Geoff Duncan

Repeat visitors to these pages might notice I've shuffled some of the furniture, added a few throw pillows, and generally updated the decor. These changes don't represent a fundamental reworking of Percolating—I assure everyone the site is using the same old decrepit technology that's been driving it for years. I've just done a little cleanup and re-jiggering to address some long-standing issues.

Folks who've been on the Web as long as I have might have noticed that my personal Web pages have had much the same form for over a decade. (Although the Internet Archive's WayBackMachine didn't pick it up until 1998 and doesn't visit very often, its snapshots are still amusing.) My attitude isn't so much "If it ain't broke, don't fix it" as "I don't want things any more complicated!" The pages were simple and worked—more or less—in a wide variety of Web browsers. That was good enough.

However, over the years the pages took on a peculiar architecture. At their core, they relied on principles from the Web's first hurrah back in the late 90's. This was mostly for the sake of backward compatibility, so folks with older browsers and operating systems would be able to view the pages largely as they were intended. Over the top of this, I'd gradually been draping a few technologies which were (gasp!) only a few years old. These were mostly a convenience for me, but also a way to spiff things up for folks who aren't as behind the times as I like to pretend to be.

However, two things spurred me to rework the pages. First, I'd been trying to chase down a problem with the systems which drive Percolating. Every so often I'd write an item and zip off in a hurry to the next thing, only to notice days (or, ack, weeks!) later that the item never made it "live" to the Web site. I'd look in the database, examine the logs, click around, find everything in order, say "huh!," and scratch my head—and typically everything would correct itself. But when the piece on Ray Evans failed to appear, I decided it was time to find and finally fix the problem.

Second, the nature of Web browsing today was tending to show my pages poorly—or, at least, the pages would often be stretched to impossible widths like so much Silly Putty, such that they were difficult or impossible to read.

Over the last several years, displays on both desktop and notebook computers have tended to take on "widescreen" dimensions, where a display might be, say, 800 pixels tall but a whopping 1,280 pixels wide. Although these screen formats support digital entertainment like immersive games and widescreen DVD movies, they aren't great for Web pages. I'd always written my Web pages to conform to the size of a user's browser window, relying on users to resize their Web browsers if a page was too narrow or too wide. However, that idea just isn't effective anymore: Microsoft Windows strongly encourages users to browse the Web (and, in fact, to use essentially all applications) with their windows "maximized" so as to take up the entire screen. (Really, the operating system out to be called Microsoft "Window"…but I digress.) And Web browser interfaces have expanded so much that users actually risk losing needed functionality if they make their windows too narrow.

All these factors meant the lengths of lines of text on my Web pages could more meaningfully be measured in feet rather than pixels, or, in more typographical terms, ems. An "em" is the width of a square piece of type in a given font and point size; in practice, this was often the width of an uppercase "M" back in the bad old days of metal type.

In running text (like this paragraph) you don't want lines to get much longer than about 40 em; otherwise, the eye has too much trouble going back to the beginning of the column to start the next line. But while folks with old operating systems were seeing decent version of these pages, folks with up-to-date systems were often seeing overly long lines and strange layouts which made the pages difficult to read. So I wanted to solve that problem, too.

It turned out the mystery of the disappearing Percolating entries was easy to solve: the entries weren't vanishing, they just weren't caching quite right. (I'd implemented a caching system so search engines and folks checking for updates using the RSS feed don't put any unneeded stress on the systems.)

And, despite a few bits of new eye-candy, re-architecting these pages to use more-modern display practices has made the underlying structure simpler and easier to manage. However, there is a bit of a downside: folks using very old browsers (say, from the year 2000 and earlier) will probably see rather unattractive pages. However, they should degrade "gracefully"—that is, users of older browsers should get all the content and functionality, but not the pretty presentation.

(More technically-inclined readers are probably thinking to themselves "Geoff just jiggered the CSS for the pages"—and they'd be partly right! The details are all behind the "View Source" commands in your browsers, although bear in mind much of these pages are generated automatically, so, from this end, not everything is as random as it might appear.)

Anyway—I'm still playing with a few things here and there, and appreciate your patience while I get items settled. And if you need to tell me the new pages are ugly, yuu can still use the handy-dandy contact form.