[XML-SIG] Functional interface to XML?
Tue, 16 Feb 1999 18:35:33 -0700
> > We
> > primarily use it to emit HTML or transformed (non-FO) XML for now, but we
> > could provide a SAX interface to make it more useful.
> If you are just emitting HTML wouldn't it be easier and more efficient to
> use a "reverse SAX" output engine instead of building and walking a DOM. I
> believe XSL is designed so that you can output everything as you generate
Probably. I'll be the first to admit that we've been working in "quickest way
from A to B" mode, and haven't seriously enough incorporated performance
concerns to our approach. This might limit 4SXL's general usefulness until we
optimize it. I'll cache this among the growing list of "excellent Prescod
suggestions" in our design process.
> > Well, I'm still not sure I get your meaning, but the pattern-matching code is
> > pretty well de-coupled from the tree-visiting portion. This is to allow
> > easier updates as the XSL WG modifies their spec. I'm not sure how usable the
> > API would be when exposed outside 4XSL, but we can certainly keep such
> > potential use in mind.
> What I am *really* interested in is answer Cee's de Groot's question
> (which is also mine) of an API that *Python programmers* can use that acts
> like XSL or DSSSL but uses Python syntax. Think of it this way: if XSL
> does 90% of what you want then you want to just extend it with Python. If
> XSL only does 20% of what you want then you want those XSL features in
> Python without the XSL stylesheet syntax.
Aah, the light dawns. And I guess this one also goes into the list of useful
suggestions. I like the idea a lot, and I can already forsee many useful
applications. The main problem will be how much effort into modifying the
current design would be required.
FourThought LLC, IT Consultants
Software engineering, project management, Intranets and Extranets