RE: forbiddenCharacters data category - related to [ACTIOn-189]

Hi Jirka, all,

> Of course implementations are also important. But as there are 
> open-source implementations of XML Schema regexps for all major 
> platforms -- for example Saxon for Java/.NET and libxml2 for C/C++
> -- I don't see any problem here. You will simply reuse existing 
> code instead of relying on default platform regexp engine.

I think there is a vast difference in using the platform's regex and a third party library: adding dependencies may be difficult or not possible in real-life scenarios. Also can we be absolutely sure that all major programming languages will have a free and working implementation of XML schema's regex (including Ruby, Python, Client-side JavaScript, etc.)?

I've seen a similar story for SRX: the regex syntax is based on ICU's. The idea was that applications could easily use either the C, C++ or Java implementations. The result wasn't that rosy. To cut the story short, today almost every application uses the platform's regex engine instead of ICU's and is neither supporting SRX properly nor provide true interoperability.

I hope we can avoid such outcome for ITS. We have the chance that a sub-set of XML Schema's regex would be enough to do the work and be interoperable with all (as far as I know) other engines... I'd say it's an attractive solution. We do think mostly about the users here: trying to prevent them to end up with interoperability issues.


Received on Tuesday, 28 August 2012 02:51:54 UTC