more. Here's what I have so far:
It seems certain that France initially denied passage over its airspace, but why? And for how long?
Did Morales' flight take off before it had clearance to cross all necessary air space before landing? That's an absolute requirement, so it seems difficult to believe that it didn't. And in particular the flight path suggests that it was avoiding France.
Did Portugal deny access to their air space? The only statements with any backing are that they were unable to refuel the flight due to unknown technical reasons, and that Morales' flight crew had been informed days in advance.
Did Spain refuse access to its air space? Some maintain this, but it does not seem to make sense.
Did Italy refuse passage? Again, there are claims, but no official statement. It's interesting that both Spain and Italy were not mentioned until long after the initial reports.
What role did Austria play? If anybody diverted the plane, it would be the Austrians, and they're the only people who—maybe—searched the plane. But nobody's blaming them for anything.
What role did Snowden play? The connections with him are remarkably vague.
What role did mechanical problems play? In the transcript of the exchange with Schwechat air traffic control there's talk of problems with the fuel gauge. Is the transcript genuine? So far nobody has claimed it's not.
This report suggests that Bolivia claimed that all four countries had refused access. Presumably this will all come to light when the UN handles the complaint. But it's becoming increasingly clear that the reporting of this incident leaves much to be desired.
How about this non-conspiracy-theory version: Edo Morales sets off home from Moscow. Back in Bolivia, people discover that he has been refused access to French air space, and that his planned refuelling in Lisbon had been changed to Fuerteventura due to unspecified “technical reasons”. They and start to wonder why, and note that Morales had been criticized for showing a preparedness to grant asylum to Snowden, and jump to conclusions. Then the plane runs into (minor) technical problems and chooses to land in Vienna to have the matter seen to, something that can only be done the following day. On landing at Vienna, the authorities visit the plane and check the identity of the passengers. Austrian media are quick on the scene and put their own spin on the story. The plane is repaired and continues to Bolivia via Fuerteventura and somewhere in Brazil. The whole time Morales is informed only by his staff, so he doesn't know of the incorrect assumptions.
Not nearly as newsworthy, is it?
|
Bindy visits
|
Topic: animals, general | Link here |
Zali from Enfield is spending the night in Melbourne and was looking for somebody to look after her Borzoi bitch Bindy, who happens to be Zhivago's daughter, so she asked Yvonne to look after her for the night. That's quite convenient for us for a couple of reasons: firstly, we were about to ask Zali to look after Zhivago while we went to Melbourne (though only for the day), and secondly we were interested to see how we would get on with a younger, more lively dog: Zhivago is over 7, and Bindy is only 20 months old.
She arrived, a little nervous, and then she saw Piccola. After her in a flash. After I caught Bindy and told her off, we found Piccola on a filing cabinet in Yvonne's office, tail three times the normal thickness. But later I showed Lilac to Bindy, and Bindy behaved herself. By the evening things were looking much better:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
In general, she's gentle and well-behaved, just like we remember Borzois. But it's good to have confirmation.
| Saturday, 6 July 2013 | Dereel | Images for 6 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Flight routes and German word games
|
Topic: general, opinion, language | Link here |
I've been looking more at http://www.flightradar24.com/, which is a surprising mixture of very useful information and frustrating user interface. Looking at flight traffic in the Near East shows some remarkable patterns:
|
|
||||||||
I'm reminded of the German word play „Wenn hinter Fliegen Fliegen fliegen, fliegen Fliegen Fliegen nach“, clearly chosen because of the repetition. It means “when behind flies flies fly, fly flies flies after”. Not a very clever choice. But German has a slang word for plane: Flieger (flyer). So here it's easy to confirm that „Wenn hinter Fliegern Flieger fliegen, fliegen Flieger Fliegern nach“
Apart from that, it seems to be possible to track specific flights, but not to get a historical view of traffic at a particular time: there's no way to enter date or time, and even when it shows traffic from older flights, it doesn't display the date. And I haven't found any real instructions. Maybe things will improve.
|
Your very first puppy
|
Topic: animals, opinion | Link here |
There was a TV programme on SC10 TV this afternoon: “Your very first puppy”. We're not exactly the target audience, but we thought it would be interesting to watch. It was, but not for the reasons I had expected. It went through the examples of three different puppies and the people who bought them. In every case the dogs misbehaved badly.
What do you do about misbehaving dogs? If your German Shepherd Dog jumps onto the table while you're eating, it seems that you buy a compressed air horn and blow it at him. If your puppy chews things, you don't remove then: you put up with it. Nowhere in the 1 hour programme did I hear the advice to say “no”.
And feeding? Dry food is good enough. Never mind that it causes the teeth to rot and the resultant stools to stink, nor that it's more expensive than healthy food: they didn't even mention that alternative.
And the cost of keeping dogs, apart from replacing chewed-up items?
|
||||||||||||||||||||||
No harm in overestimating a little, I suppose: far too many people get animals and then dump them because of the unexpected costs. But $1200 for food seemed far too high. And our vet costs are worm tablets and vaccinations, about $100 a year. And what are “Extras”, and why is the number so precise? OK, they recommend dry dog food, presumably because the manufacturers want them to, whereas we feed cheaper natural fresh food, but that wouldn't explain the difference. Afterwards I checked what we spent last year for Lilac, Piccola, Nemo and Zhivago: $1,779.76, almost exactly two thirds of that sum. That's only one dog (Zhivago replaced Nemo on the same day), but there are still 3 sets of vet fees, and both Zhivago and Nemo are big, hungry dogs. Where do they get their numbers from?
|
More weather station problems
|
Topic: technology, opinion | Link here |
My weather station has never been very reliable, and over the course of time I've been putting more and more heuristics into my software to catch the more obvious errors, most recently three months ago. But once again it seems to be getting cleverer: it's generating less obvious errors, and I can't catch them. Do I care? Yes, but not enough to drop everything and think out Yet Another Way of catching errors. Why didn't they just put a checksum in the transmitted data?
|
Beer storage times
|
Topic: brewing, opinion | Link here |
One of the things that came to light while rearranging the fridge was a number of bottles of commercial beer, all old. Throw them away? It makes sense to try them first.
The first was a bottle of Pilsner Urquell, which I bought in Canberra round 20 April 2005, and which I recall not having tasted any good at the time. It didn't taste good now either, and it was far too dark. But maybe that was the way it was at the time.
The other was a Trumer Pilsener with a best-by date of 5 May 2005. Trumer is in Salzburg, and I visited a couple of breweries there with Hubert Hanghofer on 26 October 2004, so it seemed reasonable that I bought the beer at the time. But no, the label was Australian. In any case, this beer tasted good, nice and malty, just nothing like a Pilsener. But maybe that's the way it's supposed to be. In any case, it shows that, at least with cooling, a good beer can keep for a very long time.
|
State of the art web infrastructure
|
Topic: technology, opinion | Link here |
Chris Bahlo along for dinner tonight, as usual on Saturdays. After dinner, while Yvonne went to sleep with boredom, we talked about her new job at a local web design company whose name I forgot to ask. We discussed again my incomprehension that Wordsworth had taken four days to move the Friends of the Ballarat Botanical Gardens web site from the existing, functional site to the new site—why couldn't they just have cut over the DNS when it was up and running? Chris said “updating a page on our sites normally doesn't cause more than two minutes lack of access”.
WHAT? Why that? Chris said that it was acceptable. I said that there's no reason for any inaccessibility (well, not more than the time it takes to rename two files). With rsync you can not only cut over seamlessly, but also ensure that you don't forget any files (though more than one could introduce a seam). But they copy files manually.
After I calmed down, I discovered that wasn't the only issue. They do their web design by maintaining a local copy of the site—nothing wrong with that—but it seems that their Content Management System (Joomla) doesn't do what I expected. In particular, there's no form of revision control. The only way to avoid data loss is to physically ensure that no two people work in the same area at the same time. This, too, they consider acceptable. If they need to roll back a change, they take the manual backup that they had made.
This is a problem that was solved decades ago! SCCS is over 40 years old. When I started working at my first real job, on 2 May 1973, the UNIVAC 1108 system had a file format, ELT, which incorporated up to 63 versions of a file. It must have been introduced some time round 1968. And now people are still doing this by hand?
I wonder if this is an issue specific to Chris' employer, or whether most web design companies work like this. Clearly I need to understand what CMS systems do. I had assumed that they would at the very least take care of revision control. But probably there's no reason not to use revision control in addition to content management.
| Sunday, 7 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Python learning notes
|
Topic: technology, opinion | Link here |
I'm continuing with this supposed linear algebra course, though so far I've only been learning python, and even the exercises aren't obviously related to linear algebra. I've signed a declaration of honour that I won't tell people about it—designed to ensure that people don't copy other people's results. So I can't give too much detail, but the current exercise is to build a reverse index for a text search engine.
How do you do that? The assignment documentation gives just enough information for you to be able to infer what they mean. And finally they've admitted that python has normal iteration constructs, and not just comprehensions. Spent some time doing the main exercise, building the index itself. In the process discovered a number of non-obvious things:
{1, 2} is a set. {1} is a set. {} is a dictionary. To create an empty set, you need to use the constructor (if that's what they call it) set().
Indentation determines block structure. That's a thing I have known for a long time, and one of the main reasons I have held off looking at it more closely. But how do you end a block? Outdent, of course—most of the time. But it seems that in some cases that's not enough, and you need an empty line as well. Doubtless that's in the documentation somewhere, and maybe my interpretation is inaccurate. But it confirms my prejudice that relying on indentation is bad.
My index implementation worked. But what about performance? Running it against the large collection of data supplied with the assignment used up 140 minutes of CPU time. Is that acceptable as a solution for the task? I don't know, but it's not acceptable to me. Once again I'm looking not at the algorithm, but at the constraints of the programming language.
Decades ago, round mid-1975, one of my colleagues at IBAT in Essen told us about the eight queens puzzle: Place 8 queens on a chess board so that they can't attack each other. He promptly set to and wrote a program in BASIC, which he ran on our Dietz 621 computer, not the world's fastest. After about 3 hours he had the first result, and it looked as if the whole thing would take weeks.
But Real Programmers use assembler, so I implemented the same algorithm in assembler. Faster. It only took (many) minutes for the first result. Still not good enough.
The problem was that the algorithm was iterative, and it really would have gone through 2⁶⁴ iterations. What we needed was recursion. But what machines had stacks? The legendary PDP-11, of course, but where could we get one of those? So I simulated a stack in assembler, ran the program, and it output the results as fast as the (2400 bps) terminal could display them. Clearly I had made a mistake.
Of course, I hadn't. That was the difference in speed of the algorithms. For some time to come I wrote the same algorithm in different languages to compare speeds. I last ran the latest version in early 1996, under BSD/OS, and to get any useful timing I had to run it 100 times. Bringing it up to date today gave me a run time of 0.035909 s, including outputting all the results. How times change!
Getting back to the reverse index problem: somehow I have the same feeling, that there's some elegant python solution that's not obvious to mere C programmers. It's interesting to notice that there's a solution to the queens problem written in python on the Wikipedia page. It also seems to be typical of languages that keep changing their syntax that the solution didn't work on python 3. I first had to fix it:
|
News reporting continued
|
Topic: general, opinion | Link here |
Today Asiana airlines flight OZ214 crash landed at San Francisco International Airport. Fortunately most of the passengers survived. Apart from the news coverage, it was a good test of flightradar24. It didn't do well. It did come up with a plausible time—16:43—and on at least one occasion it came up with information on the flight, but I never got to see the flight path. My guess is that it's a timing-sensitive Java application, and it doesn't handle my network connection well. If they ever build the radiation tower (should have started a couple of weeks ago, but they haven't yet), things should improve. About the only information was the URL of the playback: http://www.flightradar24.com/2013-07-07/10:05/AAR2144#!/2013-07-06/16:43/AAR214. It seems that AAR2144 and AAR214 are alternative flight numbers, and the date and time are clear. So maybe I can get more information from that.
And the news coverage was correspondingly flaky. No mention of possible causes. The only coverage I saw with any insight at all was Al Jazeera, where the newsreader asked the ex-chief of the NTSB if the issue was similar to the crash of British Airways Flight 38. And indeed it was: this plane was also a Boeing 777, and it crashed after a long, uneventful international flight by coming down too short before the runway. That incident was traced to the engines (Rolls-Royce Trent 895), but so far there has been no information either way about the engines of OZ214.
|
Room temperature distribution
|
Topic: general, opinion | Link here |
It's cold at this time of year, of course, and we're trying to save energy, so we're keeping the room temperature (measured on the wall 1.5 m above the ground) below 23°. But it feels really cold! This evening I checked: at 1.5 m, 23.4°. At head height in the armchair, about 20°. At foot height, 17°. What a pain these badly insulated houses are!
| Monday, 8 July 2013 | Dereel → Ballarat → Dereel | Images for 8 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
New carpets
|
Topic: general | Link here |
One of the consequences of the water bed leak is that the carpet in the room is both sodden and exposed. It's also old and dirty—perfect time to replace it. We've done some pricing at Bunnings, who charge a flat fee for installing any amount of carpet. But this is only about 12 m², hardly worth the cost.
It's not as if the carpets in the rest of the house are in the best of shape. So we considered replacing more. The biggest issue is that there are things on the carpets that would have to be moved, so we decided against the offices (in mine there's enough junk on the carpet that you can hardly see it anyway) and the master bedroom. But the two lounge rooms in the north of the house could be changed pretty quickly.
Off into Ballarat to check a few carpet shops. First was the Carpet & Tile Gallery in Armstrong Street South, where we were shocked by the prices—$200 for an average carpet. But they do themselves an injustice. The prices quoted are for a linear metre, laid. And the linear metre is 3.65 m (presumably 12 feet) wide, so that's really $55 per m². And what's that in terms of base price? $45/m²?
Spent some time looking and came up with a quite reasonable looking carpet for what promises to be “$120”, probably better than what we could get at Bunnings. And after the effort that the salesperson (who proved to be the managing director, Chris Cartledge) had gone to, we decided to accept that. He'll be along tomorrow to measure things—all part of the service, it seems.
While in town, went looking for water bed stuff. Nothing. Looks like we'll have to buy the parts on eBay after all.
|
Another day of python
|
Topic: technology | Link here |
Revisited my inefficient python program today, and as expected got it much more efficient—0.9 seconds instead of 140 minutes, just shy of a 1000-fold improvement in performance. It still wasn't easy, not because of program logic, but because of python strangenesses. The more I learn of python, the more I like LISP. How do you select from a composite object? It depends. Maybe there's a function to do it for you, maybe you can subscript them. Most of my modifications were related to finding the correct syntax for selecting what I wanted. Doubtless there are good reasons, but it seems so much easier just to keep everything as a list.
Funny python surprise of the day:
Can't python count? Yes, it can. It just lies about the parameters: there's an implicit self in there.
|
Are you old enough to drink?
|
Topic: general, brewing, opinion, technology | Link here |
While writing yesterday's diary, checked the Pilsner Urquell web site. What a catastrophe! You're not even allowed to read it if you're not of drinking age. And how do they check that? You have to enter your date of birth! And even that is difficult. Three drop-down choice menus:
|
|
||||||||
|
|
||||||||
|
|
||||||||
So if you're born in October to December, or after the 9th of any month, or before 2001 (i.e. if you are of drinking age almost anywhere in the world) you have to scroll down for every single selection. And for what? They can't check.
Maybe there are legal requirements for this sort of thing, which in themselves are silly: don't authenticate by anything you can't check. That's nonsense. You'd think they didn't want anybody to access their site.
| Tuesday, 9 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
More network hell
|
Topic: technology, general, opinion | Link here |
One of the most frustrating things about my wireless Internet connection is its variability. When it works well, it's better than a standard ADSL connection—I've had real-world transfer rates of up to 300 kB/s in both directions. But you can't rely on it, and since the beginning of the year even VoIP has become unreliable. And they still haven't started building the radiation tower.
Today, after nearly 5 days of connection and acceptable signal quality, things went to hell again. It's not just slowness, it's the timeouts that irritate me. For some reason DNS is a particular problem. Spent some time playing with my named configuration, in particular increasing the query timeout to 30 seconds:
That's only of limited use, though, as Edwin Groothuis pointed out: many programs have their own timeouts, normally 10 seconds. And I frequently reached the 30 second timeout.
The other issue was downloading the videos for my linear algebra course. They're generally in the order of 20 MB in size, so at 200 kB/s they should take about 100 s. When the radiation tower is operative, that should drop to 8 seconds. But my speeds today were more like 2.5 kB/s, which takes about 2¼ hours. It's enough to make you scream, especially since web downloads aren't really set up to restart a transfer. Ended up loading them to my external server (in a couple of seconds) and then rsyncing them here.
|
Linear algebra, finally
|
Topic: technology, opinion | Link here |
So I've finished my first two assignments for the linear algebra course, and finally we're getting to the subject of the course. Spent some time watching the video lectures, which are much easier than the assignments, probably because (so far) there's not much to learn. Still, it has taken a lot of time to get this far. Hopefully the assignments will become more understandable too.
|
Agurk salat
|
Topic: food and drink | Link here |
We've been watching the SBS Television series Food Safari. It's a strange mixture of excellence and sloppy research: they claimed that Rollmops are British, and that the French word for pork belly is Speck (which they pronounce “spake”), another German word. The real French word is “lard”. But some of the recipes they bring out are not bad, and Yvonne was particularly taken by the agurk salat (“cucumber salad”) in the Danish Food Safari a while back, so much that we're working out our own version. It's interesting in that it seems to fit just about any meal.
| Wednesday, 10 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Six years in Dereel
|
Topic: general | Link here |
Six years ago today we moved in to our house in Dereel. How time flies!
|
More maths
|
Topic: general, opinion | Link here |
On with the linear algebra course today, and finally it's getting towards what I was expecting. Why am I having difficulty? The thing that got me was the terminology (“onto” is an adjective in maths, it seems, though the OED doesn't seem to think so), but finally they've handled that, and I've at least got round to understanding my problem: they're talking about formal set theory, something that I have never done before. Not that that makes a difference: I have done complex numbers, but in a previous lecture they presented Euler's identity (which they call Euler's formula):
Without explanation, that's so mind-blowing that it's difficult to comprehend, as xkcd puts it:
I was sure I hadn't seen that before. But when I checked my school maths notes I discovered it on the first page of my 1964 notes, written presumably at the end of September, without name but with derivation. At least I correctly guessed that it could be proved in terms of infinite series, and when you look at it like that it no longer looks as surprising.
|
Another power failure
|
Topic: general | Link here |
Another short power failure this evening at 23:10.
| Thursday, 11 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
What's a vector?
|
Topic: technology, opinion | Link here |
On with my linear algebra course, which is getting easier. But it seems to come from a parallel universe. No wonder I was confused. I've known about vectors for over 50 years, and it's clear what they are: a magnitude and a direction. Just to be sure I dragged out my old university maths book, which showed exactly what I remember. But that's not what vectors are today, at least according to this course: they're a special kind of discrete function.
That's part of the reason why this is taking me so long: a new student would just learn and accept, but I'm stopping on every corner looking for reasoning. Wikipedia doesn't help: it comes up with dozens of different partial definitions, some self-referential, but it took me quite a while to find that it refers to my kind of vector as a Euclidean vector. I'm left wondering whether the way the course presents it is typical or not.
| Friday, 12 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Funny currency names
|
Topic: general, opinion | Link here |
Message in the mail today pointing at this page (“7 Facts You Never Knew About Our Aussie Dollar”, which says more about the author than the reader). One of the points interested me
Sir Robert Menzies suggested the new currency be called ‘royals’ – further showing his loyalty to the monarch. Eventually the dollar was agreed upon, though not after some other hysterical suggestions including: The Roo, The Digger, The Boomer, The Kanga, The Kwid, The Dinkum and more.
Imagine paying for something in Roo’s?
Indeed. What sensible country would name its currency after any kind of marsupial ?
|
Korean food again
|
Topic: food and drink, opinion | Link here |
We're planning to eat bulgogi tomorrow, and spent much of the day preparing for it. I wonder if it's worth the trouble; somehow all the accompaniments seem the same, and they certainly take longer than necessary.
One common ingredient is toasted sesame seed, which works relatively well in our toaster oven. Empirically 3 minutes on “LO” grill seem about right.
| Saturday, 13 July 2013 | Dereel | Images for 13 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Fixing Emacs
|
Topic: technology | Link here |
I've been using Emacs for ever, to the point where the key bindings come so naturally that I couldn't describe them without looking at what I do on the keyboard. So every small change made in subsequent versions of Emacs is particularly irritating.
I'm not the only one. Peter Jeremy was complaining about it on IRC recently, but unlike me, he did something about it. Here's his initialization code:
|
Sony quality and interference
|
Topic: general, opinion | Link here |
I've been tracking what appears to be Radio Frequency Interference for some time, mainly in regard to TV reception. But the little Sony mini Hi-Fi system that I bought four years ago has been emitting random hissing noises, each a few seconds long, for some time, and I've been wondering whether that's related.
Today things got a lot worse, to the point where the reception was no longer acceptable. The problem was, it was only the Sony receiver. The (26 year) old tuner in my bed works as well as it ever has. So I went looking for an alternative and found an old Marantz tuner (probably of the same vintage) which I connected up to the audio input. It works.
While investigating, discovered another thing with the Hi-Fi system: the main knob runs some digital converter that normally controls the volume. Now it seems to be getting confused about the direction: if I move it slowly in the “up” direction, it counts down. How can that happen?
One thing's clear, though: I used to think of Sony as a company that produced superior products. I've long taken a disliking to them for their business practices—abusing the GNU Public License, for example—but I was expecting a good product. This one has had three separate and unrelated failures (the other one was a failed CD drive, after less than 5 months), and the recent ones are not typical of modern electronics. For me, at any rate, it's clear that Sony no longer have any technical advantage.
And the extended warranty? When I bought it, they offered me an extended warranty for 4 years—for 35% more than the purchase price! Even if I had taken it, though, it would have expired 2 months ago.
|
Internode support?
|
Topic: technology, opinion | Link here |
My network connection has gone to hell again. After 5 days of relative peace (though not good throughput) it started again a couple of days ago: 3 disconnects on the 9th, 7 on the 10th, 3 on the 11th, 6 on the 12th, and 8 today. And in each case reconnecting my myriad TCP connections can take up to 5 minutes. I can no longer keep my MythTV programme information up to date, because the network link won't stay up long enough.
What should I do? Report it to Internode Support? That way madness lies. Once they were good. Now they don't even bother to escalate things. Is that because of the takeover by iiNet? I don't know, but when the radiation tower finally gets built, it's by no means a foregone conclusion that I'll stay with Internode.
|
Garlic sizes
|
Topic: food and drink, opinion | Link here |
While preparing dinner today, I found the largest garlic clove I have ever seen:
|
|||||||||||||||||||||||
I've already commented about the silliness of specifying things like “cloves of garlic”, but this just brings it to a head. I also found this one:
|
|||||||||||||||||||||||
That's a factor of over 25:1 between the two. And I've seen other cloves that are less than half the size of the small one.
|
Bulgogi: Australian incompatibility
|
Topic: food and drink, opinion | Link here |
As planned, ate bulgogi for dinner this evening, on the gas grill that I bought 5 years ago. I subsequently found the top part for it, but we never seem to have used it much. It worked relatively well, and it tasted good, but there was a problem: smoke. Within a few minutes of starting, the smoke alarm went off. We turned on the ceiling fan, which gave us another few minutes, but in the end we had to continue with the raclette grill, which really doesn't get hot enough. Clearly a dish for outdoors.
| Sunday, 14 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Preparing for the new carpets
|
Topic: general | Link here |
The new carpets will be laid on Tuesday, but first we need to remove things from the rooms. Here's part of the contents of Yvonne's bedroom:
|
|||||||||||||||||||||||
That room isn't finished, and there are two more to go. How I'm looking forward to it. And how happy I am that I didn't decide to do my office too.
| Monday, 15 July 2013 | Dereel | Images for 15 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
New carpets
|
Topic: general | Link here |
So we've been gradually clearing out rooms in preparation for the installation of new carpets tomorrow. And then we got a call: can they do it today? Well, we had the worst room cleared out:
|
|||||||||||||||||||||||
So we said yes, and an hour later a nameless installer showed up and knocked on the door instead of pressing the doorbell. We had been told that he would need all day, but in fact he had a dentist's appointment at 15:30, so he had to leave round 14:40. He optimistically thought he'd get it all done by then, but of course he was only half finished. He came back afterwards, though, and finally round 18:30 he was done.
Was it worth it? Well, now we have clean carpets. But what an effort!
|
Interior panoramas
|
Topic: photography, opinion | Link here |
Took some photos of two of the rooms before replacing the carpets. One's above; here's the other:
|
|||||||||||||||||||||||
That is looking down, of course, but the perspective is so extreme that it doesn't seem to be worth the trouble. I can't think of a good way to represent it. I took (almost) a full 360° × 180° view of the bedroom, so I was able to make an animated panorama of it, and that certainly looks better. But I didn't bother to do the ceiling of the lounge room, and without that I couldn't process the images at all. At the very least I need to find out how to make a full equirectangular projection of a panorama with less than 180° vertical field of view.
| Tuesday, 16 July 2013 | Dereel | Images for 16 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Carpet laying, the day after
|
Topic: general | Link here |
So now the carpets are laid, but there's still lots of work to do. The rest of the house is still a mess:
|
|||||||||||||||||||||||
Out to the post office in Napoleons to pick up the water bed accessories, and then attended to the bedroom. Yvonne has decided to rearrange the furniture in the room, which makes sense, but it means that I had to reattach the bookshelves to the wall, and of course I didn't have the right kind of plugs for this horrible plasterboard. After some cursing and swearing, gave up the attempt. Yvonne will have to buy some plugs tomorrow when she goes shopping.
|
A quarter of a lifetime
|
Topic: general, opinion | Link here |
I'm fascinated by the passage of time, and I have a script that runs once a day, giving dates an events that happened at various fractions of my life. Today I had:
A quarter of my life since I returned to Australia! How time flies!
|
Erecting radiation tower?
|
Topic: technology, opinion | Link here |
The start of construction for the radiation tower is long past the estimate of four to five weeks that I got two months ago, and yesterday I asked my sources again. No, no problems, and construction should start within days. But when I went past today, I still didn't see anything.
Well, not much. In the paddock next door there have been a number of old bales of hay. Now they're being removed and burnt (the smoke on the left):
|
|||||||||||||||||||||||
Are these the last preparations? Of course, it says nothing about the starting date for the construction.
|
More python insights
|
Topic: technology, opinion | Link here |
Still working on this linear algebra course. It's difficult enough that I'm left wondering about my capabilities, but gradually I'm coming to terms with it. Today was something more akin to traditional programming: implementing a class for manipulating the strange vector representation used in the class. That's a sparse vector consisting of two components: a list of possible values and those values that are non-0. So assigning a couple of typical vectors would look like this:
One obvious operation is vector addition. That's straightforward enough with a list comprehension:
The problem with this code is that, given the parameters above, it will return 0 values:
That's not “wrong”, but it's inefficient. Somehow I need to suppress the zero values. That, too, isn't that difficult:
The trouble here is that I'm calling the (relatively expensive) getitem() four times instead of twice, so instead of space inefficiency I get performance inefficiency. That's OK, I can assign the values the first time I call for them. So I tried this:
But no, that's a syntax error. Discussed it at length on IRC and discovered that other people were as dumbfounded by python syntax as I was, but ultimately that python assignments are not expressions, as this page explains: it allows you to make the mistake if x = 0: instead of if x == 0:.
Well, isn't that clever? So as a result I have to go and write it like this:
I suppose that's more intelligible, but somehow less pleasing to look at.
Other things that still get me are this obscure import syntax. Import a module into its own name space like this:
But to import it into the global name space, things change completely:
As if that wasn't enough, I spent quite some time reading the documentation without being any wiser at the end. Somehow the syntax doesn't match the semantics.
| Wednesday, 17 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Goodbye ACM Queue
|
Topic: technology, opinion | Link here |
It's been some years since this diary was included in the ACM Queue RSS feed. As I noted at the time, the topics in the diary range far beyond the normal subject material for ACM Queue, so I quickly modified the structure of the diary to present the content as a series of articles on various topics. ACM Queue takes the ones with a computer topic. Last month that was 33 articles out of a total of 91.
But somehow that's not restrictive enough. This is a diary, not a blog, and the main purpose is for me to make notes for myself. And lots of them, while relating to computers, must be deathly boring to most readers. So I'm further restricting content displayed on ACM Queue by explicitly excluding some things, such as my beginning steps in python. They're still there on my normal diary.
|
New English
|
Topic: language, technology, opinion | Link here |
What's a namespace? A filesystem? How do they differ from name spaces and file systems? Is it just a difference in spelling, or is it a difference in meaning? My spelling check highlights both words as incorrect.
Taking a step back: one of the biggest differences between English and German spelling is that in German nouns are written together, like „Filmempfindlichkeitseinstellung“, which looks terrifying until you split it up into Film Empfindlichkeits Einstellung (“film sensitivity setting”, which has the same number of syllables). There's a tendency in German to do this split, although it's a breach of spelling rules. And, it seems, there's the opposite tendency in English.
But discussing the matter on IRC, Callum Gibson suggested that there was a difference in meaning between “file system” and “filesystem”. I don't see that difference, though it's clear that the word itself has two distinct but related meanings: one is the code that provides the interface, usually in the kernel, and the other is the collection of data structures on the storage medium. There's probably more of a justification for “namespace”, which didn't previously have any intrinsic meaning.
So I went looking in the fountain of all knowledge. I drew a blank for “filesystem”, but “namespace” is there, attested since 1975:
1975 B. C. Goldstein & T. W. Scrutchin in Proc. 2nd Ann. Symposium on Computer Archit. (Assoc. Computing Machinery) 215/1 A local namespace is associated with each i-stream. Through the name space, the i-stream has access to other namespaces.
It's interesting to note that this quote wasn't sure whether to run the words together or not.
“File” is more interesting. OED has 7 meanings for it, including concubines, whores and worthless people (the last a separate meaning). But the one I'm looking for is the second. It derives from the (now obsolete) meaning “a thread”. There are many computer-related quotes there, but the word “file system” doesn't show up in any of them—clearly an oversight.
So: what do I do? There are many other run-together words in English, such as seafarer and bookcase, but they're older. Modern usage still seems to be to keep the words apart. Until I have a better idea, I'll stick to the OED usage.
|
Speck or bacon?
|
Topic: language, food and drink, opinion | Link here |
It's been a while since I revisited the Wikipedia Speck page, which had previously described a very limited use of the word. In the process of tidying it up, I came across this statement:
The term "Speck" became part of popular parlance only in the eighteenth century and replaced the older term "bachen", a cognate of "bacon".
The reference here is to German. OED tells me that the word “bacon” comes from the French. Yvonne has never heard of „Bachen“, though she points out that „Bache“ is a wild sow. So today I finally got round to looking it up in the Deutsches Wörterbuch. No hits, but there was one for „Bachenspeck“. The description is completely atypical of the dictionary:
BACHENSPECK, m. lardum pernae: bachenspeck mit ferkenschwenzlin herabwerfen. Garg. 81a. dem sinne nach, mit der wurst nach dem schinken werfen.
That's only a quotation. It doesn't even say what it means, beyond the Latin translation lardum pernae, which Google Translate translates as “bacon ham”. But why is this part of the DWB so superficial? By contrast, the entry for „Speck“ goes over a couple of pages. That's not related to the words, but the part of the dictionary. The part with „Bachenspeck“ was written by Jacob Grimm and completed in 1853, while „Speck“ was written after his death by people whose names say nothing, and completed in 1903.
So yes, maybe the word was once in use. But so far there's no substantiation for the claim in Wikipedia.
| Thursday, 18 July 2013 | Dereel | Images for 18 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Dining room panorama: case study
|
Topic: photography | Link here |
The 360° panorama I took of the mess in the dining room a couple of days ago may not have been of very great interest from the point of view of the subject, but it was a lot of work. The biggest problem was illumination: the room is relatively dark, but there are many windows. There's a great depth of field: the closest objects were less than 1 m away, but part of the view goes 15 m down the hallway:
|
|
|||||||||
|
|
|||||||||
|
|||||||||||||||||||||||
The distance makes flash impractical. Probably I should have done some HDR bracketing, but I'm not even sure that 5 EV would cover it. What I did instead was to take a series of images with the curtains drawn where I could. But I wasn't sure that the images in the foreground would be adequately exposed, so I went around again with the on-camera flash.
With these images I had a number of possibilities. Hugin offers a number of different ways of stitching:
Exposure corrected, low dynamic range: the images are simply blended.
Exposure fused from stacks. In this case Hugin looks for groups of images with the same viewpoint and first fuses them. This is effectively the same as what I do with my HDR preprocessing.
Exposure fused from any arrangement. Here Hugin looks for images with similar exposure and makes partial panoramas out of them. Then it joins them together. The danger here is ghosting, as I've seen in other examples.
The first time through I processed only the images taken without flash, 30 seconds at f/8. The simple LDR blending is clearly completely unacceptable. The highlights are badly burnt out:
|
|||||||||||||||||||||||
There's not really much point in running the “fused from stacks” option, because there are no stacks. But for the fun of it I tried it anyway, and compared it with the “fused from any arrangement” image (second):
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Not surprisingly, there's little to choose between the two, and they're much better than the simple LDR version. But the windows are still too bright and marginally burnt out, and in some areas the fused (“from stacks”) version isn't as good as the blended one:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
It's not clear where the jaggy in the second image came from: they were both generated with the same control points.
So what about adding flash? There was a little problem: using the built-in flash projected quite an obvious shadow of the panorama bracket at the bottom of the image:
|
|||||||||||||||||||||||
That showed up in the fused and blended images, but not in the LDR image (first), which, however, was no better than the version without flash:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Clearly there was no hope of getting anything useful from those images alone:
|
|||||||||||||||||||||||
But with appropriate masking and using all images, the result wasn't too bad:
|
|||||||||||||||||||||||
Basically, though, this kind of image is better suited to an interactive panorama, which, however, highlights some of the deficiencies in the processing. I've come some way, but it's not finished yet.
|
Filling the water bed
|
Topic: general | Link here |
So finally we have the bookcases in place, and we could fill up the water bed, all 720 odd litres of it. We decided to put in both the old and the new heating pads, which look surprisingly similar for different companies and 18 years' difference in age:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
I can only assume that that they're really from the same company, and the company changed its name in that time. Even the control units look identical.
We had the idea of putting some hot water in as well to lessen the power consumption of heating it up, but that proved to be too hot for the bladder. But to raise 720 litres of water by 10° takes 30.24 kWh of energy, which at 350 W takes—surprisingly-exactly 24 hours, so there was nothing we could do today.
|
Paying for the NBN
|
Topic: technology, opinion | Link here |
It seems that yesterday Simon Hackett gave a presentation about issues with the Australian National Broadband Network, pointing to serious deficiencies. In recent times most criticism relating to the NBN has been directed at the Federal Opposition's planned castration of the network, as I've commented in the past. But no, while Simon disagrees with that too, this time he's talking about the cost. If his calculations are right, by 2040 the NBN will cost 5 times as much as ADSL (which, strangely, will cost exactly as much as it does now, a round $20 per month).
He comes up with a number of suggestions about how to reduce the cost. Firstly, and by far the most controversially: get rid of POTS (or PSTN). That's fine in the cities—I'm planning to get rid of mine almost as soon as the radiation tower is operative—but it'll be a long time before they can get rid of it nation-wide. Still, he's talking about 2040, 27 years in the future. 27 years ago things looked very different.
Then he wants to reduce the size of the NBN connection hardware, which (for fibre) apparently includes 4 different boxes. Certainly the idea of supplying four different Ethernet connectors for 4 different ISPs appears a little baroque, but what's the chance that particular format will stay anyway? And I don't see it as being a big cost or size factor for very long. His real suggestion, though, is to limit NBN's activity to the fibre; the termination device should be supplied by the RSP (NBN's term for Retail Service Provider, i.e. an ISP), or presumably by the customer, like it is nowadays with ADSL modems. That certainly makes sense, especially to the ISP, who is now the only ISP involved.
Network infrastructure is another one. I didn't hear the presentation, but the slides appear to suggest 7 redundant POIs instead of the currently planned 121 POIs without failover. Certainly we need failover; the recent destruction of Warrnambool telephone exchange highlights that. It's interesting that a follow-up article about that event states:
The Victorian government is rolling out a backhaul fibre link to Warrnambool, and has suggested that this could be connected to the National Broadband Network (NBN). It warned that increased centralisation built into the NBN, and the decommissioning of the existing copper network, would make it more vulnerable to outages.
So: when even the government has recognized the issues, do we really need Simon to press the point? What I see from the article looks partially self-serving (only one ISP per connection, like with ADSL), but on the other hand it makes the point that the baroque design of the termination points is an unnecessary cost factor, and that allowing ISP or end user to supply the termination hardware would open the market to competition and reduce costs.
Somehow, though, I'm not convinced by this presentation. It's not up to Simon's normal standard. How could any of the changes he's proposing reduce the cost of the network by 80%?
| Friday, 19 July 2013 | Dereel | Images for 19 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Garden in mid-winter
|
Topic: gardening | Link here |
I should have taken photos of the garden on 15 July 2013, but I've been so busy lately that I only got round to it today. Despite the lack of attention and the middle of winter, things don't look too bad.
|
How much salt?
|
Topic: food and drink, opinion | Link here |
How much salt do you put in a dish? Look at just about any dish requiring salt, and it will be specified simply as “salt”, or if you're lucky “salt to taste”. A complete cop-out. It's all the worse when, as in many cases, you can't tell whether the salt was right until the dish is finished.
An example: as the result of a couple of misunderstandings, we have many kilograms of dried beans, peas and lentils, so I'm planning to eat more of them. Today we had a salade de haricots (white bean salad), which I've eaten before. But the recipe didn't look convincing: 500 g dried beans, “10 g (?)” salt? That seemed wrong, so I added much more, ultimately 35 g. And at the end, Yvonne told me there wasn't enough salt.
Not just that: the sauce tasted boring. Last time it didn't. But just a little salt brought it to life. It's surprising how much difference it makes.
Looking at the online version of the recipe, it seems that I quickly revised the 10 g (?) to 20 g (?), but didn't note the fact in the printed recipe. Still, that's also far too little. It looks like 45 g would be about appropriate—and that means, effectively, 9% salt by weight for beans. I should start keeping a list of how much salt each kind of dish requires.
| Saturday, 20 July 2013 | Dereel | Images for 20 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Linear algebra pain
|
Topic: technology, opinion | Link here |
Finally finished last week's assignments for my linear algebra course. Algebra? Statistical analysis using vectors. They're not due until next week, but the next assignments are there already, and I want to keep up to date.
So on with the next week's lectures. No question, we're talking mathematics here. And I don't understand any of it! At the beginning of the series it seemed that they paid too much attention to things that were obvious; now it's very much the other way round. What's a vector space? I have a vague idea, but only a vague one, and the concept is central to the whole course. Maybe part of the problem is that I want to understand where the story is going. Another is that the vector representation is so strange. The lecturer justifies this with the claim that it's easy to understand vectors if you have a concrete representation, rather than using the abstract concepts. I think it's probably the other way round: until you understand the concepts, you can't implement. Looks like I have a lot of reading to do elsewhere. And why are these lectures so short? Even the longest are less than 20 minutes long.
|
Hugin stitching methods again
|
Topic: photography | Link here |
House photos again today. Since discovering the alternative stitching methods (blended and fused), I've been using them more and more. But it's becoming ever clearer that they only work well when the panorama matches up really well. Today I did the verandah panorama with “fused from any arrangement”, and got a flash panorama looking like this:
|
||||||||||||||||||||||
|
|
||||||||
That's because the images are blended rather than cut. With the default “LDR” stitching, things are greatly improved:
|
|
|||||||||
|
|
|||||||||
So: blending only works when the images match really perfectly.
| Sunday, 21 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Debugging web pages, the hard way
|
Topic: technology, opinion | Link here |
I'm somewhat self-conscious about my slow progress in the linear algebra course, to the point where I modified my PHP scripts to exclude certain topics, such as this one, from publication on ACM Queue. And then I wrote another article on the topic, and it showed up in ACM Queue anyway. It didn't take long to find the bug—but that should never have happened. Egg on my face.
|
Morales affair: pots and kettles?
|
Topic: general, opinion | Link here |
I was horrified by the report a few weeks back that Evo Morales, the president of Bolivia, had been refused transit through Europe from Moscow to La Paz because he was suspected of transporting Edward Snowden. But further research suggested that the reporting was very inaccurate: there was conflicting information, the countries that allegedly denied passage weren't even on the way, the country in which they landed was not included in the list of rogue countries, and there was clear evidence of equipment malfunction. As a result, I'm watching how the matter develops, and whether we will ever hear an official reaction from Italy, which I suspect was only included in the list because it was necessary to support the story.
I've subscribed to Google alerts looking for the text “Morales complaint”. It's amazing how many false positives there are, mainly related to people called Morales in trouble with the US law. Today I found one that related to Evo Morales and intercepted planes—but not what I expected. It seems that the Bolivian Air Force searched a Brazilian military plane assigned to the Brazilian Defence Minister two years ago and searched it for drugs. Another report suggests they were really searching for a stowaway. Now isn't that interesting? Is the concept of intercepting and searching planes a well-known concept in Bolivia?
|
More watch timing
|
Topic: general | Link here |
It's been over a month since I last set the time on my new watch. At the time I had estimated that it was gaining about 6 seconds a month. But since that time it has gained 15 seconds, more like 12 seconds a month. Set it to 4 seconds slow again.
|
Brewing goodbye
|
Topic: brewing | Link here |
It's been over 3 years since I last brewed any beer. And I don't see it happening again any time soon; I've become too lazy for that. But I had plenty of supplies lying around—3 to 4 year old malt, barely still usable, and what proved to be several kilograms of hops. Offered them to a good home on Freecycle a while back, and finally found a taker who knew what to do with them. They came from Bacchus Marsh today and picked them up. Now to get rid of the equipment, which should be worth some money.
| Monday, 22 July 2013 | Dereel | Images for 22 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Apps for cooking
|
Topic: food and drink, opinion, technology | Link here |
A lot of discussion about cooking on IRC today. Does it make sense to use a tablet computer to view your recipes? That was one of the reasons I bought a tablet last year. My experience with this particular device was so negative that I didn't even try it in the kitchen: it went back. But that was that specific tablet, and potentially a tablet could be useful in the kitchen; it's just that it seems to be a lot of money for one small application. If I were to do that, a relatively modern laptop seems preferable.
Then Gregory Orange came up singing the praises of My CookBook, a tablet App that he uses extensively in the kitchen. Went looking, and found a description with screen shots:
|
|
||||||||
That goes against a lot of things I consider holy: measurements in cups, tiny display. Gregory tells me that this is a phone display, and that a “real” tablet with 1024x768 display would look better. And you can edit the recipes.
On the positive side, though, the idea of displaying the ingredients on the left and the method on the right makes a lot of sense. This example does it wrong: it refers to apples in the directions, but they don't show (yet) in the ingredients. Still, it's a good idea, and something that I might be able to do in my own recipes.
Apart from that, found some interesting features:
The first point is unclear, but the other two make a lot of sense. And not for the first time, I was left wondering whether my “I did it my way” approach is not limiting what I can do. Did some more investigation and found that there's a normal web interface, so signed up for that.
The results were more sobering. The first thing is to find a recipe. OK, I was already looking for a recipe for Bibimbap, so did a search. Eight recipes—not bad, until I realized that they were nearly all from taste.com.au. It seems that the choice of language also limits your search to sites with that “language”. It makes sense to limit the recipes to a specific natural language, but this application distinguishes between five different forms of “English”. In my case, it was “English (australia)” [sic], and as a result I only got Australian recipes. With “English (united states)” I got no less than 318,000 results—far too many. Tried importing a couple, including a French and a German version. The web interface doesn't make it easy: after finding the recipes, you have to click on “Import recipes”, and then you discover that no context gets passed—you have to copy the URL to enter.
After that, things aren't too bad. It strips the normal random junk that so many online recipes have, and the resultant format isn't bad. But the content is still an unchanged online recipe. It should be possible to edit the recipe, but all I can see is “my comments”, which is rather too little for my liking.
One decidedly good thing about the site, though, is the list of sites from which the application can import recipes. That and the search function have a higher signal to noise ratio than a random search with Google.
|
Pollo ai funghi
|
Topic: food and drink | Link here |
We had some risotto left over from a meal last week, and some chicken breast as well, so I went looking for an Italian style chicken dish which would go with it. Found a recipe for Pollo in cacciucco in Lonernza De' Medici's “Tuscany: The beautiful Cookbook”, but that was for a whole chicken. So in the end I called what I made simply pollo ai funghi (chicken with mushrooms).
| Tuesday, 23 July 2013 | Dereel | Images for 23 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Installing shelves
|
Topic: general | Link here |
When we moved in to the house six years ago, it was clear that we needed more shelf space, and on 1 August 2007 I bought some shelving for the four wardrobes in various rooms. I didn't get them installed quickly—the first ones were a year later, and to date that's all I have done.
For some reason, Yvonne got impatient, so I've finally bit the bullet and asked Bryan Jackson to put the shelves in the other three wardrobes. Finally we have space!
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
|
|||||||||||||||||||||
|
More July anniversaries
|
Topic: general, opinion, technology | Link here |
I've just finished reading the end of my paper diaries, from January 1968 via the end of contiguous diaries on 1 October 1970 until the final isolated entry on 14 June 1977. It hasn't been as pleasant a reading as I had expected—you remember the good things, but you tend to write down the bad things as well, and there's so much of that that I'll probably never type them in.
But one thing I discovered was that my original assumptions about various anniversaries were wrong. I've already noted that a number of anniversaries fall in mid to late July—in two days' time, for example, I will have known Yvonne for 31 years—but now I discover that the end of July seems to have been a good time to meet girls. 44 years ago I met Teen Rozalla, who was important to me at the time, and it seems that I met my first wife, Doris née Pischke, on 19 July 1973, just over 40 years ago, though at the time it was so important that I didn't make any mention of it for a month. Previously I had thought that this had happened a month earlier.
In fact, it was a month earlier, on 28 June 1973
And then there are the three houses we moved in to on 27 July 1982, 17 July 1997 and 10 July 2007, and my first arrival in Germany on 27 July 1967. They say time stops things all happening at once, but something seems to have gone wrong here.
|
Blurring computer history
|
Topic: technology, history, opinion | Link here |
Seen on Pinterest while looking for something different and useful:
Rear Admiral Grace Hopper. She worked for UNIVAC in 1949 who made some of the first computers ever. In 1951 she discovered the first computer “bug.”. In 1952 she had an operational compiler. “Nobody believed that,” she said. “I had a running compiler and nobody would touch it. They told me computers could only do arithmetic.” A compiler is the reason you have an Operating System with programs, a phone with apps. There would be no Windows or Apple or facebook or twitter or tumblr without her.
OK, I don't know much about Grace Hopper, but I did know that she found her bug in the 1940s, and that UNIVAC was introduced in 1951. And compilers in 1952? Where does this nonsense come from?
Further investigation was instructive. Yes, Hopper did join the Eckert and Mauchly Company Computer, which later became UNIVAC, in 1949. And yes, she created a “compiler” for the A-0 System in 1952. It seems that calling it a compiler required a stretch of the imagination: autocoder might have been a better term—indeed, it seems that she used this term herself. But it also seems that she coined the term “compiler”, so it can't be wrong. The only thing that's completely wrong was the date of the bug, which was found in 1947.
But: “A compiler is the reason you have an Operating System with programs”. Like OS/360 or even the earliest versions of Unix, no doubt. But It Must Be True. I read it on the Internet.
And that's the problem. There's a germ of veracity in the article, but it's so lopsided that I initially thought it was all wrong. God protect us from this level of “information”.
| Wednesday, 24 July 2013 | Dereel → Ballarat → Dereel | Images for 24 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
A new vacuum cleaner
|
Topic: general, opinion | Link here |
Two years ago I bought a new vacuum cleaner for Yvonne. I was very happy with the choice, but not so Yvonne, who kept complaining that it didn't work well. After the new carpets were laid, I used it myself. She was right. So we already need a new vacuum cleaner. This time Yvonne took me to a specialist, Vacuums Ballarat in Norwich Plaza, along with Zhivago to illustrate one of the issues. Small company, husband (Bill) and wife (Viv), but Bill knows his stuff and ended up selling us a semi-professional device with easy cleaning, for 4 times the price I paid two years ago.
Back home, tried the thing out. Yes, much better. But the junk it pulls out of the carpets still clogs things, and the instructions are some of the worst I have seen. It has a big paper filter, about 35 cm in diameter, and it needs removing. How? They don't say. If Bill hadn't demonstrated one to me, I wouldn't have been able to work it out. As it was, it seems that the filter was wedged in the upper part of the device and needed a tug with a pair of pliers to remove it. But that should be explained in the instructions.
|
Upgrading Napoleons General Store
|
Topic: general | Link here |
On the way into town, stopped at the Napoleons General Store. Small shops in remote areas have problems, and many have shut down, but they seem to be doing something right in Napoleons. So it was a little disappointing to see that they have removed the petrol pumps:
|
|||||||||||||||||||||||
The petrol there was relatively expensive, of course, but occasionally I've found it made sense to top up the tank there rather than go the further 12 km into Sebastopol. It looked ominous that they are removing it. Such proved not to be the case, though: they're just modernizing it.
|
Haircut for Zhivago
|
Topic: animals | Link here |
We took Zhivago to town primarily because he didn't want to stay at home alone, though it was good to demonstrate him for the vacuum cleaner. On the way home, I went for my quarterly haircut and took Zhivago with me. The barber's eyes lit up:
|
|||||||||||||||||||||||
|
My CookBook in practice
|
Topic: food and drink, opinion | Link here |
One of the things that I found while investigating My CookBook a couple of days ago was the Cuisine et vins de France web site. CVF is also a magazine, or it was 20 years ago, when we subscribed to it for a couple of years. We still have the magazines—they were frustrating because many recipes required ingredients we couldn't get even in Germany—and today I pulled them out and took a look. One of the recipes, tajine de bœuf aux pois chiches, caught our eye, and we decided to cook it this weekend.
So I went looking on My CookBook, and sure enough, I found the recipe pretty much verbatim, strangely not on CVF's site but http://www.cuisineaz.com/. So I fought the “import” interface and “imported” it, to a somewhat unnecessarily password protected page. What did it really do? It removed the advertisements, but also some of the comments (“Difficulty”, for example), and it reformatted it, somewhat worse than the original. Do I need that? Maybe it's useful for the app, but I don't see much added value in this particular case. I'm still much happier with my version, though all I really did was to rearrange the ingredients to match the sequence of preparation.
|
Disk problems with a difference
|
Topic: technology | Link here |
Trying to back up my photos today, I ran into a problem I hadn't seen before:
=== root@eureka (/dev/pts/11) /home/grog 20 -> mount /dev/da2p1 /photobackup
Huh? I had just plugged in the (USB) disk and confirmed that it had been probed successfully:
Disk label problems? But it seemed to be OK:
=== root@eureka (/dev/pts/11) /home/grog 16 -> gpart show
So I plugged it in to lagoon, Yvonne's computer and ran fsck, which found a minor issue. But it still wouldn't mount on eureka.
After some investigation, discovered:
=== root@eureka (/dev/pts/11) /home/grog 22 -> df
Is this Schrödinger's cat? It's mounted and it's not mounted?
No, what had happened was that it appears that I had forgotten to umount the other backup disk before removing it. So the mount remained, but that device was no longer there (ENODEV). On earlier versions of FreeBSD I would have had a panic. Now I only had a confusing error message.
Clearly there's scope for improvement here. The most obvious one is a less confusing error message, but how about a forced umount of the old device under these circumstances? And it would be interesting to consider an automatic soft umount after a certain period of inactivity: the disk remains mounted, but the data structures on disk are synchronized in the same way as umount would do, so that if it gets removed, it's completely consistent and doesn't need an fsck.
| Thursday, 25 July 2013 | Dereel | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Spammers getting even more stupid?
|
Topic: technology, opinion | Link here |
Spam is bad at the moment, and I'm continually wondering whether I shouldn't do something draconian like whitelists. But the spammers are not getting any cleverer. Or maybe they're catering to the toy MUA crowd who can't see the nonsense they're sending:
There's lots of stuff that looks plausible, and normally you don't look at the headers that closely. But sending a message like that to three people? What are they thinking?
|
Symlinks with Microsoft
|
Topic: technology | Link here |
One of the results of tidying up the house was that I found a whole lot of old photos that I need to scan in. And after my experience with SANE, I've decided to use my Microsoft box to do scanning. That works about as well as you can expect with a Microsoft box, but one irritating thing is that it saves the scanned data on the Microsoft machine, and I have to move it manually to eureka. What I need is a symlink.
But doesn't Microsoft have symlink functionality? Does it work to external file systems? Asked on IRC and was told that it was called a “shortcut” (another modern joining of two words), and that I could make one by pushing mice between “Windows Explorer” windows.
I feel uneasy with this sort of thing: I want to be able to write down how to do it, and the obvious way is to use an appropriate command. But the combined wisdom of the Microsoft-heads on IRC only managed to find an almost appropriate web page. But it did describe how to do it: a command called MKLINK:
Microsoft's language is interesting. On the one hand they call directories “folders”, and symlinks appear to be “shortcuts”, but in other places they use the correct names, like above.
| Friday, 26 July 2013 | Dereel | Images for 26 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Linear algebra: the pain continues
|
Topic: technology, opinion | Link here |
On with my linear algebra course today. I'm gradually catching up with the schedule, but I'm still having great difficulty understanding the material. Much of it seems to be a matter of inadequate explanation. For example, in the same lecture I get these two statements:
Span {[1,2],[3,4]}: All points in the plane. Two-dimensional
Span {[1,3],[2,6]} is 1-dimensional.
Huh? Is this stuff correct? How can that be? No explanation. It took me a while to realize that the issue here is that [2,6] is in the span of [1,3], whereas [3,4] is not in the span of [1,2]. But why not explain it?
I now understand that one, but in the next week's lectures the descriptions of multiplying matrices with vectors are so hazy that I'm still not sure that I've understood them. For some reason, the Wikipedia page also confuses me, maybe because it uses notations that are unfamiliar to me. And yet all this stuff is so simple! I've found I already have the sources for the operations, written in FORTRAN IV and presumably over 40 years old:
Admire the comments. Admittedly, this comes from a package (UNIVAC, I think) that came with a detailed manual explaining the parameters. It's also interesting to see how FORTRAN's lack of array metadata significantly complicates the calling sequence.
|
Enblend refuses to stitch more than 9 images
|
Topic: photography, technology | Link here |
Some weeks ago Thomas Zenker reported a problem with the FreeBSD port of enblend, which I maintain: he couldn't get it to stitch more than 9 images at a time. It aborted with the messages:
He thought this was a general restriction, but of course I have been stitching many more than that, coincidentally with an almost identical configuration. He sent me his images and I was able to stitch them with no problems. What was the difference? I sent him some more suggestions about what to do, but he wanted to go off in another direction: since the error was reported by vigra, he wanted to try various versions of the dependent ports.
In my experience, that's the wrong approach. But it was Thomas' choice, so I left him to it. Today I got a message back: yes, it was vigra. Not just that—it depended on what compiler it was compiled with! Currently if you include numpy, it gets compiled with gcc 4.6; otherwise it gets compiled with gcc 4.2.
That can't be true! So I tried it. It is true. What a horrible situation! Thomas suggests that some structures get allocated differently by the two compilers. That's possible, but one way or another it's a real concern.
|
Nele Kömle visits
|
Topic: general, music | Link here |
Nele Kömle is doing some horsey stuff at Chris Bahlo's at the moment, and we thought it prudent for her to stay with us. It seems she's taking classical singing lessons—one of the things she's working on at the moment is „Ruhe sanft, mein holdes Leben“ from Mozart's Zaide. She knows this recording with Lucia Popp:
It reminded me of the aria “Che farò senza Euridice?” in Gluck's “Orfeo ed Euridice”, so I pointed her at the Naxos CD library at the State Library of Victoria, so we tried things out. It didn't work well. My notwork connection showed itself from its worst side, so I dug out a CD of Orfeo and tried to play it. How do you play a CD with mplayer? I've forgotten; talking to cd://1 didn't work. But I have a CD player there, so I tried that. And since the last time I used it, it has developed mechanical problems and wouldn't open. Another repair job to look at some other time.
So back to SLV. This time the network connection was better, but SLV were having trouble with their system. After about 30 minutes we finally got it to play and discovered that the recording we found wasn't nearly as good as the one with Popp. Somehow the Naxos library needs significantly better search functions.
And only later did it occur to me that „Ruhe sanft“ bears quite a strong resemblance to „Schlummert ein, ihr matten Augen“ from Bach's „Ich habe genug“. Since they both talk about sleeping, I wonder if they might have been derived from an old lullaby.
| Saturday, 27 July 2013 | Dereel | Images for 27 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Enblend insider joke?
|
Topic: technology, photography, opinion | Link here |
While looking at the enblend home page today, I noticed an interesting detail:
|
||||||||||||||||||||||
|
|
||||||||
That jaggy in the white stripe (which is part of a SVG image) is exactly the kind of thing that enblend is supposed to eliminate. I wonder what the thought processes behind it are.
|
Tajine de bœuf aux pois chiches
|
Topic: food and drink, opinion | Link here |
As planned, cooked our tajine de bœuf aux pois chiches from the recipe that we got from cuisineaz.com. It required more frobbing than I had expected:
No specifications for quantities of salt, of course. In total we used 35 g.
The spices seemed inappropriate for North African food, in particular quatre épices, and we ended up changing them. Left out the quatre épices and added cumminseed and cinnamon. Not enough, as it turned out; I've adjusted the recipe for “next time”.
Why do so many recipes cook food until it disintegrates? This recipe wanted the chick peas cooked for an hour before even starting the main preparation (requiring cooking for another 90 minutes). After the first hour they were cooked through.
A Tajine in an oven? The whole purpose of the thing is to keep the pyramid at the top cool. We cooked it on the stove.
And the result? Quite good, not spectacular. But the quantities! It was only 500 g of meat, but between the three of us (with Chris Bahlo) we didn't eat more than a third of it, and in the end we were completely bloated. We had served it with Cous-cous, which is really not necessary: the chick peas do the job.
|
Darah with colic?
|
Topic: animals, opinion | Link here |
Yvonne came in round dusk to say that Darah was lying on the ground and wouldn't stay up. There's nothing wrong with horses lying down from time to time, but this was too long. Went out to take a look. Her neck muscles were very tense, something that she's had before and which could lead to Laminitis. But she seemed to be in no pain, and I was able to get her to stand up for a while, but then she lay down again, rolled over and kicked, and then turned back on her stomach. Colic? She had just eaten, which horses with colic normally don't to. But it would have to happen on a Saturday evening.
Yvonne called Chris Bahlo, who was due for dinner anyway in an hour, and she came and took a look. This time Darah had her forelegs stretched out, which Chris thought might be to press her stomach on the ground to cool it. But she wasn't sweating, which is usually an issue with colic. We checked her for muscle pain and found nothing there either, but her gums were very pale, suggesting some kind of circulatory issue.
In the end we decided to leave it until after dinner. Yvonne didn't want to come out to look, but she needn't have worried. Darah was standing up and looking relatively normal; a couple of hours later she still was, so whatever it was, it doesn't seem to have been as serious as we feared.
|
Four spices?
|
Topic: food and drink, opinion | Link here |
One of the ingredients in the original tajine de bœuf aux pois chiches recipe was “quatre épices (“four spices”). But which? Went looking in our cookbooks, and found information only in Stephanie Alexander's The Cook's Companion. There she wanted cloves, pepper, nutmeg and powdered dried ginger. That doesn't sound very French. So I went looking on Wikipedia and discovered that the number four is a coincidence: it's really Pimenta dioica, known in English as Allspice, along with a surprising number of names in various languages. Interestingly, the page describes the flavour as being exactly what Stephanie Alexander suggests. And it is used in Morocco, and Mohamed Ifadir confirms that it's quite popular.
|
Cooking cous-cous
|
Topic: food and drink, opinion | Link here |
We served Cous-cous with the dinner. I've always had problems getting the kernels fluffy and not sticking together, even with pre-cooked cous-cous (which is about all we can buy round here). Tried mixing it in my Kenwood Chef mixer, but that's far too difficult for it. It relies on things sticking together, and if they don't, it just pushes them to the edge:
|
|||||||||||||||||||||||
Yet another case where the mixer fails.
| Sunday, 28 July 2013 | Dereel | Images for 28 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Mystery bassoon
|
Topic: music | Link here |
Received mail from Donald Rumph enclosing a number of photos of the remains of a bassoon that he bought recently. It's in very poor condition, probably not worth restoring, but it's interesting because of its construction:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The bell and keywork appear to be missing, which doesn't make it any easier to identify, but looking at the construction I'd guess it to be round 1810. The back of the butt shows a number of surprising things:
|
|||||||||||||||||||||||
There's some kind of floral decoration round the top and bottom, an oval piece of wood missing above the F hole, F♯ key missing, and what looks like a metal flap between the F hole and where the F♯ hole would have been.
Who made it? The wood reminds me of Savary jeune, but that's about all that does. On the bass joint there's a saddle and hole to the left of the D key, obviously for E♭.
|
|
|||||||||
But nearly all bassoons have that key on the other side, like my Sautermeister bassoon:
|
|||||||||||||||||||||||
The only ones I could find that didn't are English, like the George Astor one depicted in this thumbnail and the Thomas Key instrument here.
And the stuff on the butt? Hard to say. I'd guess that the oval hole once contained an ornate manufacturer's plate, something along the lines of the plate on my Thibouville oboe:
|
|||||||||||||||||||||||
And the metal flap? No idea. The strange thing is that the instrument really doesn't seem to have an F♯ key. The flap is far too high to cover the hole. But I've never seen a bassoon of that era without an F♯ key before.
|
Darah on the mend
|
Topic: animals | Link here |
By chance Caroline Hamilton came by this afternoon, mainly to look at Wotan (who is part Iceland pony and should thus really be called Óðinn), who has some back problems. While she was here she took a look at Darah and came to the conclusion that yes, she had probably had a mild colic. She found her stomach to be sore to the touch—exactly what I had tried the previous evening without noticing anything—and gave her some treatment to loosen her stomach. It looks like it was a relatively mild attack.
|
Linear algebra: dirty tricks
|
Topic: technology, opinion | Link here |
I have my doubts about this linear algebra course that I'm doing, but I'm really determined to finish it anyway. They've brought out a book to match the course. Is it any better? It has a total of 528 pages of American “letter” paper, which certainly makes it a lot more detailed than the lectures, and it includes material not in the course. I'm not convinced, but it seems a good investment, though the price relationships are irritating: $30 for the book and $26 for postage so that it will get here before the course is over.
On with the lectures, and it's beginning to make more sense. Solving matrix equations using standard generators. One of the nice things about watching lectures on video is that you can pause them and try to first-guess what the lecturer is saying. For a while it worked. First example: stretch an image 2× horizontally. Conversion matrix is [[2, 0], [0,1]]. Makes perfect sense.
Next example: rotate the image 90° counter-clockwise. [1, 0] becomes [0, 1], [0, 1] becomes [-1, 0]. So the conversion matrix is [[0, -1], [1, 0]].
Next rotate counterclockwise by θ°. [1, 0] becomes [cos θ, sin θ], and [0, 1] becomes [- sin θ, cos θ], so the matrix is [[cos θ, - sin θ], [sin θ, cos θ]]. Simple once you get used to it, and I was able to write it down here without referring to the notes.
Next one seems simple enough: translate the vectors by [1, 2]. That's additive. How do you represent it with multiplication? I sat there for a while and couldn't work out how to do it. Then I looked at the example. He came up with a conversion matrix [[2, 1], [2, 3]]. And yes, put [1, 0] or [0, 1] into that matrix and you come out with the correct translation. But that's the only way you get good results. [1, 1] should give [2. 3], but in fact it gives [3, 5].
What didn't I understand? Went looking elsewhere and found that it shouldn't work at all, because all this stuff assumes that the origin doesn't change.
But despite shortcomings in the presentation, there seems to be no reason to doubt the lecturer's understanding of his subject, especially in such a basic area. Spent over an hour trying to understand, and finally gave up and moved on to the next lecture. Surprise, surprise! “What I said in the previous lecture wasn't strictly true”. In fact, it was Just Plain Wrong. Yes, working through incorrect assumptions is not necessarily bad, but it needs faster resolution, certainly in the same lecture, and preferably immediately after presenting the calculation (he moved on to other topics afterwards in that lecture). No wonder I'm having such difficulty.
|
Fixing the tajine recipe
|
Topic: food and drink, opinion | Link here |
Based on yesterday's experience (a French word that also means “experiment”) with the tajine de bœuf aux pois chiches, typed in my changes. It took me 90 minutes and generated 162 lines of diffs. I wonder how long that would take with a tablet and without a keyboard.
| Monday, 29 July 2013 | Dereel | Images for 29 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Using wake on LAN
|
Topic: technology | Link here |
On Saturday evening Chris Bahlo were looking at the web site of her new employer, ruadvertising.com.au. First question: does it render correctly? Well, sort of, modulo overrun at the bottom, caused by guessing that I would use the standard character size. We were looking at the page on the TV, 58" diagonal, but some distance away. I've already noted that resolution isn't the issue: it's angle of view. At default sizes, it's illegible on the TV.
Chris took that on board and then asked “And what is it like under Internet Explorer?”. I knew the answer, but of course the real challenge was getting “Internet Explorer” to display on the TV. For that I had to go into the office and wake dxo, my Microsoft box. And of course “Internet Explorer” displayed correctly, but illegibly small.
The real issue, though: why should I have gone into the office? Doesn't this thing have Wake on LAN? It handles suspend very well. Today I did some investigation and ran into lots of dead ends. You'd expect to find the information on the Microsoft web site somewhere, but all I could find was related to “Windows” 7 (I'm using “Vista”). Even when I searched specifically for “Vista” documentation, the first link was to the ”Windows” 7 documentation, and the second, though it looked promising, had no information on the topic. Thank you, Microsoft.
So I went out looking for details on the web, and came up with the usual selection, most relating to problems with the facility. This page looked almost OK until I got to point 7:
Enable the following options by highlighting them and selecting "Enable" in the drop-down menu: "Wake on LAN," "Wake on Magic Packet," "Wake on Pattern Match," "Wake From Shutdown" and "Wake on LAN From Power Off."
Problem: none of these settings were there. Wrong NIC? Possible. But there's a setting under “Power Management”: “Allow this device to wake the computer”. So I set that.
Next point: how do I send a wakeup packet? That proved to be much easier than I thought. FreeBSD has a program wake(1) that does just that. There are different ways, but it seemed the cleanest was to add the name of the machine in /etc/ethers:
The man page claims that the name has to be fully qualified, but it lies. And how about that, it worked! Total work about 10 minutes.
And then I heard a noise. Following up, discovered that dxo was running again. I hadn't done anything. Put it back to sleep and pinged it. It woke up.
More searching the web finally came up with some documentation from Microsoft, which said:
One kind of special data packet contains a wake-up pattern. By default, Windows 7 and Windows Vista listen for the following packets when you enable WOL:
A directed packet to the MAC address of the network adapter
A NetBIOS name resolution broadcast for the local computer name
An Address Resolution Protocol (ARP) packet for the IPv4 address of the network adapter
An IPv6 Neighbor Discovery packet for the network adapter's solicited-node multicast address
A Magic Packet can also wake a remote computer. A Magic Packet is a standard wake-up frame that targets a specific network interface.
That's actually not a bad piece of documentation, and I can even understand that those packets could reasonably be expected to wake the machine. And ita carries on to say that you can get “normal” Wake on LAN behaviour by setting the “Only allow management stations to wake the computer” box in the power management tab:
|
|
||||||||
And sure enough, that fixed it. It's good to notice that the text has changed in “Windows” 7. It now reads “Only allow a magic packet to wake the computer”.
|
Musical instrument photography
|
Topic: photography, opinion, music | Link here |
While writing up my article on the mystery bassoon yesterday, I went looking for the image of the manufacturer's plate on my Thibouville oboe. It didn't look good:
|
|||||||||||||||||||||||
Why did I make such a mess of it? It was taken with my old Nikon “Coolpix” 880, but it had a usable viewfinder. But looking back at my instrument photos, few of them look much good. Were my standards that much lower in those days?
In any case, I still have the oboe, so it's relatively simple to take another photo. Years ago I bought a “light tent” for taking this kind of photo:
|
|||||||||||||||||||||||
So I took a new photo, which came out considerably better:
|
|||||||||||||||||||||||
While I was at it, noted that the photos of my old bassoons weren't that flash either. It seems that this is all I have taken myself of the Sautermeister bassoon:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Aligi Voltan had already sent me some better images, but why not take some more photos while the tent was set up? The tent is too small for a bassoon, so I took the Sautermeister apart and took photos of the joints. They didn't look bad, but they had obvious flare:
|
|||||||||||||||||||||||
Where did that come from? The tent, obviously. Part of the function is to spread light in all directions, including into the lens. How about a lens hood? On the whole I'm not overly convinced by lens hoods, especially on zoom lenses where they can only work properly at the shortest focal length. But it's worth a try:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Still flare. And it's surprising how much difference there was between these two photos, taken from an identical position under identical circumstances. The only difference was the lens: the first image was taken with the Zuiko Digital ED 12-60mm F2.8-4.0 SWD, and the second with the Zuiko Digital ED 50mm F2.0 Macro. I had expected the image quality of the second photo to be better.
After a bit of thinking, decided to take photos without the tent, just with reflectors on the flash units, to compare them. I suppose the photo is marginally better:
|
|||||||||||||||||||||||
But somehow it's still not what I was expecting. More head-scratching to do.
|
My CookBook on Android
|
Topic: food and drink, opinion, technology | Link here |
Gregory Orange reported back today. He had taken the original version of the recipe for tajine de bœuf aux pois chiches and put it on the My CookBook app on his Android device. It's difficult to get a screen shot, but what he saw was:
|
|||||||||||||||||||||||
How much use is that? I suppose it's about as good as the original recipe. But in the meantime I've changed it, and I wonder how long it would take Gregory to update it accordingly. At least it seems to be possible, unlike the web version.
| Tuesday, 30 July 2013 | Dereel | Images for 30 July 2013 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
|
Wake on LAN: the rest
|
Topic: technology | Link here |
So yesterday I configured Wake on LAN on dxo, my Microsoft box. Modulo some unexpected behaviour (wake on any LAN event), it went remarkably smoothly. Today I had more photos to process, so I tried to wake up dxo again. Nothing.
Further investigation showed that wake(1) wasn't working:
=== grog@eureka (/dev/pts/11) ~ 42 -> wake dxo
It worked fine as root. A clear case for setuid:
=== root@eureka (/dev/pts/20) ~ 7 -> chmod 4555 /usr/sbin/wake
=== grog@eureka (/dev/pts/11) ~ 43 -> wake dxo
=== grog@eureka (/dev/pts/11) ~ 44 -> ping -c 1 dxo
|
More Coursera courses
|
Topic: technology, opinion | Link here |
I'm still making heavy weather of the linear algebra course from Coursera that I started at the beginning of the month, and I'm devoting more time to it than I expected. So the last thing I need is another course, right? But I have signed up for no less than three others, in the hope that they'll be much easier.
Today the first one started, Introduction to Mathematical philosophy from the Ludwig-Maximilian-Universität in München. I was right: it's much easier, but quite interesting, and it might help me understand what the other course is on about. Hopefully the others (A brief history of humankind from the Hebrew University of Jerusalem and Climate Change from the University of Melbourne) will also be easier. In passing, it's interesting how international these courses are.
| Wednesday, 31 July 2013 | Dereel | |
| Top of page | ||
| previous day |
|
Google: don't be evil?
|
Topic: technology, opinion | Link here |
Like many people, I've been watching the increasing influence of Google with a mixture of admiration and concern. How can a company of that size remain true to its motto? And so I get more concerned every time I see something pointing away from this premise.
Today I read an article in Wired about unnecessary restrictions on use of Google Fiber. It seems that the terms of service prohibit “servers”, whatever they may be. The article goes on to assume evil intent behind these limitations.
That's possible, but the article doesn't make it plausible enough. They forget Hanlon's razor. What's a server? We know, of course, and even the article finally gets beyond the initial statement (with dead link):
The problem is that a server, by definition, doesn’t have to be a dedicated expensive computer. Any PC or Mac can be a server, as can all sorts of computing devices.
Of course, a server doesn't have to be hardware at all. But even the correct link doesn't make that very clear. Can you be sure that the legal people who put together the terms and conditions understand that?
There are two questions here: what are they prohibiting, and why? The Wired article quotes a submission from Google which contains a reference to the terms of service:
Your Google Fiber account is for your use and the reasonable use of your guests. Unless you have a written agreement with Google Fiber permitting you do so, you should not host any type of server using your Google Fiber connection, use your Google Fiber account to provide a large number of people with Internet access, or use your Google Fiber account to provide commercial services to third parties (including, but not limited to, selling Internet access to third parties).
But that's not what I read now, and it's instructive to include the heading in the quote:
Resale and Redistribution
The Fiber Services are intended for the personal use of you and others with whom you share your residence (including, within reason, guests who are visiting you). You agree not to resell or repackage the Services for use by people other than those with whom you share your residence. If you wish to use the Google Fiber Services to provide Internet service to others, you must enter a separate agreement with Google Fiber that specifically authorizes you to do so.
So it seems that Google has changed the wording since the article was written. It doesn't just remove the reference to “servers”: it also tidies up the text in other areas. That makes sense, and it supports my argument about Hanlon's razor. The concept of “server”, especially in the Microsoft space, is badly understood. If this incident has helped Google improve the intelligibility of their legalese, all the better.
And the reason? Under the circumstances, that's moot, but given the context, it seems clear that this clause distinguishes end-user use from commercial resale. I don't see that limitation as being “evil”. And even the old wording didn't explicitly prohibit “... running their own mail server, using a remotely accessible media server, SSHing into a home computer from work to retrieve files, ...”. That prohibition relies on an agreement on the meaning of the word “server”. I certainly wouldn't like to rely on the interpretation, but the changes to the policy seem to bear it out.
Of course, there are lots of people who see the motto as “Don't be evil”, as
borne out by my initial search. But that's at least partially human nature.
|
Sausages without splitting
|
Topic: food and drink, opinion | Link here |
We've been making our own Thüringer Rostbratwurst for several months now, and we're quite happy with the taste. The appearance is another matter: every time so far the skins of at least some of the sausages split.
I've gone through the potential reasons before, and tried several solutions, but the results are always the same. Today I first half-cooked them in the microwave oven until they were firm, about 5 minutes on 330 W for 500 g of sausage, and then grilled them. And, as always, they split on the underside. Why? Is it the way the skins shrink when grilled? Is it a problem with the casings? We're running out of these collagen casings, so next time I'll try natural gut casings.
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 |