Allright.
> How is that condescending?
Well, to me it implied that I couldn't express my opinion about error
message reporting, except if I was part of the browser coding WWW
hard-core. Well, I'm sorry if I got it wrong.
> And how's your browser coming?
Well, I work at the Euromath Center (as you can see from my
signature), which is part of the Euromath Project. Euromath develops
an environment for research mathematicians, for editing math
documents, searching on-line databases, sending mail, stuff like
that. Everything is based on SGML, at the core of Euromath we have a
commercial SGML editing engine, GATE, which we bought from (and partly
co-developed with) Grif S.A. in Paris, France.
Euromath lets you do WYSIWYG editing (well, you probably know about
the qualifications you need to make that a meaningful proposition.
Anyway, in Grif/GATE you a presentation specification for each DTD,
which is then used for editing, printing and so on.) of SGML. We have
developed our own Euromath DTD, and are part of the standardisation
effort to get a common Math DTD fragment.
I'm responsible for the development of a prototype Network Information
Service in Euromath. So far I have developed a Gopher browser, and I
have built a client which downloads SGML (Euromath DTD) files from our
WWW server using HTTP.
I've also written a DTD equivalent to the current HTML, which lets me
do WYSIWYG editing of WWW documents, and then export them to HTML. I'm
considering writing an export facility from the Euromath DTD to
HTML[+].
What I HAVEN'T done, is write a translator from HTML (or HTML+) to my
HTML-equivalent DTD. Therefor, at the moment, I can't import HTML docs
into Euromath. -- This is where all moaning and groaning about bad
markup comes in.
WWW was perfect for us: It was there, lots of interested information
providers (there is already quite a bit out there, as you know), AND
it uses (claimed to use) SGML. -- Should be a piece of cake, 3 months'
work, and we'd provide the Euromath community with access to all sorts
of information from within their own environment.
The catch, of course, is that we can't handle bad markup, our editing
engine is TRUE SGML; there is no way it will handle tags in all kinds
of strange places. Never. Of course, writing the translator from
HTML[+] to our HTML-equivalent DTD, I could anticipate some problems,
and handle them, but there's no prospect that I can handle
everything.
So how's my browser coming along? Well, right now I'm waiting to see
what HTML+ will look like in the end (soon, I hope?). If the result is
reasonable, I'll try make a translator. Otherwise, we'll probably
build our own little distributed hypertext, using our own DTDs (but,
perhaps, using HTTP). That outcome would be a real shame, since there
is so much attractive (sic!) information available through WWW/HTML.
Enough said.
Cheers,
Klaus
PS. I thought I'd post this on list as well... -K.