[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
SGML vs other standards
Haven't been around for a while, but I noticed some comments
about why is SGML the standard? Why not others too.
I'm sure we aren't going to get into the great debate again
but I thought I'd share the following thoughts...
Once upon a time there was plain old text.
Then came basic systems that permitted words to be
made in bold or italics (if our daisy wheel
or dot matrix driver supported such things - for the older ones
amongst us remember EPSON FX80 codes etc)
Then WYSIWYG came along - brave new world stuff.
We all jumped on the band wagon (in the PC world at least)
and had all manner of fonts everywhere.
[You've seen the results of someone new to word processors
a different font on every line...]
When the web came about the natural extension was for a
simple wysiwyg system, hence HTML (based on the foundation of
Meanwhile in unix land they stuck with nroff, troff etc
then came tex and latex. Only recently (eg StarOffice etc)
have we gone all gooey.
The problem is that there are now millions of pages on the
web and on corporate intranets. Sure they're all pretty and
individual but that really doesn't contribute to the
problem of turning data into information and thence knowledge
Almost none of this mass of info is properly indexed. The
results of search engines demonstrates some of the issues.
And yes I do have a point :)...
We now have too much data. We need data mining to sift
through it all. To do this we need structured data.
PRESENTATION IS ALMOST IRRELEVANT (I'm not saying unimportant
I'm saying irrelevant to this problem. given a well structured
document presentation is realtively trivial and will look better
than a hand crafted solution. IMHO most latex documents look
more uniform and elegant than any Word document, and you don't
spend hours agonising over tab stops etc).
We are all now just starting to move down the path of
structured information content. A relatively simple extensible
system that is human readable is required. SGML provides
a good environment to do this.
Now we can really start filtering and indexing data properly
and presenting in an appropriate manner on anything from a
TV set to a computer screen to a mobile phone. Why because
we know what the data is about, so we can be more intelligent
about how we present it.
This technology although reasonably well understood is still
in its infancy. Sure SGML editors seem like a step backward
because they don't focus on looks, but the power is greater.
It is a bit like windowing system vs command prompts, the latter
tend to be uglier but a re far more powerful.
Things will improve, we'll get good (free) SGML tools running
that strike a balance between structure and looks. A little bit
of visual aid and indentation is of benefit as a visual
cue to structure.
The last thing we need is to have another input format.
Sometimes democracy is bad. IMHO this one of those times.
One input standard is best for now whilst we develop this
technology. Multiple standards in this case will only hamper us.
BTW IMHO I disagree with someone's earlier comment that pictures
are bad. Pictures are good, very very good. If you have ever
tried describing something even simple like typing a bowline
without pictures you'll know what I mean. Computer things tend
to be much harder. More pictures. Sure maybe we restrict
ourselves to B&W vector formats for now. But we should be
pursuing some good drawing standards.
Datafusion Systems Pty Ltd
To UNSUBSCRIBE, email to firstname.lastname@example.org
with a subject of "unsubscribe". Trouble? Contact email@example.com