Monthly Archives: October 2007

Fixing mod_proxy

I just realised, I hit a kind of milestone in my work on fixing the finer points of HTTP protocol compliance in mod_proxy.  With my today’s fix to PR#41798, all the dependencies of my tracking bug PR#43454 are now either fully fixed, or fixed in /trunk/ and just awaiting sufficient review from my fellow developers to backport.

I’m confidently predicting a much-improved proxy in Apache 2.2.7, over the current version (2.2.6) or any earlier version.

I shall now make a celebratory cuppa tea to take to bed with me.  Goodnight.

Advertisements

Black mark to Ubuntu

So, I power up the desktop (from hibernation) this morning.  And it’s cocked up the time change away from Daylight Saving Time.

Not just a simple “got it wrong”, but the system clock has been moved the hour.  So it gives me the same time as my watch (with no attempt at automation) says, but calls it “GMT”.  That’s seriously wrong!

Last time I saw such breakage was in Windows 95.  I’ve taken great care not to let Windows try to do anything complicated (like, say, tell any time other than GMT, at any time of year) since then.  To see Ubuntu doing it is deeply shocking.

Humbug is here

Went to a lunchtime concert: Devon Sinfonietta at the United Reformed Church, playing Haydn and Gershwin.  This is the first time I’ve heard of the orchestra, and they turned out to be a youth group.  Or less charitably, a fancy school orchestra, though I think in practice they are selected players from a number of schools.  Anyway, I quite enjoyed the music.  Worth going to on a wet saturday afternoon ten minutes walk away, but not too much more.

The church is much smaller inside than it looks from the outside, but pleasant  and comfortable.  That’s a contrast to Tavistock’s main church, which is much bigger and gets the top performers (such as the Devon Baroque last week) but a is terrible place to sit, even if you get one of the relatively-few seats with a decent view.

But more disturbingly, the wretched place has “christmas” crap, and big banners outside advertising the sale of christmas cards.  Doesn’t the church notionally follow the teachings of a man who took violent direct action against merchants trading in a temple?  I’d have been happy to forgive them that, but NOT this humbuggery before at least December.  And it’s only bloody October, ferchrissake!

Bah, Humbug.

Getting our river back

Now that the summer weather is past, we appear to be getting our river back.  Over this week, much of the building crap has been removed. There are what looks like some permanent changes: a third level of weir, and a new concrete construction around where the water gets siphoned off for the canal. The canal is also looking healthier, though that could be in part because it was dark (well, deep twilight) when I last looked, and the canal definitely looks it’s best (beautiful) in those conditions.

Hopefully they’ll reopen the path within a few days. They’ve now had the full 90 days closure they claimed, so we bloomin’ well should get it back. No excuse for delay, either: the weather has been very good during the time it’s been closed, in contrast to the wet weather earlier in the summer.

Galileo and the Inquisition (1)

A couple of recent events demonstrate chillingly that the human dynamic that drove the Inquisition, and ruthlessly suppressed the scientist Galileo, is alive and well in the UK today. The subjects are of course different: astronomy and even cosmology no longer violate deep taboos.

I’m too tired tonight to do justice to the Big One, so let’s just have a little rant about a normal-size scandal. Nothing abnormally controversial or perilous to a blogger. Just another proposed genocide. Call it a prelude. It doesn’t do justice to the title, but what the heck?

OK, the background. There is tuberculosis. It’s no longer a serious problem amongst humans, but cattle suffer from it. Badgers also suffer from it. The hypothesis is that badgers spread it amongst cattle (vice versa is not an issue, because nobody cares). The proposal advanced by the farmers union, representing big, multi-millionaire farmers (not least the landed aristocracy)[1] is that badgers be scapegoated and exterminated.

So far, the question has been left open. A scientific study was conducted over nine years, to determine whether killing badgers would protect cows. It concluded that anything short of a Final Solution would in fact have the opposite effect, not least because it would increase mobility amongst badgers, and hence the likelihood of their carrying infection to cattle that weren’t already infected.

But that’s the “wrong” conclusion for those whose agenda is extermination. Now the government’s chief scientist has spoken in favour of it, throwing up a clear gulf between science and Politicised Science. I find myself somewhat bemused by this, but it reeks of political agenda: if the science doesn’t support our agenda, let’s fudge it (hmm, anyone for Weapons of Mass Destruction?).

An heretical hypothesis

Since the subject of infection falls way outside my expertise, I can only rant. But one thing that looks very relevant is the nature and history of the disease itself. Historically it was a major killer amongst humans. But it was always a disease of poverty. It was the advance of living standards, notably slum clearance, that caused it’s decline to negligible levels in humans. Could it be that the disgraceful conditions in which cattle are kept (for economic reasons, of course), are the real culprit?

That’s an heretical hypothesis. It cannot be discussed anywhere more significant than a blog. Let alone investigated. Because the consequences of the wrong outcome would be unthinkable.

[1] Not all NFU members fit that description. But they’re the ones with influence in high places. Others are incidental.

Things that come around

Several years ago, I tried proposing to Google that they should incorporate accessibility analysis into their search rankings. Their (eventual) reply was, not interested.

I’ve just heard the BBC’s In Touch program, which deals with issues affecting blind and partially-sighted people. Today we had a lengthy interview, with a blind Indian engineer working at Google on exactly that problem. He explained that the accessibility-enhanced search will as first priority select the best/most relevant pages by google’s standard closely-guarded-secret algorithms, but then order those results to ensure that the highest-placed results are accessible.

He even gave some technical details of how the accessibility assessment works. The perennial subject of alt attributes was mentioned (without details on how they assess them), but more interestingly, he referred to well-structured pages, and clearly uses HTML heading markup as a criterion.

It’s all happening very quietly, but it’s gratifying to those of us who have been banging on about this for years. Of course, it would’ve been far better if they’d used Site Valet (customised as necessary to integrate with their systems) for this analysis.

Verifying Ubuntu

I’ve just downloaded an .iso of the new Ubuntu (7.10). Actually, that’s kubuntu, though I understand it’s from the same stable.

With it comes an MD5SUMS file. The MD5 sum of my .iso checks. So far, so good.

Finally, check the MD5SUMS with the PGP key in MD5SUMS.gpg. Unknown key – oops. Import it, try again. No chain of trust – can’t verify. List the sigs: strewth, this is a *tiny* list for such an important key. Import keys of the signatories, and all but two have no bloody signatures on!

Right, Ubuntu’s release signing key has exactly two meaningful signatures. I don’t have an adequate chain of trust to them, but there are some familiar names in their keychains, including several debian.org folks, which I should stand a reasonable chance of verifying independently. But that’s a helluva lot of effort to get even a minimum level of security. Aaargh!

Ubuntu – don’t you believe in security?

MSDN

I have recently received email from Microsoft. They’ve given me an MSDN subscription number, which works to log me on to their site. A bunch of MS developer resources are apparently in the post.

Those who know me know that I’m not a Windows user. I rarely – maybe once in two or three months – boot a computer into windows. I dislike what I get, and see it as good for games (which I often find horribly addictive) but nothing else. So what am I doing with MSDN?

The first part of the answer is, I’m a developer. Specifically, an Apache developer. And Windows is an important platform for Apache and for some of its users. I already support my own Apache modules on Windows, though that doesn’t mean design or programming – only compilation and minimal testing on the assumption that the modules and any bugs therein are platform-neutral. In fact, since my old laptop (with a windows partition) died about 17 months ago, that’s the only thing I’ve used windows for.

But there’s more to it than that. There are good reasons I might want to undertake more complex work on Windows. For example, having written a driver for FreeTDS under Apache DBD, I can hope to maintain it if and only if I have access to the resources I expect the MSDN to bring to me.

The second part of the answer is, MS needs developers like me on-side. They know they command little goodwill amongst developers, and attract a good deal of suspicion.  MS appears to be moving towards open source in some areas, and would presumably like to develop relationships with open source projects that are beneficial or neutral to their own interests.  Apache is firmly vendor-neutral, and has a range of projects of interest to Windows users.

So, a little while ago, Microsoft made an offer to the ASF, to provide us with complimentary MSDN licenses for those of us who may benefit from them.  It is, or should be, mutually beneficial, for reasons that should be clear.

How much use I’ll make of it remains a big unknown.  I expect that any substantial upgrade to my windows-fu will be a big learning curve with a lot of cursing and swearing, and may happen only when a client demands and pays for it.  But at a minimum, I should hope to be able to simplify the labour-intensive chore of compiling modules on windows, and move to compiling my own APR and httpd there.  And with a bit of luck, I might find time to play with the databases we support with DBD.

Royalties

I have my first royalty cheque for The Book, covering sales up to the end of June.  It was waiting for me when I got home yesterday.

Eligible sales were 1300 copies, which I understand is not too bad for a specialised techie book’s first six months.  The good reviews it’s got can’t hurt!

Alas, it’s not just a conventionally-miserly royalty: it’s rendered nearly-worthless by being denominated in US dollars.   A payment that might’ve been worth £2400 at the time I was writing the book is reduced to £1800.  Not that I was expecting to get rich on it, but that kind of drop feels like a third-world currency in crisis.  Worse than getting paid in b***** Lire, as I was for most of the 1990s (when I became a millionaire each month, without ever getting rich).

Still, it’s a satisfying outcome to the effort that’s gone into the project: the fun of writing it, the stress of the editing/publishing process, and the pain of the ITIN epic.

Where political correctness cannot go

Just watched Wozzeck. Alas, only on DVD, but IMHO a very fine performance, from the Wiener Staatsoper. It’s the first time I’ve seen Wozzeck, and it reaffirms my view of Berg as the outstanding avant-garde composer of the inter-war years.

I couldn’t help wondering how, in our times, anyone gets this work past the PC brigade. The story of bullying, adultery/prostitution, and ultimately murder and madness, is clearly adult. Or at least something that would get an intermediate classification if it were a film. Yet the young child is indispensible, and must be a real child: you might substitute a doll in the early scenes, but not playing in the street at the end when they report the discovery of his mother’s body.

Perhaps the fact that the Nazis banned it as degenerate makes it immune from criticism by the Politically Correct?