RE: Future of XQuery and XQuery Update Faclilty

[Hans-Juergen Rennau]

> Do you like the current draft, or are there things we should change? 

More than liking it, I am enthusiastic about it, admiring the elegance with which seemingly incompatible concepts (values and actions) are forged into one whole. And from the usage point of view, the new expressions feel as natural, straight and slick as I could imagine.

There is only one point which worries me about the draft, and this is not the content, but rather an aspect of its presentation: I would wish for a still more explicit emphasis on the fact that the XQUF is in the *first* place about updating XDM instances - not about firing updating intention into the external environment. The latter possibility is built on the first - but it is not at all the raison d'etre for it, rather an application, isn't it? My view is that the updating directions can be "poured" either into a within-query XDM instance (transform expression; from the user's point of view it is irrelevant that - strictly speaking - the copy source is not modified), or into the query environment. So we get the notions of a within-query update, and of an external update. The stormy discussion seems to reveal that the within-query update is not very much considered in its own right, and the external update regarded as the heart of the matter. (Or am I
 wrong?)

Frankly, it makes me nervous that some certainly non-trivial problems connected with external update should stop the vitally important possibility of within-query update.

Therefore, I wonder if it should not be considered to define two conformance levels: 
a) level 1 restricting updating expressions to occur within modify clauses and to target nodes of the respective copied node sequences; 
b) level 2 without such restriction.

Level 1 seems to me so perfectly integrated into the current model - all the sparks fly within the transform's modify clause, and behind it everything is back to normal. And the positive decision to include it into the standard language seems to me of strategic importance for the language: it would be a decisive move toward a *comprehensive* ability to deal with structured data of any complexity. (As mentioned earlier: to perform local modification of highly complex documents can become next to unfeasible - what with the necessity to reconstruct node by node all the complexity embedding the region of interest. What is our view of the scope of XQuery if  we accepted such a limitation?)

Hans-Juergen Rennau


      Jetzt Mails schnell in einem Vorschaufenster überfliegen. Dies und viel mehr bietet das neue Yahoo! Mail - www.yahoo.de/mail