W3C home > Mailing lists > Public > public-qt-comments@w3.org > October 2002

RE: comment on Xpath 2.0 - no strong typing!!

From: Kay, Michael <Michael.Kay@softwareag.com>
Date: Thu, 3 Oct 2002 13:08:07 +0200
Message-ID: <DFF2AC9E3583D511A21F0008C7E621060453DC08@daemsg02.software-ag.de>
To: bryan <bry@itnisk.com>, public-qt-comments@w3.org

A personal response:

Bryan, thanks for telling us your views. I'm sure some people on the working
groups have similar feelings. However, a note like this isn't going to be
persuasive unless it provides some concrete evidence of problems caused by
the current design, and positive proposals for an alternative. 

It would be helpful to know whether you disagree with the published
requirements for XPath 2.0 [1] (and if so, which specific requirements you
dispute), or whether you are suggesting that the current design is not the
most appropriate response to those requirements, in which case it would be
nice to see alternative suggestions.

The term "strong typing" means different things to different people. There
are some type rules in XPath 1.0, for example you cannot supply a number as
an argument to the count() function or as an operand of the "/" operator. We
have tried to ensure that XPath 2.0 remains usable in a world where there
are no schemas and all data coming from source documents is untyped. Do you
feel that we haven't catered adequately for this scenario, and if so, can
you be specific about where you see the problems? Or do you think that the
alternative scenario, where data has been validated and typed, is one that
XPath 2.0 should not cater for?

Michael Kay

[1] http://www.w3.org/TR/xpath20req

> Subject line says it all,  it is rather hard for me to put 
> together everything I would want to say here, so let me point 
> to the Subject line once more.
> Seriously, lots of people seem to have problems with the 
> whole intertwining of Xpath 2.0, Xquery, XSD. It seems to me 
> like the various groups want to achieve some sort of synergy 
> and they want to draw it from the success of Xpath, not 
> worrying that they might instead drive people from Xpath.
> I've complained other places about XSD, not here however 
> cause I didn't realize how harmful I felt it to be until I 
> actually started working with it.
> Obviously if I dislike XSD I cannot feel totally happy with a 
> technology closely intertwined with it.  Xpath 2.0 is such a 
> technology, All the various things I dislike about Xpath 2.0 
> can be boiled down to the rallying cry of no strong typing! 
> Of course if there were no strong typing in Xpath 2.0 the 
> integration with Xml Schema would have to be handled 
> altogether differently, or thrown out. 
> A point about strong typing in Xpath 2.0; there is a lot of 
> argumentation in the programming community devoted to the 
> superiority of either Strong Typing or Weak Typing, Static 
> Typing or Dynamic Typing. Pun coming her: "Xslt is bad cause 
> I don't like typing so much" arguments as well. Anyhow I was 
> not aware that these arguments were as yet resolved in 
> anyone's favor. For your information I am not a fan of Strong 
> Typing in anything although I find that I sometimes have to 
> use languages that enforce this. 
> I am sure that fans of Strong Typing need to use Weakly Typed 
> languages at times and they resent it as much as I do the 
> opposite, undoubtedly they have complained very loudly 
> regarding the lack of a type system for XML,  but let us 
> consider that XML rose to quite a high level of success with 
> no Typing to speak of. To win the strong typers over might be 
> a laudable goal, I think however that if you drive off the 
> weak typers to do it you will find yourselves with just as 
> divided a community as before. If there were some way to turn 
> off typing I for one would be happy, regarding this as a 
> suitable compromise. 
> As a reference I point you to the thread XPath/XSLT 2.0 
> concerns  at 
> http://lists.xml.org/archives/xml-> dev/200210/msg00014.html 
> wherein most respondents seem to be 
> of the opinion that there should be no strong typing.
Received on Thursday, 3 October 2002 07:08:20 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:56:43 UTC