Re: About Constructions rules

[Brandon Ibach <bibach@xxxxxxxxxxxxxx>]

Quoting Matthias Clasen <clasen@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>:
> >    On that note, I'm starting to think *very* seriously about making
> > an attempt at implementing the transformation language in Jade.
> 
> That would be a very very nice thing. What do we have to do to seduce you
> into starting this project ?
> 
   Well, as I said, my time is pretty short right now, so I'll
probably just be playing with DSC 2.0 for a bit, and digging through
its internals to help form a plan for implementation within Jade.
Once I've got a basic plan, I'll need to see how it would fit into
Jade, then lay out what pieces need to be done, and hopefully divide
them up amongst all those talented C++ programmers who will have
magically appeared to help out with this. :)
   Seriously, if I'm to do most of the coding, this will be a long and
slow process, but I'm fairly committed to see it through, as long as I
can make sense of what Henry's done in DSC enough to be able to at
least come up with a basic implementation plan.

> Yes, definitively, but maybe as a separate program built upon the same
> libs (libsp, libgrove, libspgrove). 
> 
   While I can see the benefit of this, I have to agree with Avi that
they should be together, as that is where the power will be most
useful, in being able to specify a transformation (or two, three...)
followed by a styling step.
   One other possibility that occurred to me would be to provide
alternative forms of the (process-*) procedures which would allow you
to specify which transformation or style spec you'd like to do the
processing with.  That way, you could sort of do your chain of specs
from the "end", where you have an application of a style spec which
triggers other processing in order to get its input.  Just a thought,
anyway...

-Brandon :)


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist