Lists Home |
Date Index |
Dare Obasanjo wrote:
> If you are not one of our customers then I doubt that you care what my opinion or that of my colleagues is about how to process and manipulate XML.
On the contrary, I care very much indeed. XML-DEV is, at its best, the forum for those with the passion and the experience to debate the nature of XML-at-its-best. Let us pass over in silence the rudeness of referring an XML-DEV thread to continuation in a separate forum that is a) proprietary, b) structured as petitions from the great unwashed to
the (self?-appointed) experts of the Microsoft Corporation, and c) by its charter and your own admission closed to discussion of those very issues of this thread which the Microsoft Corporation has not embraced in its idiosyncratic (possibly solipsistic) view of XML. Let us instead concern ourselves with what it means to have the half, or more, of
what XML can do that no other tool I have seen in twenty years can match snatched from us because it does not comport with a datamodel which is nowhere to be found in the specification of XML. That of course we cannot discuss on the forum to which you would direct me because the narrow-minded premises of that forum do not comprehend the necessary
I wrote here on XML-DEV (http://lists.xml.org/archives/xml-dev/199808/msg00330.html) more than four years ago that this schism would break this community. I am very sorry not to have been proved wrong. I will not, however, refer you to some other forum where I, or anyone else, redefines XML to suit some fluid preverbal Gestalt, or some cockamamie
attempt to structure semantics, meaning, or models. That effort is, and always was, moonshine. The best example in all of scholarship of where we are headed if models dictate syntax, instead of being elaborated from autonomous instances of it, is the entire nineteenth century course of Homeric scholarship in the hands of the 'learned Germans'. They
were absolutely precise in their models, nuanced to a fare-thee-well in semantic distinctions, and--as Milman Parry proved with real instances of real oral poetry--absolutely wrong in every conclusion. And how could they not have been? Like the XML datamodelers their chief method of advancing their science was petitio principii. Once they had
concluded that the text self-evidently indicated a particular model they could then use the undoubted correctness of that model to impeach the accuracy or authenticity--or to use terms closer to home here, the validity--of this or that bit of inconvenient text, and brush it away to draconian termination. The danger, then and now, to such an approach
is that the price of admission to scholarship in the field becomes the rote memorization of the conclusions, the well-modelled semantics, of every great man who has left his name upon this or that construct or theorem. As the field moves thus inexorably further from the day-to-day stuff on which it operates--the instances of text (unless, of course,
you will sneer at mere empirica and confine your theorizing purely to abstractions)--the 'certifiable' practitioners become ever more removed from the grunt work that now gets done in manipulating documents as documents and ever less concerned with what needs doing, caring only how what they deign to allow done should be done.
I did not come here for governance from the top, nor instruction out of abstractions, nor motivation by models. I have real work to do with real numerical financial data and the legally-privileged texts which decree what must and may be done with it. My job is to read the one in para materia with the other. That is inevitably mixed content, and the
processing which it demands depends in every instance on the specific interweaving of the two in an instance text. That is content, and whether or not it is a content model is utterly beside the point.
Do you have anything to offer as XML, or are we now to move on to a different topic on this list?
Respectfully, as always, because I am here privileged to debate in the company of those I would aspire to have call me their peer.