Re: DOCBOOK-APPS: Re: SGML vs XML

[Trevor Jenkins <trevor dot jenkins at suneidesis dot com>]

On Tue, 9 Jul 2002, Norman Walsh <ndw@nwalsh.com> wrote:

> / fyl2xp1 <vnhu38f93@subdimension.com> was heard to say:
> | Advantages:
> |
> | XML is the "wave of the future", we can either ride it or sink?

Always reminds me of a Simpson's episode in which Bart is having trouble
at school and (pre-school) Maggie is deemed to be "the wave of future".
Everyone, especially Homer, choruses "wave of the future". There wasn't
anything wrong (!) with Bart just that someone else thought his time had
come. I shall remain convinced that SGML still has life in it.

It is of course nothing more than marketing speak.

> | XML is more Internet oriented, the future is more Internet oriented?
> | XML is simpler and perhaps less cumbersome than SGML?

:-O With namespaces, scehema, XPath, etc in XML I don't hink it's simpler.
One of the earliest documents on XML, which was circulated amongst ISO
committee members, gave a high priority justification for XML as "it can
be implemented by a computing science undergraduate in two weeks". I
wasn't an undergrad (rather the equiv of post-doctoral) and I had a
struggle to write an "XML" parser in two weeks --- well okay it wasn't
helped that it was in programmed in VAX BASIC. That was when XML was
really just SGML without minimisation and the "&" operator in content
models.

That same document suggested that programmers were driving these changes
through and that "we" knew better than the users what should be in a
markup language. I always felt that the reason that SMGL was disliked was
its origina in the printing/publishing arena rather than the result of a
as a piece of pure computing science research.

> | XML is easier to write tools for so the future will bring a bounty of them?

With all the extra bells and whistles that XML has I don't think that
claim can be substantiated.

> Certainly far more XML tools exist now and will exist in the future. I
> won't say that "no new SGML tool will ever be written" because someone
> will surely point one out, but I'm confident in saying that the number
> of XML tools being developed vastly overshadows the number of SGML
> tools.

Perhaps that is a comment on the quality of the present XML tools that a
new one is always needed. The SGML toolchain has been well served by sp
and jade for sometime.

I suspect that the real reasons for more XML tools are its bogus link with
Java and definitely the publishers' bandwagon of producing XML books. If
as many SGML books had been written in the same length of time then there
would be no need for XML.

> | The SGML DocBook will eventually dissappear.
>
> Well, I doubt it'll disappear in my lifetime. Legacy lasts. But new
> development of the SGML version will probably cease eventually (for
> some value of eventually).

I'm glad to hear it. How many SGML DocBook documents are there out there?
That's a huge legacy to convert to XML.

> | Disadvantages:
> |
> | SGML is a maturer technology hence is regarded as better?
>
> If you stick to the bits of XML that have been proven by 10+ years of
> SGML experience, I don't think this argument holds.
>
> | The SGML DocBook toolchain is easier to configure and setup?
>
> That's not my reading of the situation.

'Tis mine.

> | The XML tools are mostly written in JAVA which is slow?
>
> And portable. :-)

And slow. And proprietary, which SGML and XML are not.

> | There are tools available to convert SGML to XML anyway so this is no bother?
>
> Because SGML is dead.

Nah its only resting.

Regards, Trevor

British Sign Language is not inarticulate handwaving; it's a living language.
Support the campaign for formal recognition by the British government now!
Details at http://www.fdp.org.uk/

-- 

<>< Re: deemed!