RE: Scheme Programming Reference

Subject: RE: Scheme Programming Reference
From: "Frank A. Christoph" <christo@xxxxxxxxxxxxxxxxxx>
Date: Thu, 24 Jun 1999 08:57:44 +0900
BTW, there is a sense in which I think DSSSL makes a very bad standard.

A standard is almost by definition an arbitrary thing. We standardize wire
protocols because, really, modulo some efficiency concerns, it doesn't
matter how you encode a message, just so long as the sender and receiver are
both using the same encoding. The same goes for metric systems and
electrical current.

In this respect, DSSSL is a very bad standard because it standardizes a
relatively broad field, whose parameters and boundaries are not
well-defined. It has too much innovation, and people probably would have
been served better if it had standardized only the most basic things, and
left the innovation to the interested parties. The design space is too
large, and different people have different ideas on how things should be
done, especially when it comes to programming languages. (Although I must
admit, I think the lambda-calculus is a much better candidate for a
universal programming language than anything OO.)

Also, a lot of people who write documents don't know anything about
publishing or typesetting. When that happens, they think about formatting in
imperative terms, in little chunks, rather than declarative terms, where you
see the whole picture, and the meaning and necessity of many DSSSL features
is not going to be as apparent to them as it is to publishing specialists.
It's easy for such users to dismiss DSSSL outright as being too complex or
unwieldy because, frankly, they are novices and aren't able to anticipate
what they will need in the future, for a slightly more complex formatting
job, or a slightly different document, or whatever.

And, yes, I'm feeling cranky today... :)

--FC


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


Current Thread