- From: Felix Sasaki <fsasaki@w3.org>
- Date: Thu, 11 Oct 2012 16:30:17 +0200
- To: Yves Savourel <ysavourel@enlaso.com>
- Cc: public-multilingualweb-lt@w3.org
- Message-ID: <CAL58czq4-WR3xAY71NRZx549i1uhce+QEvxedbDEgEqmV3uPNg@mail.gmail.com>
ok ... but what to do about CSS selectors? How does its:param interact with these? We just end up with a lot of combinations of ITS processors: implements global rules and XPath 1.0 implements global rules with XPath 1.0 and param implements global rules and XPath 2.0 implements global rules with XPath 2.0 and param implements global rules using CSS implements global rules using CSS and param I'm not so much worried about the implementation choices - a well documented implementation will make the choice clear -, but rather about re-usage of rules across implementations. With ITS 1.0, each rule could easily be re-used, because there is just one conformance class: global rules. With the above, in fact we have six classes. Also, how to tell rule authors what of the choice the best practice is? Or in summary, the flexibility creates drawbacks on the rule re-usage and best practices side ... Best, Felix 2012/10/11 Yves Savourel <ysavourel@enlaso.com> > > How about dropping it completely? It is handy, but it > > adds complexity (e.g. how does it interact with query language?), > > and others might have the same implementation issues like you. > > I'd rather make it optional than drop it. > > The feature does add flexibility in real-life situations. So I don't think > we should give it up simply because a badly designed API does not support > properly XPath. > > Cheers, > -yves > > > -- Felix Sasaki DFKI / W3C Fellow
Received on Thursday, 11 October 2012 14:30:42 UTC