W3C home > Mailing lists > Public > www-tag@w3.org > July 2011

Re: Dropping RDF mapping from microdata spec

From: Jeni Tennison <jeni@jenitennison.com>
Date: Fri, 29 Jul 2011 09:10:41 +0100
Cc: "www-tag@w3.org List" <www-tag@w3.org>
Message-Id: <F54D9A29-40D7-4643-B375-245E0CCA9EF4@jenitennison.com>
To: Bjoern Hoehrmann <derhoermi@gmx.net>

On 28 Jul 2011, at 23:08, Bjoern Hoehrmann wrote:
> * Jeni Tennison wrote:
>> I don't think that there's anything stopping such a declarative schema
>> language for microdata vocabularies being developed or reusing a
>> validation technology from elsewhere. Actually, Philip Jägenstedt was
>> talking on IRC earlier about possibly generating XML from microdata and
>> then using RELAX NG to validate over it:
>> http://krijnhoetmer.nl/irc-logs/whatwg/20110728#l-549
>> but I'm not sure how serious that suggestion was...
> I would have thought "hsivonen" would be Henri Sivonen, but then again I
> tell visitors every now and then to post things under my accounts.

No, you're right, it was Henri -- I misread the log (sorry Henri, Philip, if you're reading).

> I don't know if recognizing some subset of context-free languages is
> useful when validating "microdata", but I would recommend against trying
> to implement this through conversion to other formats (importantly there
> are usability issues to consider, like maintaining the right column and
> line information, which can be quite difficult with some libraries). The
> idea was appealing to me in 2005 mainly because I hadn't yet implemented
> various regular expression engines at the time and doing it seemed hard.

Interesting. I did something similar to implement a validator for LMNL (a markup language with overlapping structures); the implementation went via XML but only as a method of producing a format that I could work with. You're right -- it isn't outrageously hard to implement a validator of this sort.

For microdata, the obvious validation language to look for would be something that could operate over JSON, which I guess would be JSON schema [1], but that doesn't look to support some things that a RELAX-NG-like validation language would provide, such as choices between (sets of) properties.



[1]: http://json-schema.org/
Jeni Tennison
Received on Friday, 29 July 2011 08:11:06 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:39 UTC