[etech] Peace, Love and XML
Don Box of Microsoft, responsible for Longhorn Indigo (communications tech for building Web services, or “the SOAP messaging stack” that Don works on), is talking about how Microsoft is going to support standards, really this time. He says the following:
WordML, the current Word format, is optimal if you’re a Word author, but is unusable if you are trying to do interesting XML-y things with it, like write an app to process it. It’s designed to work well for Word. When Microsoft shipped it, people had a normal, human, emotional reaction: They hated it. Microsoft said that it didn’t expect you to author it, only to process it.
[A whole bunch of stuff I don’t understand it, and then:] We will be able to extend the Microsoft file system by providing our own schema. (Marc Cantor calls out that this is “really coolio, dude.”) “This isn’t just about the API. This is about data extensibility,” says Don.
Indigo is about “service-orientation” rather than object orientation. “We don’t want you to run .Net on your Linux box.”
[More stuff I didn’t understand. I’m not complaining, mind you.]
[Since I am obviously in over my head – feel free to explain it to me – I should perhaps report that both Bob Frankston and Marc Canter, off line, were favorably impressed with the direction Don sketched.]
Categories: Uncategorized dw
I wouldn’t count on Microsoft doing anything for the cause of open standards. Best off if they ignore the open standard. Worst case scenario is if they adopt it. You’re only a step or two away from them coopting it, controling it…and killing it.?
Thank you very much–also for your time and assistance with my blog design.
http://www.microsoft.com/mscorp/ip/format/xmlpatentlicense.asp
The part that Marc Cantor got excited about was the extensibility of the schemas that WinFS will use for metadata definition. So Microsoft said since they’re using the eXtensible Markup Language, that they will in fact continue to let it be extensible by letting their users extend the schemas. Brilliant! It’s like saying that even though we don’t like the words “open source,” we’ll let our users continue to use those words in Microsoft Office because they can type any number of words with a keyboard. If Microsoft hadn’t made their XML-based metadata thingamagij extensible, what’s the point of the whole thing?
Chris, thanks. You’ve moved me from bewilderment to puzzlement. Care to take one more shot at it?
Is this the point: I’ll be able to extend, say, WordML. I’ll be able to add a tag to capture the mood of a sentence or flag that a word should be triple underlined. Word, of course, won’t know what to do with those tags. Is this what MSFT is saying? If so, why would we want to extend their schema? So that we can then extend Word’s functionality through VBA (or whatever) to make use of the altered schema? Is that it?
If so, or even if not, in what sense do we need MSFT’s permission to extend their schemas? I think I’m missing something fundamental.
I think you got my WordML discussion exactly wrong.
WordML was indeed not designed to optimize human authorability.
Rather, the optimization point was ease of processing using standard XML tools.
The non-use of mixed content is one of several examples I drew attention to. Mixed content is lovely if your authoring XML in emacs. It’s not so great if you’re processing the XML using standard XML tools.
As to David’s comments, there are several mechanisms for adding “semantic markup” to WordML. The easiest technique is to use named styles (similar to using the class=”sad” attribute in XHTML). Word also supports user-defined schemas interleaved with WordML.
DB
Doh! I shoulda been clearer. I wasn’t even touching WordML. Microsoft doesn’t want you to extend WordML — it’s a contract for specific content parsing, not for content generation, set by Microsoft. As Don said, don’t even try to make Word docs with WordML, cuz it’s ugly for that purpose. Good for doing things like XSLT (and XQuery) transforms into other presentation vehicles.
The schema we (and now with Microsoft’s blessing) want to extend are the ones included with WinFS (that describe a song, a person, a picture, etc.) in Longhorn for the metadata filesystem indexing features. If we have some other goodies to include about a person (say favorite food, best side or FOAF data) we can extend the built-in schema to include our own elements and even propogate that out to the community. In that part of the talk, Don was letting us know that Microsoft is talking the ‘extensible’ part of XML at face value and letting us effectively hack the metadata layer they’re building. I’ll bet one they’ll have a schema to parse all of the Office format (like WordML) to pull out interesting info for us to do local filesystem searches.
Sorry for the confusion. I owe you a beer.
Damn wrong again. Ignore everything i say today — my brain is not functioning due to long flight home.
Don, thanks for the multi-level corrections of my truly profound and complex ignorance.
God Bless Don Box
[etech] Peace, Love and XML .