can design 2.7

can design 2.3

later that day

It’s been nearly eight months since v2 went live and I think I can safely say Radiant is the perfect match I thought it might be when it was chosen as the backend for v2. I’m slowly getting the hang of radius which just makes me like Radiant even more. I’d guess that the move to getting all content as Markdown is about 99% complete, with the notable hold out being the cv. Having all the content in Markdown makes simple content edits (the most common type of edits) really simple.

Since v2.1 I’ve put up the first Radiant-backed client site (with four more in the works). I learned a great deal more about Radiant during that project, and by the end had scraped the nearly ready in December v2.3 update. The scope became broader. Then grander. Then out of control. By January I knew I needed a solid, not to be severely modified, plan of action.

By the time I decided a plan was necessary I had v2.3 slated to include basically everything I could think of that was wrong with (or I wanted to change about) about v2.2; css updates, content edits, javascript updates, browser support updates, html updates, navigational updates and more I’m not even bothering to mention. It was just too much to focus on, but more importantly it was a wrong minded approach, counter to how I work on client sites. So why not treat myself like a client?

the plan

It’s simple, effective and repeatable. 3 × 3 =

v2.3: the essential (1st quarter)

  1. clustered navigation
  2. better browser support
  3. better resolution support

v2.4: the accessible (1st quarter)

  1. 100% no javascript accessibility
  2. better html semantics
  3. html5 preparedness

v2.5: the accentual (2nd quarter)

  1. css updates/better fonts
  2. javascript re-write for Mootools 1.2
  3. re-implement switch like functionality

v2.3, finally

As you might imply from the above this update is simply the setup for the next two versions. But you might not be able to deduce how far this update really stretches. For example, this update drops all javascript except the one Mint needs; that pretty much knocks #1 out for v2.4. v2.3 also comes in the not so special arial, helvetica, sans-serif or courier, monaco, monospace font combinations making it possible to complete #1 for v2.5 at basically anytime with a simple find/replace.

Wow, huh? Temporarily degrading the experience to increase the speed with which the experience can be upgraded. Just to show things really can be upgraded quicker I’ve already made the v2.5 font changes. It’s now more like, use DejaVu first, but if you’re missing it use Helvetica Neue (Andale Mono for monospace) or the browser preferred fonts, whatever is available.

Of course, v2.3 also delivers on all it’s “deliverables”.

  1. clustered navigation
    • re-factored v2.2 navigation code
    • changed css rules
    • added the search cluster
  2. better browser support
    • clean slate css re-write
    • added special css to correct Internet Explorer 7 issues, very minor issues remaining targeted for v2.5
    • Safari 2 & 3 now just work
    • mobile devices might not receive any styling issue targeted for v2.5
    • updated time is now properly set in the feeds workbook, notebook
  3. better resolution support
    • came free with #2
    • works well in my testing from 1024 × 768 to 1680 × 1050, “extreme” size issues targeted for v2.5

Did I mention it’s also really fast. Firebug/Yslow claim 30-50% decrease in page load times across a sampling of pages (with the average around 30%). This is, probably, in large part due to trimming the javascript to 26 bytes or whatever is required by Mint. Of course that’s not to say that I like the site better without javascript; I already miss the fancy hover effects. However, it is an interesting observation that I’ll have to keep in mind as I continue work on v2.5.

2008/02/12