[xquery-talk] Xanadu

Hans-Juergen Rennau hrennau at yahoo.de
Sun Feb 15 05:30:55 PST 2009

Hello Ghislain,

your wrote:
"The question here seems to be: can XQuery do more, e.g., do what Java/C++ do --  
without reducing performance on pure XML processing -- and would this be desirable at all?"

In response to your words I had a daydream, but before telling it I would like to remind you of the poet S T Coleridge who once in a dream saw a poem ("In Xanadu did Kubla Khan / a pleasure dome decree ..."), vivid and beautiful, who after waking up started to write it down, was disturbed by a visitor and afterwards, alas, could not any more reconstruct anything.

I dreamed that the time we presently call the present time is deeply steeped in anachronism. The OO- and the X - way of thought exist in serene oblivion of each other. One day people will remember with amusement and compassion that once there was a time when it was not evident that objects may extend, besides their behavioral interfaces, informational interfaces, each one having a name and as content an element information item constrained by a schema type. But that, so they will remember, was only the edge of darkness, concerning a single object's interface. Even worse, perhaps, was the darkness steeping the inter-object relationships. With unbelief they will remember that there was a time when objects could call each other only immediately, had to know each other personally, so to speak, in order to invoke each other's methods, when there was not yet a system tree looming in the background of any program execution which you can extend, rearrange and
 prune but never delete and never create, to which objects may be attached via dedicated attributes whose type is object sequence and whose default value is an empty sequence. Yes, there was no way to delegate a method call to the system tree, prefixing the method parameters with two further parameters, one being an XPath expression selecting the objects, the other one being an object type defending type safety and acting as a second filter superimposed on the filter implied by the XPath. The strangest thing of all, however, was that programmers, even brilliant ones, accepted that state of affairs calmly, without protest, just wrote their programs, some of which were pretty complex and some of which were pretty reliable.

But I think I heard a bell, so I hurry to come to a close. Certainly there will be another mail giving all the details. Good bye, with kind regards -

Hans-Juergen Rennau


More information about the talk mailing list