Thanks. I was not trying to promote some return to SGML*: that is the opposite of the specific direction we need to go. It was just an example. Anyone who needs SGML's particular features has various SGML tools available already.
For tools, this issue of generalist versus specialist versus occasional versus expert will always be with us, and it is merely the arrogance of ignorance to claim that what one (as an individual or a representative of the masses or elite) finds optimal is universally true. Like Woody Allen's joke, that his parents would fight over anything, including whether the Atlantic Ocean was mightier than the Pacific.
Cheers
Rick
* The fact that no sooner had the dictum that
"terseness was of minimal importance" been applied to SGML to produce
XML, the MarkDowns arrived, where terseness is of maximal importance.
Now I do think that it would have been good for XML to have then had an
error-handling stage defined for it:
to allow tag implication, the kinds of whitespace movement SGML
required, and the kinds of tag movement we see in HTML5. But what
MarkDown shows is that with the advent of HTML ~4, SGML's presupposition
that we need a generalized tool that can support any document-type
dried up: all MarkDown users needed was a facade on HTML: going through
SGML (or any other compiler compiler like ANTRLY, PEG etc) offered no
advantage.
