I’ve been writing about identifiers lately, and I used an example of two identifiers pointing to one resource in one of my posts. It’s a very nice way of addressing the problem of versions on the Web, and while I can’t say for sure that the W3C was the originator of the technique, it is the first place I recall seeing it done.
The basic model is:
Latest Copy: http://[domain]/Local_ID
A particular version http://[domain]/ Local_ID-DateStamp
I don’t keep old versions of articles I write, or at least not in any systematic way, but talking with my new-found-literary friends, Mark Kelly and Brian Taylor, one of them remarked… yeah, most authors don’t keep more than 3 or 4 versions of their manuscripts…. This is a big deal in the humanities, but something that is probably not managed with much diligence by writers who use computers to write.
Keeping multiple versions of documents in a systematic way is easy with paper – a shoebox, or salvaged copier paper box is the system, and it works fine for the most part. It is no surprise that archives are organized around boxes’o’stuff, and that is largely sufficient to the task. Scholars who peel back the layers of this author or that artist will have a much harder time (and a lot less fun) peeling back the… what? File structure of someone’s hard disk? Especially since we throw out our old computer every few years (or lose it to a disk crash), the digital equivalent of throwing away our basement full of memorabilia. What seems a virtue in the moment (a chance to reorganize and rationalize a now-chaotic old file system) may be a travesty in the eyes of some future cyberhistorian.
So, the naming approach that the W3C uses for its official documents gives us a simple, understandable way to name collections of versions, and a transparent way to understand their evolution. Most of us won’t do it, even recognizing its value. We need digital shoeboxes, systems that understand versioning (that is, that have an inherent recognition of workflows of various types), and which help us to achieve the goals of such workflows.
In a previous post I noted a personal interest in the provenance of a particular document currently under development at the W3C. As luck would have it, the latest-version of the URL for this document is intact, but the version identifier is not, helping me make the point that even in organizations with a diligant commitment to the importance of version chains, it is easy to make mistakes.
Copied directly from the header information of a W3C Tag Finding URNs, Namespaces, and Registries:
As part of my explorations of identifier issues I’ve been reviewing this document carefully, as it touches closely on other work I’ve been involved in.
As it happens, my views are at odds in certain respects with the prevailing views of the W3C Architecture Group, (the TAG), and in particular I’ve sparred in public meetings with one of the authors of this document (Henry Thompson). I have high regard for Henry’s intellect, and I further believe his motivations to be of the highest order. So, I was at once flattered and annoyed to find our disagreement memorialized in an editorial comment in this TAG finding:
Editorial note: HST 2006-06-06
HST now owns
oclcnum.info, will sell to Stuart Weibel for a modest consideration :-)
Better to be talked about than ignored. Now, I can only guess that this tender of two potentially important properties in cyberspace is unlikely to make the final editorial cut in this official TAG Finding(1). But, I want to be sure that this offer remains part of the formal W3C record. Who knows…I might want them someday. Though, Henry… isn’t this cyber squatting? I’m thinking if you hang on to them long enough, you might be willing to pay me to take them! I wonder if smiley faces have standing in contract law?
Now... could someone at W3C please fix the URL?
(1) A Tag finding is a formal pronouncement on the part of the WsC Technical Architecture Group:
The primary activity of the TAG is to develop Architectural Recommendations. The TAG findings listed below document fundamental principles that should be adhered to by all Web components. The TAG expects to include these findings in the TAG's Architectural Recommendations, to be published according to the requirements of the W3C Recommendation Track process.
Image: Orcas Island Panorama, March 2006 (click to enlarge). The day I took this picture, you could faintly make out, from the highest point on Orcas, Mt. Rainier, over 200 kilometers distant.