Why ALA’s “JavaScript Image Replacement” Sucksppk blasts Christian Heilmann’s JavaScript Image Replacement article. Overly harsh maybe, but for the most part accurate. The arraying and O2 looping is fugly, and ppk makes a really good point about replacement after image-loading (this would solve the Mozilla incompatibilities as well, yeah?)

ppk also launches into an attack on ALA’s JS articles, which again, while harsh, is… true. I love the design stuff, but the dev/tech stuff is hit or miss, which probably is most attributable to the lack of actual domain expertise of the editor(s).

My Treo’s touchscreen was freaking out on me for a little while, but it seems ok for now. I’m thinking I’ll need to give Handspring a call and see what’s up. In any case, I think I’ll do a little blogroll list a-la what I did for the hiptop (to this day btw, most of the list remains accurate; Danger never fixed anything).

The Blazer 3.0 browser has fairly complete CSS + JS support; and for the most part renders stuff well. It does take a while to render stuff (it’s not efficient about letting you navigate partial pages – still not bad, avg time to navigation is about 20s from hitting the request button, full page typically loads in about a minute), and doesn’t seem to be very smart about its caching (it’ll try to reconnect the last page when you don’t have a signal but won’t load up the cache). I’ve included the page sizes that it gets (spoiled by broadband). These are of optimized views (wide layouts seem to be even better; congrats Blazer team):

+ randomfoo.net - no header in optimized mode; 270.5K
+ a.wholelottanothing.org - header compressed, otherwise good;  221.1K
+ lyd - looks real good, miniblog shows up first; 47.4K
+ waxy.org - looks real good; 55.2K
+ torrez.org looks real good; 10.3K
+ sixfoot6 - looks real good; 192.2K
+ onfocus - loads, eventually; 334.7K
+ megnut - looks real good; 48.8K
! kottke - causes a soft reset!  no joke; tried multiple times
+ anil - legible, has two columns (?)
~ Simon Willison - wraps just a little wide; 37.3K
- 0xDECAFBAD - content space is a thin sliver, funky even in wide mode; 216.5K
+ Zeldman - nav first, otherwise good; 103.7K
+ Mezzoblue - long nav at top, works; 108.7K
+ whatdoiknow - scaled headers, otherwise good; 74.4K
~ Clagnut - text column half-width; otherwise ok; 139.4K
+ meyerweb - some design overlap, but quite legible; 77.6K
+ eatonweb - all nav on top, otherwise good; 45.0K
+ unoriginal creativity - nav appears first, otherwise good; 53.8K
+ disastro - no header, nav text/bg color rendered the same, otherwise good; 14.9K
+ boingboing - looks real good; 298.6K
~ mefi -  slightly thin text columns; 63.6K
+ /. - looks really good; 82.9K
Friendster, Tribe, and Upcoming.org load up fine.  Friendster sometimes renders
weird (ads or frames space weirdness?), but I was able to use it well enough 
over the weekend to look up an event location from a bulletin board posting.  

Browser improvements I’d like:

  • Threaded browsing, heck being able to access the menu while it redraws would be nice
  • gzip support
  • Bookmarks that remember wide/optimized pref
  • easier switching from wide/optimized mode
  • image zooming
  • easier to switch images on/off, CSS and JS while we’re at it
  • Typing goes directly into URL bar, a la hiptop
  • drop-down type auto-complete
  • Better Cache Control
  • can’t be fixed really, but the 11-bit screen is *really* annoying

To look into: more into browser info, does it support handheld stylesheets?

AT&T Wireless Re-Launches mMode

AT&T Wireless today announced the re-launch of its mMode mobile Internet service. Using the new xHTML standard on phones that support it, the service offers richer graphics and is designed to be easier to use. New services include Mobil Traffic, a location-enabled application that provides color-coded multimedia traffic updates for 40 metro areas, including traffic alerts, speed flows, and incident reports.

Ryan has a Nokie 3650, so I decided to chedk out the Mobil Traffic. After several minutes of confusing navigation, I did get to it, where I was confronted w/ a 30-day trial offer. OK, well, I wasn’t going to subscribe him onto it, but I did take a look at the sample. Not bad. Although being able to get my TANN and LADOT reports directly beats that out I think.

ok, here's a pic; TANN on the Treo

Recently I’ve been spending a fair amount of my work time working on WebISO (Web Initial Sign-on) and portal integration, specifically Pubcookie, which we’ve adopted as part of the NMI R3 recommendations, and uPortal [#4 in InfoWorld’s Top 100 IT Projects of 2003] integration (my preferred portal framework).

Unfortunately, previous Pubcookie integration efforts have stalled out, buton the bright side, uPortal’s security framework looks pretty adaptable…

Here’s a great quote:

Middleware is the intersection of what the Network Engineers and the Application Programmers don’t want to do.