Friday, January 18, 2008

The Difference Between Theory and Practice

Specific constructive criticism in a comment from Dave X inspired this blog entry. (See Dave's comments on Susan's January 17, 2008 blog entry.) We like to hear what others are thinking. And specific examples are something we can more easily address and use for improving the College web presence.

Dave X references this W3C document. In theory, I agree that URIs should not be as disposable as we often treat them. Our circumstances, however, remind me of Yogi Berra's famous insight: "In theory there is no difference between theory and practice. In practice there is."

Here's an unfortunate truth from the W3C page:
[Excuse] We just reorganized our website to make it better.
[Response]Do you really feel that the old URIs cannot be kept running? If so, you chose them very badly. Think of your new ones so that you will be able to keep then running after the next redesign.

We are moving from a very decentralized structure that has grown organically over time with no central planning or oversight. As we move into a new web environment, there will be broken links. I cannot say how many - probably lots. They may have been chosen badly. Others may result from decoupling content from our org chart (another way to say they were chosen badly). For high profile sites and pages, we may leave redirects behind. But there's a practical limit to creating redirects.

What are our hopes moving forward from re.web?

Some content changes over time. The URL may not change, but the content will. The CMS allows for content versioning, histories, and scheduling to move in current information to existing pages. This doesn't break bookmarks, but may make the bookmark irrelevant for the intended use. But I think this is within the spirit of the W3C article in question.

Other content is more static and can remain in its posted form indefinitely (like a news story about Her Majesty's visit). Blogs have the notion of "permalink", and Wikipedia has the notion of web-accessible versions - both of which are designed to handle some of these challenges. Our hope is that much of our content will live in this form for quite some time.

There is a small subset of content that we ought to treat as an "authoritative source" - that is, the one place that we look and link for specific information. These links should be treated as more "valued" than others, and should be widely referenced. If it should happen that this "authoritative source" moves, a redirect would be warranted.

But these three types of content don't seem to capture all possibilities. Web application (such as facebook or youtube) provide both services and content - and there is no guarantee of continuance of deep links within these systems. If I remove my profile from facebook, somebody's bookmarks might fail. If a video clip is removed from YouTube, the link in the latest re.web blog post will be rendered dead.

As the College web presence embraces some of the elements often lumped under the "Web 2.0" moniker, we may have similarly transient content on our site. Certainly no "authoritative source" should be treated this way, nor the main news stories about Her Majesty's visit. But student profiles, posts within an admission blog, or the like may be. That's just the nature of the web.

I'll make a final point. Because we choose to release content on a web site, does that obligate the site to host the content forever? In the non-web world we see companies like Disney release movies for sale for a limited time only. Should the College never remove stale content? FOIA requires, and the CMS provides, storage and retrieval of historical documents. But I'd argue for the College to have the option to post and remove Web pages when we find it appropriate to do so.

I don't know whether I've convinced any of you of my own opinion on the subject - but I hope you know we are taking your points seriously and will make the efforts we can to keep some continuity for our audience as we roll out a new College web presence.

posted by Andrew Bauserman


Mark W (re.web) said...

As far as relocated content, there will be an initial "hit" but we should also remember that there will be a better, more logical information architecture for the new site moving forward. If the location of a specific page or document changes, what may be left behind, the dreaded "404- page not found" error page should be customized and dressed within the main site template including, at least, the top level navigation. If the navigation is as intuitive as we hope it will be, visitors should be able to fairly easily find the relocated content. Between this and careful consideration of "authoritative sources" and redirects we hope to keep frustrations to a minimum.

On a related note- our new content management system will also allow us to validate internal W&M links. "Dead" links between departments and organizations will be a thing of the past!

Dave X said...

In an older blog post, it looks like the current plan is expected to change in 4-5 years. Will the reweb plan for now help eliminate possible redirects or dead links in future designs?

It is possible: Blogger future-proofs their URLs by including a date in the string. If you expect to be radically different in 2012, are you dooming what we write in the interim to be part of some inevitable "hit" in 2012?

Dave X said...

Re: "There is a small subset of content that we ought to treat as an "authoritative source" - that is, the one place that we look and link for specific information. These links should be treated as more "valued" than others, and should be widely referenced. If it should happen that this "authoritative source" moves, a redirect would be warranted."

I think the major point of the W3C article is that a good design would not require this subset of content to ever be moved. The other major point is that this category is actually much larger than you might expect. There is no reason to change the content of 'Spring 2001 syllabus for Philosophy 101', which may not be widely referenced, but way out in the long tail of your potential donors are alumni who spent lots of money and time with that material. If you design your web well, you won't have to shuffle old pages around with redirects and expensive redesign projects, and can mostly ignore the $0.00001 worth of storage it would cost you to keep something that could connect you to your alumni.

I'm not arguing that we should make all the old syllabi accessible, just that the transience of the web is a product of bad architecture. If you want good and useful content, design for posterity. If you tell your content providers that the work they put into some WYSIWYG CMS editor won't last more than 2-5 years, why would they get excited to contribute?