Who You Calling A Jesse?

Shortly after I started working in the Higher Education web space (2001) I came across the brilliant post by Jeffrey Zeldman on A List Apart that lead down all sorts of paths towards web standards and accessibility. I wasn’t alone. I think many web folks that were dealing with the internet bubble bursting were inspired by Zeldman’s call to arms to change things and many had the time to explore the possibilities. I did what I could in my position to influence the University of Waterloo web space and in 2004 we had a XHTML/CSS layout that was clean and accessible which was finally let loose on campus in early 2005.

Things changed on campus and I spent more time on usability testing and meeting with the few students that relied on adaptive technology. I wasn’t put off by the fact only two people might notice the enhancements as I knew UW was doing the right thing by fixing things. However, all of the applications students and staff rely upon were not going to be fixed or changed with even the course management system saying it was section 508 compliant but that version was even less usable than the main user interface. A problem that I have observed is that accessibility laws or regulations seem to force people under the covers in the HTML to make things work in screen readers (sometimes) but people ignored how usable the content or the application actually is.

It gets stranger by the day, developers demand unit tests they can meet to make the app accessible but there aren’t any… I don’t think. Laws and guidelines just compound the problem by giving people a false sense of compliance. In the case of learning environments most aren’t even all that usable but golly gee they are 508 compliant. It starts to drain hope.

Blame technology or developers?

A developer most certainly should make browser based apps (HTML/JavaScript/CSS apps) ‘professional’ grade by using semantic HTML, unobtrusive JavaScript, and sensible colour contrasts. That checks off a lot boxes in terms of Search Engine Optimization, re-usable code, dealing with rendering fun, and accessibility. There are different ideas of what it takes to make a web app or page accessible however, and I am not sure a developer should kill a load of time on certain things (that change with the project) like zoomable layouts—especially when browsers are implementing features that make that time wasted.

I am not sure that is inline with that Derek brings up in his post When is the right time for accessibility? as I think some (or a lot) of the things that are generally seen as making HTML as ‘accessible’ really should be the responsibility of a different team of developers (mainly those that make web browsers). I don’t disagree with the strategy of implementing accessibility later based on need and I think Derek’s post offers a bit of an olive branch to developers. You shouldn’t be expected to be all that accessible until you actually know that (a) people will use your product and (b) knowing how people will use your product.

What is my problem?

Honestly I don’t know. Call it a long winter, annoying problems repeating themselves for years, and new experts making the same mistakes.

I started this post sometime after I saw the small torrent of comments about a JavaScript framework which was summed up in Drew’s post The Cost of Accessibility. Drew is on that fence of innovation needs to take into account the reality of the web browser right now and I am not sure I agree.

At the same time I got into a few insane conversations about making the new job system for co-op on campus accessible and IE6 isn’t dead (like we had hoped) for an important 5% of our user base. Our development process makes it insanely difficult to spend time testing, fixing, and tweaking for accessibility (application is hiding behind a VPN and has a few other features that make it hard to access outside of our network). We use jQuery wisely, CSS, and HTML to the browser. AJAX is sprinkled in parts but nothing should depend on it. For a first version that isn’t really ready for user testing it has some good fundamentals but someone pulled the ‘yes but you are missing x’ and I just got deflated.

The problem, in my mind, goes back to the way people think about accessibility. The whole issue is an Art not a Science and certainly not engineering. Engineers have left us with this problem, with HTML, a stateless browser, and a crippled feature set that forces hacks, short cuts, etc. They are doing their part, slowly. Code artists like Derek Featherstone and Drew McLellan help spread the word and lower the barriers through simplified approaches and keeping the dialogue going.

HTML 5 gives me hope even if the Engineers aren’t too quick to drop it in our browsers! I am not really cynical ;)

Quick question. If you have multiple devs working on a few different screens, each monkeying with CSS, it takes very little time to end up with a huge mess of CSS. How do you deal with that? Do you:

1. Delete the CSS and start again defining a common sheet?
2. Try to optimize the CSS.
3. Live with it.
4. Don’t ever let dev’s touch CSS… they are dirty.

I ran into a 6000+ line CSS file for a dozen pages. They each have some heavy js UI going on but 6000 lines? An auto-optimizer cut it to 2200 or so pretty quickly but you can’t work with that file. I decided to start again, clean.

The upside is that I know the site mostly works without CSS and it exposed some odd decisions with some of the HTML (yay for nekkid web sites). The downside is that we may have to deal with browser bugs all over again—but then again we do not support IE 6. Only IE 7+, FF (latest), Safari (latest).

Feature request for Dreamweaver CS5 – something to optimize my CSS!

I have been on thing about figuring out coding patterns in HTML. Since I did the UW CLF back in 2004, I have been thinking about a macro-format for content generated on higher education web sites. Any CSS framework uses some abstract naming convention now—so I guess what I have been looking at is a “blueprint” that works specifically for higher ed.

What I did today was grab the code structure from about 10 higher ed web sites (three each from the UK, US, and Canada plus one more). It is just amazing how different HTML can be. Most sites are similar design wise, they have very similar content, and they supposedly trying to provide the same type of experience to the exact same audience.

Only three had Microformats on them, one had errors, and all are ‘valid’ HTML/XHTML. Good and bad ;) Well time for a break then on to more research and maybe even some prototyping. You can call what I am researching is a possible Macroformat for higher ed…

There has been a pretty interesting flame war that has erupted over a posting by Jeff Croft entitled What’s not to love about CSS frameworks? It seems like it has been quite a while since a good flame over web standards and best practices has played out. The tone of the post likely has really fueled the war but the topic itself seems to truly polarize some in the web standards community. Why is that? The devil is likely in the definition and I see it as the less formal art world colliding with the engineering world (something that has been slowly happening for a while with web development I believe).

Jeff Croft posted some follow ups: A follow up on CSS frameworks and The final word on frameworks, from someone way smarter than me. Andy Clarke interjected a comical What’s not to love about instant cake mixes in between that offered some satirical insight. The comments on the posts are shocking in some ways but once the definitions were clarified I think it comes down to artistic approaches meeting formal engineering process.

If you agree a framework is just a collection of reusable code that offers enough abstraction that you could apply it to whatever project you are working on then you have probably some engineering exposure ;) Reusing things is common practice, if you have a problem with that then you are just plain dumb with your time. This reuse of code features is part of what makes Dreamweaver CS3 such a good tool for rapid development. The CSS templates that come with it offer a powerful ‘framework’ to start with. Would you consider that a framework? I dunno. The ‘CSS Framework’ proper that is implied (blueprintCSS ) is in fact a more extensive framework that tries to solve more problems.

I think frameworks are great. I am building one now along with my GUI team of co-op students for a new system here. We are using a more formal engineering process to approach it but what we are essentially doing is creating a framework of GUI elements along with their HTML and JavaScript. Love them or hate them frameworks are just another thing the web dev world ‘re-invented’ from the software engineering world.

Over the weekend I sat down and went through Dave Shea’s and Molly Holzschlag’s new book The Zen of CSS Design (which you can get at the UW bookstore). What a great book. It gives some good insight into what is behind the CSS Zen Garden and some design concepts. Really worth a read if you can squeeze it into those end of year budgets ;)

There were a couple things that did stand out while I read it though. The first bit is more of a ‘this is the code in the garden and why’ which has caused a bit of stir in places. One discussion that spilled over into emails to Dave Shea caused him to post a response on his web site about the AAA statement on the Zen Garden. Another one, The Zen of Disinformation brings to light some issues with the accuracy of some parts of the book. Molly and Dave have comments in the thread of that article. Interesting thread.

Even guru’s make mistakes and they really shouldn’t be tarnished too much for it… the book is about the Zen Garden and design, for that it is really good book. Just be sure to grab Web Standards Solutions or Zeldman’s Designing with Web Standards (both should be in the bookstore here) if you are looking for more accurate and extensive information on web standards ;)

I have never been all that sure about using XHTML Strict for anything as long as DW did not support it. Why? Because you start of strict but as soon as you hand over the page to someone else there is no way to ensure it stays that way. Transitional languages are better suited for that scenerio. When my co-op student from term needed a way to learn CSS I told him to make a pure CSS UW home page .. and he did.

Today (until around 12pm) the current UW home page was XHTML strict and standard CSS. There are, however, some limitations:

1. The left nav is not ‘clickable’ in text mode
2. The search form is not accessible
3. IE 5.1 on OS X has some issues

We were trying it out until we got complaints, which I was certain it would happen by noon and it did. But it is an attempt to make the UW home page accessible without changing the design – which I am not entirely sure is possible.

The page can be found:

http://www.uwaterloo.ca/indexstrict.html

The first problem that was noticeable was the drop down navigation, it failed in IE 6. But what else is there? Other than the drop downs, was there anything else wrong with it? I am wondering if anyone can duplicate the IE 5.1 on OS X issues?