Posts in the Tech Category

Things are relatively quiet for the first time in several weeks.  Outside, there’s close to a foot of powdery snow covering everything, which was fairly easy to clean off the driveway—it’s light enough that I used a push-broom instead of a shovel.  Inside, Kat and I have been enjoying hot cider in front of our fireplace and reconnecting with each other after all the stress of the last few weeks.  Occasionally I play with Gravity, the household cat.  It’s a markedly more peaceful mode of living, and I’m enjoying it while I can, because it won’t last.

It seems like there were things I was going to post, really cool stuff, but it all got buried while I was off at conferences, memorial services, and so forth.  I did notice that Tantek has redesigned his weblog, and the new look was broken in Mozilla for a few hours.  It’s fixed now, but I wonder if that was due to him working around browser bugs, or just tightening up his CSS?  Knowing Tantek, it could very easily be either one.  Regardless, it’s a very interesting design; very paperish.

I’d dig through my e-mail for more stuff, but the fireplace is softly calling my name, and I hate to disappoint anything that could theoretically burn down my house.

• Digging Out was published on Wednesday, December 4th, 2002.
• It was assigned to the Personal and Redesign categories.
• There have been no replies.

As I work my way through an enormous backlog of e-mail, I found a message from last month telling me about a utility called pngslice.  Apparently J.J. Green‘s girlfriend was really impressed by the Ragged Float demo but didn’t want to spend her time slicing stuff up in Photoshop.  So, in the manner of good boyfriends everywhere, J.J. stepped up to help out by writing a Unix utility and then released it to the world.  Better than flowers, I tell you.

Ooh la la: “Faites bonne impression avec les CSS,” a translation of “Going to Print.”  Thanks to Stephanie Booth and Samuel Latchman for helping my work get en Français!

• Slice ‘n’ Dice was published on Saturday, November 23rd, 2002.
• It was assigned to the CSS and Writing categories.
• There have been no replies.

Meryl Evans has recreated the missing WebNouveau list of tableless CSS site on her own Web site, and is looking for other CSS resources.

Security through obscurity never really works; when I use it, I at least know what I’m doing, and that it could bite me.  I had thought any halfway informed administrator knew that same thing.  Apparently, some folks still don’t get it.  Let’s see… a company puts a file on its Web site that isn’t in any way protected except in the sense that there are no links to the file, and someone else figures out the URL, which leads to advance publication of the information.  That’s not breaking into your site, it’s being smarter than you.  From Intentia’s own press release: “The incident has severely damaged confidence in us as individuals and in Intentia as a company,’ says Björn Algkvist, CEO of Intentia International AB.”  That’s almost certainly true.  I know I wouldn’t trust my company’s data to a firm that made so obvious a mistake.

• A Rescued Resource was published on Wednesday, October 30th, 2002.
• It was assigned to the CSS and Standards categories.
• There have been no replies.

The W3C has released a public beta of a major upgrade to their HTML validator, and authors are very much encouraged to try it out—I did!  It adds a lot of new things to the service, including better handling of document MIME types (like application/xhtml+xml) and more.  One of the validator’s lead developers e-mailed me about it, and then dropped this comment in at the end of the message:

The new CSS is loosely inspired by your stuff (well, that’s probably true for most CSS these days I suppose).

Pardon me while my ego starts beating its chest and howling.  (And Kat’s not home to keep it under control, either.)

• (X)HTML Validator Upgrade in Beta was published on Wednesday, October 23rd, 2002.
• It was assigned to the CSS, Personal, Standards, and W3C categories.
• There have been no replies.

Wired News has redesigned their site, and not just the front end, either: the really important stuff all happened behind the scenes.  Using no tables to lay out the page, but instead applying CSS to XHTML, the site is a stunning example of how standards can be made to work today.  They have an article with some details (and a few quotes from yours truly).

There are a few flies in this ointment, but they’re fairly understandable.  The pages don’t always validate, in part because of third-party advertisement code (which is notoriously horrible) and in part because converting seven-plus years of pages isn’t a simple task.  Actually, most of their validation errors seem to involve unencoded ampersands in URLs, which ought to be easy enough to fix.

The Web Standards Project calls this a gutsy move, and I agree.  A site with their kind of traffic has to make a big commitment to do something like this and stick to it.  The management of Wired is to be applauded for approving this move, and the men behind the scenes deserve even more applause for their work.  Look for a DevEdge article soon where we interview Douglas Bowman, the point man on the Wired redesign.

• Wired With Standards was published on Friday, October 11th, 2002.
• It was assigned to the CSS and Standards categories.
• There have been no replies.

In all the head-pounding over learning XSLT last week, I let some things slide by without comment, so I’ll try to cover them all in a single post.  (And remember, if you have an RSS aggregator, you can syndicate these posts via my RSS feed!)

In early November, I’ll be appearing at Meet The Makers New York on a “standards mini-panel” with Jeffrey Zeldman, so I’d better get around to calling Moishe.  There will also be a San Francisco Meet The Makers where my co-worker Arun will be on a panel with Tantek Çelik of Microsoft.  You might be able to score a free VIP ticket to either event if you hurry (and are willing to fill out the questionnaire).

I’ve added more information to the upcoming events on my Speaking page, including promotional codes for events that have them.  I disclose when using a code will make me money, and have been thinking about ways to turn those into community-building exercises.  Maybe I’ll take everyone who used my code(s) to a group dinner, assuming I can come up with a way to verify code use.

Last week, we published a CSS2.1 Quick Reference sidebar tab for Gecko-based browsers, and French translations of the CSS2 and DOM2 sidebar tabs, to the Sidebars area of the DevEdge Toolbox.  I also published a technical note on fixing list-item marker size in the NS6.x series.

Over the weekend, I not only dug into more XSLT (which almost made me pound my head against a wall, again), but I wrote some Javascript bookmarklets to help manage the administration of css-discuss.  It’s been a while since I thought of myself as a programmer, and I certainly am no expert—but it’s been good to stretch those mental muscles again, after so long.  The neural paths needed for exploring and using CSS and structural markup aren’t the same as those needed for programming.  The sense of achievement I felt when I figured out how to do what I wanted to do was a welcome change of pace.

It’s really cold in our house right now, but at least the shaking and banging of workmen dismantling our 82-year-old boiler has stopped.  Kat and I are sort of sad to see the old beast go, but since it had suddenly started leaking enough carbon monoxide to form its own atmospheric system, we don’t exactly regret replacing it.  The replacement boiler is almost ridiculously smaller than our old boiler.  I have trouble believing that it can heat the basement, let alone the whole house.

• Catching Up was published on Monday, October 7th, 2002.
• It was assigned to the CSS, Speaking, Standards, and Travel categories.
• There have been no replies.

Recently, at the urging of a well-known Web dude, I added “permalinks” to the entries here.  That’s what the little paragraph symbols are all about.  At that point, I figured that I ought to drop all of my past posts into some kind of system that could automatically generate said “permalinks,” as well as allow me to finally publish an RSS feed.

Most people would say “I’ll use Blogger!” or “time to install Moveable Type!” but I didn’t really want to run a CMS package just to store and syndicate my random mutterings.  So I of course did the obvious thing, and decided to pour the whole archive into XML, using my own private markup, and then teach myself XSLT so I could take that archive and produce whatever output I wanted—the home page’s five most recent entries, the permalink archives, an RSS feed, whatever.  Okay, there was an actual work-related reason: learning XSLT will make me better able to help manage DevEdge, which is driven by XML, XSLT, and CSS.  But mostly I was doing my usual thing, and effectively learning how to fly a helicopter just to get to the corner grocery store.

So here it is!  You syndicators can pick up the RSS feed if you like, and be automatically informed whenever I work up the energy to write another entry.  Aren’t you lucky?  I know you’ve just been dying to see this happen.  I even went to the effort of writing entry titles, which I didn’t have before, just to make the feed content look better.  Who loves ya?

Now I’ll probably decide to teach myself HTTP and Perl just to build my own skeletal RSS aggregator.  Yeah, that sounds reasonable…

• A Transformation In My Thinking was published on Friday, October 4th, 2002.
• It was assigned to the Tech category.
• There have been no replies.

A point of followup on the KPMG fix: It turns out that the fix works almost completely in Opera 6/Win, even when it identifies itself as Opera (as my copy does).  The little yellow-box navbar thing zips along quite speedily, but the drop-downs for “Search,” “Contact Us,” and “Country Selector” can be really slow, while other times very zippy.  Also, one of the “close dropdown” buttons doesn’t work.  I don’t know why, but I suspect these are easy to fix.

Here’s the kicker: I didn’t do any Opera testing until this afternoon.  As I carried out my fixes, I didn’t make one single coding change with Opera in mind, and yet the page is 95% fixed for that browser.  That’s the whole point of using standards—you can be almost completely browser agnostic.  The other 5% that doesn’t work in Opera is probably due to either a DOM bug in Opera—no browser is perfect—or (more likely) my fumbling attempts to get the code based on the W3C DOM wasn’t a complete job, and I left some non-standards stuff in there.

Will it work in Konqueror?  In OmniWeb?  I don’t know, but if they support the correct W3C standards, then the answer is “yes.”  It’s the same answer for any browser that supports Web standards.

I’d also like to reiterate for those of you planning to dig into the source of the fix that it’s not an example of completely  standards compliant design.  It’s merely an example of how one person, with a modicum of effort, was able to take an outdated design method and hack in some semblance of standards support in order to take a broken site and make it work in multiple browsers.  It isn’t perfect, but maybe it’s a start.  Share and enjoy.

• KPMG Revisited was published on Monday, September 23rd, 2002.
• It was assigned to the Standards category.
• There have been no replies.