- From: Jim Davis <jrd3@alum.mit.edu>
- Date: Wed, 28 Feb 2001 13:34:28 -0800
- To: www-ql@w3.org
At 02:02 PM 2/28/01 -0500, Michael Kifer wrote: >Jim Davis <jrd3@alum.mit.edu> wrote: > > .... parsing XML queries is > > undoubtedly made more difficult by the non-XML syntax. > >Jim, >are you referring to the speed of parsing or to the difficulty of parsing. The difficulty of creating the software that does the parsing. At present, we already have XML parsers, so it's zero effort. By contrast I at least do not have access to a parser for FLWR syntax. Doubtless such will be created, but it all seems to me to be useless effort. (We will need FLWR parsers in Java, perl, python, C and all the other languages, and perhaps multiple bindings for some (for XML n Java I have to choose between SAX and DOM.) These parsers will have to be designed, debugged, optimized, and checked for security holes and flaws that might expose one to attack. There will be one set for the free software crowd, and another for the paying customers. There will doubtless be platform-specific parsers, and long flame wars. We can avoid all that by using XML. And I see no benefit - none whatsoever - to departing from XML. Wasn't the very point of XML to at least provide a common low-level syntax for encoding messages, so we would escape the babel of ad-hoc message formats previously used? Note that at present we already have Z39.50, LDAP, and SQL as three different query languages, each requiring its own parser. Why add yet another? What would happen if each W3C group created its own ad-hoc language to express the messages it was creating? We'd end up with dozens of different parsers >..., XML is a data exchange >format, and standardizing is a good thing, because you need to query and >possibly reason about that data. However, exchanging queries is a different >matter. ... and a user application doesn't reason about >queries --- query optimizers do (and users don't write them). So, I fail >to see the rationale for the XML-ization requirement for the XML query >languages. This makes clear that our disagreement stems from differences on our expectations of the number and nature of programs that will use XML query. Do you agree that that's the fundamental point? If so, we can attempt to resolve this, and then see how it affects our opinions about syntax. If I thought that XML queries were only to be parsed by query optimizers, and there will be only a few such created, then indeed the syntax would not matter. But in fact I believe there will be many applications constructed that either create or process XML queries. I think there will be hundreds or thousands of (separately authored) programs that construct queries. And there will likewise be many programs that process XML queries. Consider for example how many meta-search engines there are for current web-searches. Perhaps others on this list could provide their estimates of the likely number of (separately written) software systems that either create or process XML queries? if there are only going to be three or four, then this point is not worth debating. with best regards Jim
Received on Wednesday, 28 February 2001 16:35:43 UTC