procrastination diagram - December 2003

So, perhaps in consonance with my earlier outburst of the holiday spirit, while eating dinner with myself, Garry and Marc at an expensive restaurant, Alexis and Assar had their house broken into. Their cats are all right, they should be adequately insured. Now they just have to get over their sense of violation.

Of course, you probably have one question: how was the food? Very good, although I wasn't paying attention and probably consumed my year's allowance of cream and other dairy fat, as well as my food budget for the rest of the week.

Anyhow, I suppose that's one advantage of living in a house with lots of loosely affiliated people with desperately random sleep schedules. Which is my clumsy segue into the USB alarm clock (via Gizmodo) I think I want one, if only to encourage companies that are busy hooking everything with a microcontroller in it to a USB port. Of course, I wonder if you can set the time from the computer? :-)

The other link I want to drop is an article from Security Focus about a guy that stole a copy of a huge consumer information database. People are worrying these days about identity theft near as much as the real sorts of them mentioned above, and don't stop to consider that someone in general doesn't need to come in contact with you at all to obtain that sort of information.

Bah Humbug.

So, I'm flying down to North Carolina so I can visit with my parents (a little) for Christmas. I'm actually writing this on the plane, and I'd like to share some observations:

• Flying on the twenty-third of December during an Orange Alert seems to mean surprisingly short security lines. As such, I got in disgustingly early for when my flight was originally supposed to leave. (It was further delayed an hour while they flew our plane to Boston from Philadelphia.)

• Exit rows in an American Eagle Embraer ERJ145 rock. At least twice as much legroom as the other seats. I am using my laptop more or less comfortably on an airplane for the first time ever, which is surprising since the whole plane is the size of a postage stamp.

• I miss having a laptop small enough that I can hibernate it and stick it in the seat-back pocket in front of me. Come to think of it, I miss having a laptop that I can hibernate.

• When you are building a new satellite terminal for your commuter airline, kind of like American Eagle's B30A-J gates at Logan, to pick an example not at all at random, if you don't want your customers to think you are petty and/or stupid, install more than a single two-gang power outlet, and install them somewhere other than up near the ceiling. While I was sitting and using my laptop (dialed in via GPRS (airlines, you should install WiFi while you're at it, but I realize that's a bit harder)) for an hour or two via the single outlet, two other people came by and charged their phones. There's obviously a desire for this, and accommodating people won't cost you more than about a hundred dollars in recurring costs over the next few years, and will probably affect your ticket sales more than that as people will be happier with your brand and less likely to defect to airlines whose logos are splattered all over terminals with power (of course, now that you failed to put the outlets in when you were building it, it'll cost you a few thousand dollars to put them in, but that's probably still worth it). Sam suggested that airlines might prefer you to buy an Admiral's Club membership, but I'll bet on average that most the people who want to charge their phones and are flying on regional jets probably aren't the sort of people who fly enough for that to make sense.

Sorry, that turned into a rant. Anyway, have yourself a merry season of greed.

Things like this are why revision control systems should be mandatory in free software projects. It would be nice if we weren't relying entirely on Linus's memory.

Wow. Josh Schacter does it again, this time with a social bookmark manager. I haven't quite figured out what it is yet, but it's really cool. (via corante)

I've been feeling the urge to explain these pages for a while. I suppose it's also a random excuse to generate content, but now that I've raised the question...

One big part of it, at Mark's urging, but I agree with him, is to get writing practice. At work, and at other contexts, my written communication skills aren't what they could be. They're probably acceptable once I produce something, but actually getting me to do the writing is difficult. I will use any excuse to put it off, and if I've gotten past that (which is rare) I often spend a fair amount of time staring at an empty buffer if I don't have a good idea of what I'm going to say before hand. Of course, if I do have something to say and I can get myself to do it, I can often blather incessantly. It's not like I'm a slow typist.

Another thing, which hadn't occurred to me at the time I started, is keeping people I don't talk to as much as I should reassured that I still exist. My mother (Hi, Mom!) will be happy to see this entry, for instance.

This also serves as a place to note facts and links to things with more context than a bookmark file somewhere where I won't lose it. [Paper] notebooks? They disappear. Random files in my home directory get lost and pushed into successive corners. This, doing double duty as my web page, is at least something I won't lose.

And, of course, everybody's doing it. I'm not going to flatter myself into thinking that many people are ever going to read this, or that I might become a Dave Winer or InstaPundit, but I might make my small contribution to this distributed reinvention of the BBS with more uniformly competent writing. There is also small incremental value in participating in Google's global reputation server input set.

I could've used livejournal, but the temptation to abandon would be much greater, and using livejournal is sort of embarrassing for someone with the resources to run his own machine. I'm writing my own software because it's a diversion, and nothing out there did precisely what I wanted, and it probably wasn't any more work to adapt other people's code than to write my own. (I'll just keep telling myself that...)

And I'm really not writing for this as much as I should, to achieve the above ends. I actually have a collection of links that I want to blather about, I just seed to sit down and assemble them into entries. And on consideration, there are various technical ways I could make things easier. And, of course, I'm going to tell you about them (see above, about a place to write things down):

At the moment, I create an entry by editing it in emacs in psgml-html-mode, as part of an html file that gets thrown away (this in theory provides somewhat correct html, and provide a good opportunity to remember to run a spelling checker). Then I log into my server, and run a command-line utility and paste the entry into it. Then I run the command to regenerate the html files. Then I reread the (live) results in a web browser, and invariably find grammar errors, typos and paste-os that made it past ispell. Then, if I remember, I hit the bookmarks that ping technorati and weblogs.com.

Causes the phrase "fraught with peril" to come to mind, don't it?

What I want is something that starts with an emacs buffer running something like psgml-html-mode, but that does some amount of validation without needing <html> and <body> tags (the ability to cleverly hide the attributes of, say, <a> tags would nice too). Automatically turning on flyspell-mode is simple and easy to forget. When I hit C-c C-c, it should do an extra-paranoia xhtml validation step, then ship the text somewhere that causes it to get rendered like my weblog in whatever web browser I have handy. Then I should have a "publish" button that causes the entry to written to the final output location, ping the meta-sites, and eventually do auto-trackback.

All this should reduce the friction for entries like this one as much as possible. For link dumps and their ilk, I want a "blog this" button in my browser and my RSS reader. It should let me type in a small amount of accompanying text (in a dialog, or on a web page) and stash the results and whatever other metadata is has handy somewhere. This information could then be retrieved into the above utility and processed into a proper entry.

Ooh, stories on code to write. Now I just need to find the time. I should at least get another entry out of the details. :-)

Driving to the institute for my MITSFS hours today, I was struck by how cold and dead everything looks with a coating of rain-refrozen snow on everything. You'd think it would look dead enough in uncovered late-fall mode, with dead brown leaves and bare trees, but the ice covers most of the smaller signs of life or life recently departed, leaving only dead looking trees and mounds of featureless white.

I now understand why, in some cultures, white is the color of death.

Of course, that white will be ugly mottled grey/brown in a few days, because I live in the city, and of course everything will spring back to life in a few months, but the actual winter will have to start and end first. And it's going to be cold, and long, and cold, and depressing, and cold, and lonely, and cold, and did I mention cold?

I don't hate all winter weather. Snowstorms are really nice, except for the shovelling part. And it's not going to stay that cold just yet, it's even supposed to get into the mid-40s for the next couple of days. But it might not get above 30 during February. And the anticipation is the worst part.

Dave Winer suggests a complex scheme (not unlike the Radio coffee mug links) involving localhost urls and arbitrary ports for OS rss subscription management.

This seems overly complex, especially if you've got control of the browser. Much simpler would be for the browser to recognize an rss file (by mime type, or possibly involving some slightly cleverer handling of xml) and hand the URL off to your designated subscription manager. I'm pretty sure I could do this now with only configuration files (other than adding command line subscription support to my rss reader).

(Yes, I've noticed that my syndicate link comes back as text/plain. I expect to be fixing that in the near future. :-)

Mark convinced me that link dumps serve a useful purpose. So, here goes:

• The Dead Media Project
• Alexis's Dead Fish Page Circa 1999
• How to break out of a chroot() jail
• 2004 Spam Conference Registration
• Paul Boutin's Will the Broadcast Flag Break Your TiVo?
• pyrple (python rdf api)
• IBM Research's Reinventing Email

Inspired by Simon Willison's code to HTMLify user input, converting urls into links, I wrote a hundred lines of python today, intending to drop them into the comment system. Only problem is, having written it, I don't trust it. I know there are bugs in there, code being code. Because I don't know, specifically, what those bugs are, much less understand their implications, I am reluctant to actually deploy the code.

As it happens, my first feelings of mistrust are fading as time passes.. but I think I need to figure out whether this is because I'm deciding that the potential problems are not that great or because I've already distancing myself from the code while shoveling snow and playing Mario Kart.

Basically, you can class the potential problems thusly:

1. Bugs that allow a malicious attacker to execute code of their choice on my system. (extremely unlikely)
2. Bugs that allow a malicious attacker to cause viewer's browsers to execute code (javascript) on their system. (unlikely, I think)
3. Bugs that produce ugly comments. (likely)

Certainly the issues are weighted towards the latter, but the first two issues are sufficiently serious or embarrassing that I think I need to let the issue simmer for a while.

Once upon a time, someone posted an absolutely adorable picture of a kitten leaping/falling off of a windowsill to ratemykitten. Then people started photoshopping it...

Thanks to perbac for the pointer.

link

I've decided to go with a old-school single number version scheme. So this is bleat 3, the next "release" will be bleat 4, etc.

Audience visible changes:

• comments

And that's about it. Under the hood, the templating system got the first of what I'm sure will be several major revamps, the database schema was shaken (not stirred) slightly, entries can have arbitrary text attributes, and there's a cgi script (which is used by the comments) but will mutate in the future to rule several provinces.

Immediate future plans: trackback (should be easy with current scheme of things) and comments with actual formatting other than the current "what you type is what you see" setup.