Re: Several questions to sgml-tools
On Thu, 29 Oct 1998, Antti-Juhani Kaijanaho wrote:
> can't find a good tag in linuxdoc, you can either file a wishlist bug
> or develop your own DTD and use (IIRC) Jade. In the former case, you
> would be better served by writing in TeX and *roff directly, and
> forget about SGML and HTML, since they cannot help you.
I had some experiences with HyperLatex. I think this will be my friend
instead of SGML, because I don't think that I have to write my own
> Your ideas do show that you haven't grasped the idea of separating
> content from presentation, which is fundamental to SGML, HTML and
> LaTeX (in decreasing order of the idea's importance). However, that
> does not imply that you are a beginner: I know many professional HTML
> authors that make the same mistake.
OK, so far. I ask my question other way around. Content is a
formula equation (not to complicated so that it could be expressed
with ASCII). I think an appropriate representation would be to
center this formula on a single line. Question: Which tag is the
right one to tell my style that this is an equation.
... so far to the centering problem ;-).
> On the second point, I'm not sure what you mean with "literate
> programming". In the Knuthian sense: no, SGML is not a literate
> programming tool.
To be more precise: I've read in the GTK list that DocBook is a tool
which uses SGML to enable literate programming (more or less in
the Knuthian sense). I havn't the faintest idea what kind of a
beast DocBook is but having a very short look at the documentation
it seems to have a complete set of LaTex equivalent tags. Seemed
me rather interesting.
> Yes. In fact, you are asking too much of HTML, too. You cannot
> expect Lynx to show larger fonts, yet Lynx is (at least trying to be)
> a conforming HTML implementation. Likewise, you cannot expect a
> speech syntesizer to implement text centering - the whole idea is
> meaningless in that context! Yet, many visually impaired people do
> use speech synthesizers for reading HTML documents.
But HTML contains certain Tags for changing the font size and for instance
you can include images too. Thats well defined in the HTML standard.
For poor browsers like lynx exist default methods like showing the
"alt" field of the img tag or printing the text with bigger font in
different color for instance.
I thing speech synthesizers will have several problems to express
Headings of sections too. Do they shout at the listener with
screaming voice at <h1> headings and shout a little bit more moderate
There ARE content elements that produce exactly bigger fonts
and so they are available.
But not all things which should be in bigger fonts are section headings.
For instance, I want to print a warning that my software is Alpha.
It wouldn't be good style for a typesetter to change the fontsize
for that purpose but in a HTML document I consider this to be
necessary for reasons I don't want to explain here.
Is there a tag <warning>? --- No.
How can I define rules to get such a tag? --- Write my own DTD.
I can't imagine that I'm the first who need this feature and so
my question is: Has anybody found a solution for this, because
I will definitely not write any DTD. I want to write a C program
and if the effort for writing the documentation exceeds the effort
of programming I simply leave out the documentation or write
a short README:
Look at the source. It is selfexplaining. Have fun.
> Thus it is more important to tell the computer about titles, headings,
> emphasis, trivia, summaries and other logical concepts than about how
> the text should be rendered. That way, even if the looks is not "just
> right", the text will understandable even in most unusual
> circumstances, which is the most important thing.
Of course, but if there are missing some logical concepts, how to
get the right markup with a reasonable effort.
> Probably. Last time I looked at that side of sgml-tools, it was quite
> simplistic. You will probably get better results if you start using
Can you explain in short words what jade is. Does it contain
styles in the sense I described above or is it an interpreter
to work with selfwritten DTDs?