XUL Talk:Templates Plan: Difference between revisions

Line 1: Line 1:
== Comments [[User:Conor325]] ==
== Comments [[User:Conor325]] ==
==== Separation of query and condition ====
==== why a separate condition step? ====
Separating the query from the conditions would seem to remove efficiency, for example, for a server applying conditions to reduce a scoped data set to a managable size. The current template rules combine query and condition in the condition clause. Perhaps, rather than view this as a short hand for two separate things, let it stay atomic and let new mechanisms emulate the combination. Separating conditions also raises the bar for version 2. Certainly RDF templates could use better condition support, but is it a blocker before deploying a new data-format independent mechanism? Leaving conditions to the query mechanism would isolate this shortcoming in the RDF mechanism, an infrastructure that could be considered separately from making a data format agnostic template builder. Perhaps query-wise, it need never change as it may be superceded by a SPARQL query processor for comprehensive RDF processing.
As query's are free-form, most of the time, conditions will be applied as part of a query. The conditions in the rule would follow on, sifting out data to match generation. The question is, is this overkill and overly complex? Why not make query scope match what must be built. If the query returns data then content will be built around it - there is no nuance per se. Want variations of content then vary the query. But isn't this less efficient? What if you want to share most of a query and add nuance in multiple rules? Well, yes, this could be supported but it adds complexity to the syntax and does it really give much more USEFUL power?
 
BTW, in this view (query applies all conditions), the current template's conditions is the query section. There is no separation.


==== RDF caching and refresh ====
==== RDF caching and refresh ====
40

edits