This is as days pass by, by Stuart Langridge

Since bloody Mozilla Thunderbird seems to (a) be screwing up on me something vicious (if you sent me a mail and were expecting a response, you might want to send it again) and (b) won’t let you filter messages on body content (why? why? why?), I’ve added a .forward file that looks like this:

# Exim filter
if $message_body matches "name="[^"]+\.zip"" or
  $message_body matches "name="[^"]+\.scr"" or
  $message_body matches "name="[^"]+\.exe"" or
  $message_body matches "name="[^"]+\.pif""
then
  save /home/aquarius/mail/Dodgy-Attachments
endif

That should kill off most of the virus mail I get.

My firm, Mills & Reeve, is up for an award for its intranet. Now, I designed and built the intranet, and I’m really rather proud of it. So I thought we were in with a good chance of winning said award on technical merit. However, it turns out that the process for this award is that it’s awarded based on number of votes cast, not on any kind of judgement as far as I can tell. (I might be wrong about this, but I don’t think I am.) Needless to say, I was somewhat disillusioned by this; we’re not a very big firm, so it’s unlikely we’d be able to muster up lots of support from our clients or suppliers like some of the larger firms can. Again, the underdog gets stamped on. So, here’s my plan. The M&R intranet, because I built it, was built entirely around open technologies; it’s shipping XML around, it’s using structural semantic HTML, it’s styled with CSS (and the css-design mailing list helped me with some of the finer details!), it’s cross-browser. It’s not the typical works-in-IE-only, font-tags-everywhere, JavaScript-heavy design, it’s not using some horrendous overpriced “portal” product or content management system. It’s all as open as I can make it. It’s a critical part of the firm’s knowledge management, and it really helped to explain to people here that we can build systems without payng consultants to do it or buying boxed products for a fortune. While I appreciate that none of you here have seen this system, neither has anyone else outside the firm—it’s our intranet, after all. So the fact that there’s a category in this award for intranets obviously means that people are supposed to vote on it sight unseen. Leaving aside how silly this is, my plea is this: if you support systems being developed using open technologies and modern web design techniques, and you’re opposed to the perception that companies must buy overpriced overengineered overcomplex “software solutions” to do good work, then vote for Descport, the Mills & Reeve intranet, in the “Best knowledge management / information portal” section of the 2004 Inbrief e-Loties awards. There’s also more detail on the awards themselves, and the shortlist of candidates for each award.
If your reaction here is indignation that I might even consider asking you to do such a thing, then I pretty much agree with you, and I certainly won’t take it askance if you refuse entirely and never read here again. My only defence is, “this is how this award is won“. If it were on pure technical merit, I’d be disappointed if we hadn’t already won.

So, I’ve had xmule recommended. Herewith, some notes as I start it up for the first time. I’m apparently running v1.8.4, from the Debian unstable archive.

Ooh, “for system tray integration to work, you must specify which desktop you are using“. On the one hand, it’s nice to see that they care about integration. Also, Gnome 2.x is already selected. On the other hand, it might be selected because it’s first on the list rather than because it’s what I’m running, and secondly the whole damned point of the freedesktop.org stuff is to make system trays work fine with either desktop, no? Nothing else—Gaim, Skype, Rhythmbox—has ever had to ask. Not good.

Ah. This happened to me last time I tried a P2P client; there are no servers in the server list, so I can’t connect. How do I get a list of servers? Ah, “update server.met from URL“, and an “Update” button. Good stuff. On the other hand, why doesn’t it do this by default the first time you load it? Not good.

“You have a lowid. Please review your network config and/or your settings“. Wonder what that is? I suspect that this is a “forward a port on your firewall” issue. (fx: reads @/usr/share/doc/xmule/README.gz@) Ah-ha! “If you are behind a firewall or router, be sure that ports 4662 (TCP) and 4672 (UDP) are unblocked, or you will always receive a low ID.” Don’t know what a “low ID” is, but I suspect it means “people can’t connect to you“. I’ll forward the ports.

And searching seems to work! Good stuff!

Stone me, there are a lot of buttons here. Does anyone really want more from a P2P application than a textbox to enter searches and a listbox for clickable results, I wonder? And probably one “preferences” window, which doesn’t need very much in it at all. I am spoiled by Gnome, I think. Not good.

There’s a “method” dropdown for searches, with things like “FileDonkey (Web)” in it. No idea what that means.

OK, I lied, you need a separate bit showing you what you’re downloading, perhaps. Starting a download seems to work. Good stuff. No progress being made on said download, though. The thing I’m downloading has “Availability” of 2, which I assume means “two people have it“?

So, how do I hide the window, just leaving the notification area icon? (NB, developers: it’s not a system tray! It wasn’t even called that under Windows! It’s the notification area.) If I click the cross, will it shut down the application? Blimey, the notification-area-icon menu is a bit odd, isn’t it? Not at all a standard looking menu. And “Hide” in it doesn’t work, either, which isn’t all that promising. Not good.

Bugger. Clicking the cross closes the whole application. Not good. Restarting, it seems to have remembered downloads in progress, though, although that doesn’t deserve a point because it’s supposed to do that.

So, so far, more “not good“s than “good stuff“s. Investigation will continue.

-----

A few people (Richard for one) have been asking about DOM scripting resources; specifically, whether there are any decent books out there which cover modern JavaScript and the DOM in detail (rather than JS-as-she-was, with image rollover scripts and onclick handlers in the code and other obtrusive DHTML techniques, which we don’t like). Well, there certainly will be one before Christmas this year, because I’m writing it. Due for publication around November time. I’m not quite sure how much I’m allowed to say about it, but I think it’s pretty good so far (I’m about halfway through it). Of course, I would say that, so all I can do is invite you to buy a copy for yourselves when published and make your own judgement. If enough of you do this then I will happily do a weblog entry from Barbados saying thanks, I promise.

So, I was at tEC7 and I am now back. It was sunny in Durham (!), but there ain’t no sunshine here. It was a really excellent weekend. Roll on the next one! I’m going to go and see if anyone’s done a meet report :-)

-----

Linux.Ars at Ars Technica currently have up a guide to Mono (no permalinks! if it’s not there, then it might be at this link), going on about how stuff will run on Windows and Linux, and demonstrating how you can do a regular expression matcher in a few lines of code, and a Gtk# app in a similarly short amount. I have two questions, a small one and a big one. The small one is: that Gtk app running on Windows—does it do it if you haven’t installed Gtk on Windows? I bet it doesn’t. Wouldn’t it have been more useful to demonstrate an application which used the native widget set on each platform and didn’t require extra installation stuff?
The big question is: why is this in any way better than Python? Python has all the benefits that they’ve trumpeted about Mono in that article: it’ll run on Windows and Linux (and the Mac, which Mono won’t, not properly), it’s got bindings to a shitload of stuff that Mono hasn’t, it’s got a ton of libraries with it for XML and regular expressions and everything that .net has, and it’s not owned by MS. Yes, yes, perhaps they are going to keep it an open standard and not have some way of using it to bolster their OS monopoly at the expense of Linux. Then again, perhaps they’re not. Since it’s potentially the future of Linux we’re talking about here, I think I’d prefer to play it safe and not use .net. Other people may disagree. So, those of you who think it’s a good idea: please can someone explain to me why it’s better than Python?

Ahem! http://www.livejournal.com/users/kryogenix_blog/
Where did that come from?
I mean, I’m not entirely opposed to the idea; this is what RSS is for, after all, and quite a few people have asked me to syndicate my site on LJ and I didn’t bother because it was too much effort. Someone’s saved me that effort, so OK, thanks. I just wonder who it was!

So, I thought, I gotta actually get into Project Utopia rather than just talking about how good it is. You’d think something as potentially revolutionary as this would be difficult to get installed, right?

sudo apt-get install kernel-image-2.6.5-k7 udev hal gnome-volume-manager

That was it. I love Debian. And Gnome.

Create a new mail in Outlook. Look at its message ID. (You’ll need Outlook Redemption for this.) Then send it. Get the recipient to look at the message ID of the message they receive. (You can even send it to yourself, and check yourself.) Now, they’re the same message, yes? So they have the same message ID, yes?
Nope. They’re different.
Why? Why in Christ’s name would you do this? They’re the same bloody message! Why have they got different message IDs? For God’s sake!
I am very annoyed by this. That means that Outlook hands off the message to Exchange, with a message ID in it, and Exchange throws away that message ID and puts a different one on it. Why the hell would that ever be a good idea? Bloody hell!
One day I’m going to find the guy who wrote Outlook and I’m going to scream at him for things like this.

OK, so adpb has said for ages that comments are Textile parsed, and there’s a JavaScript preview. I never released the JS Textile code, but Sam Newman hassled me for it and released it himself, and good on him. It’s been broken here for a while, and I’ve finally got around to fixing it. Moreover, it turns out that comments were never and have never been Textile parsed; the comment preview was not what you got when the comment was posted. For example, links weren’t done, as can be seen on some old posts. That’s now fixed. I think.

Fuck me, today has been a long day. yawn

When I was a kid, I thought that the epitome of being a groen-up, being a man, was how my dad would take a new, unopen bottle of milk from the fridge and shake it in one hand to mix the cream before opening it. I could never do that; my hands were too small. Shaking a milk-bottle was a two-handed job all the way, involving serious concentration to avoid the bottle slipping out from between my hands and smashing into splinters on the floor. That only happened once: I distinctly remember crying about it, although my parents were decent enough to not make the obvious joke. But the way my dad could just pick up a bottle and shake it about in one hand while his attention was on something else, hand oscillating back-and-forth almost as an afterthought, summed up being adult in my seven-year-old eyes. I’d wish I could do it.

Now, of course, I am a grown-up. I’ve got a car and a mortgage. And I can shake a bottle of milk in one hand. But I don’t need to, because I only have skimmed milk, and there’s no cream to mix.

There’s a message here, I’m sure.

Andrew Clover:

But customers—other large organisations—love this. They would, apparently, much rather spend millions of dollars on an over-specified ‘platform’ and a team of horrifically expensive consultants to attempt to build an application out of it (eventually succeeding months behind schedule and with half the feature set missing) than simply to get a few in-house programmer nerds to hack up a bespoke system in half the time and one hundredth of the cost. I certainly could have done such myself, and I’m hardly the world’s brightest programmer. (No, really.)

Tom’s been talking about metadata as it relates to libraries and cataloguers. I don’t know how much contact there is between web metadata people and library metadata people, but I get the distinct impression, having talked to him a bit, that a lot of the problems that people struggle with in our sphere were all solved a long time ago in theirs. This is the same as Dorothea Salo’s common complaint that some of the problems that the web are struggling with were solved a long time ago in the markup world, too. I don’t know how true this all is; there’s a big case of Not Invented Here around the web, true, and there are altogether too many people talking about “forging a new path through uncharted territory” and suchlike when it is all charted: we just haven’t bothered to look for the charts. On the other hand, I think that the people who are involved in all this stuff, the Sam Rubys and Tim Brays of this world, do know about all this prior art and they’re busy weaving it all together. I trust them to get it right.
On a different note, one suggested title for an “advocate of structured metadata” (quite why you’d want to advocate a different type of metadata, or no metadata, escapes me) was “metaphile“, which was rejected (among other reasons) because it might mean “one who is beyond being a lover“, “a lover of change“, or “King Aegeas“, who was married to Meta. That last one made me laugh, although I should probably be embarrassed about laughing at librarians’ jokes.