Wrappers, injectors, and writing tools

I gather that this way of representing my RSS feed is ready to declare victory over this way. Wake me up when it's over. At the end of the day, any XML metadata wrapper around the content of our blog entries will do the job, and it's trivial to transform one flavor of wrapper to another. If there were no legacy to consider, it'd be a toss-up as to which I'd prefer. Since there is a legacy, I'd rather preserve it, but that's a complicated matter about which too much has been said, and I'm only one of many voices.

Similarly, I gather that this way of injecting an item into blogspace is preparing to declare victory over this way or this way. Again, wake me up when it's over. Were there no legacy to consider, I'd much prefer the new approach. I like its RESTian purity, though I'd also be open to a SOAP variant that could optionally leverage all of the authorization and routing machinery that's finding its way into SOAP headers. Of course there is a legacy here as well, and in this case, it seems to carry some weight.

Oddly, despite all my blogging, I seem to depend on none of these injector APIs. When the "Blog This" bookmarklet first surfaced I used it for a while, but soon lost interest. I never wanted to blog just a link with a sentence of description. I always wanted to write something more substantial. The same holds true for comments, another major use case for the injector APIs. No matter which API wins, we will still -- so far as I can see -- be dumped into HTML TEXTAREA widgets to compose the content that is the ultimate purpose of all this blogging. Isn't it?

So while wiser heads than I debate the pros and cons of various wrapper and injector strategies, I've decided to try to achieve some forward motion on a different tack. For reasons and in ways I've recently been demonstrating, blog content would be a lot more valuable if it were easier for non-emacs-using civilians to write XHTML. I'm particularly interested in finding ways to relate the style vocabulary of a standard wordprocessor, which is the only kind of granular metadata that people will consistently apply, to an emergent semantic vocabulary in blogspace. But that's a long-term thing. In the short term, I'm just looking for ways to empower regular folks to create well-formed blog content.

My first experimental subject will be Word 2003. I've just received the "beta refresh" so this is a good time to explore whether it's practical to write in WordML (Word 2003's XML vocabulary) and then transform to XHTML. If you've already gone down that path and have experiences and/or XSLT code to share, I'd love to hear from you.

Former URL: http://weblog.infoworld.com/udell/2003/07/10.html#a740