Re: [Groff] Eric Raymond on groff and TeX

[Larry Kollar]

Anton Shepelev <address@hidden> wrote: 

> I think, groff and TeX macro packages do  provide  a
> means  for  structural mark-up, and, considering the
> example above, it is of course possible to  redefine
> the  macro  .B  to  achieve the desired result?  For
> clarity, it could also be renamed as "EMPH".
> 
> In my understanding, a package  provides  both  con-
> structs  for  structural mark-up and means to modify
> their underlying "presentation", and the one is very
> loosely  coupled  with the other, allowing to change
> "presentation" without affecting the "structure" and
> vice versa...

Personally, I don't think markup type is a binary function.
Yes, there is presentational markup, that can be
represented by (for example) low-level *roff commands.
And yes, there is structural markup, like DocBook or DITA.
Both have problems:

Presentation markup provides full control over the
appearance of a document, but do not leverage the
strengths of computerized typesetting — consistency, and
the ability to easily transform the markup to another type.

Structural markup addresses consistency and transformation
issues, but take all control away from the humans using
the system. To my knowledge, it's pretty difficult to
implement a feature like keeps in DocBook, or to squeeze
spacing to eliminate ending a chapter with a page containing
only two lines.

But I think there's room for a third kind of markup. I
call it *humanist* markup. Humanist markup has structure —
headings, lists, paragraphs, are easy to denote and
separated from presentation. The markup is simple to
transform to other languages. But in the end, the human can
step in and override things when necessary, because in
the end the humans know what they want. Macro packages
can provide that kind of flexibility, where BDSM markup
languages won't.

-- Larry