[Prev][Next][Index][Thread]

Re: DSSSL style editing (was: RE: Positioning...)



>Once we have figured out the 95% of stylesheet features that 95% of all
>people need (i.e. more than CSS but substantially less than "anything 
>you want to do"), we can define an editable, interchangable DSSSL subset 
>and even a DSSSL class library. In fact, simply by defining more flow
>objects to handle the cases people need handled, we can make actual
>programming completely unnecessary for the same set of style features that
>will be available in "CSS 98".

If one parameterises the sylesheets correctly (read, *use*
programming to determine the output device characteristics, and react
to them), there is no need for choosing a subset. The programmatical
control DSSSL provides is not a barrier to implementation.

I once had a huge argument with a few people Jon knows well over
adding the capability to query the enviornment to DSSSL, and was
soundly toasted, but I still believe it to be the *only* way one can
get truly interoperable stylesheets (and I note that some things did
creep in afterward).



References: