- From: Jeni Tennison <jeni@jenitennison.com>
- Date: Sun, 7 Dec 2008 16:47:45 +0000
- To: James Fuller <james.fuller.2007@gmail.com>
- Cc: public-xml-processing-model-comments@w3.org
Jim, On 7 Dec 2008, at 15:14, James Fuller wrote: > On Fri, Dec 5, 2008 at 8:44 PM, Jeni Tennison > <jeni@jenitennison.com> wrote: >> I think a better guide would be that any process that returns an >> atomic >> value (eg string, number) should be an XPath function; any process >> that >> returns XML should be an XProc step. > > this could be made to be true (and I like the characterization); but > this is not saying much as the difference between a string, number and > xml could just be an interceding c:result root element. Yes. I just think it's wasteful to generate nodes when all you really want is an atomic value. And that it's more work for users to invoke a step than a function. > I think the less xproc ordained functions the better ... I would > propose leaving it to future versions to figure out. I agree that it's less intrusive for XProc to define steps than to define XPath extension functions; I think it'll prove less usable, but that is something we can correct later if we need to. Jeni -- Jeni Tennison http://www.jenitennison.com
Received on Sunday, 7 December 2008 16:48:21 UTC