Re: Frames & WWW

Joe English (
Fri, 18 Nov 1994 04:55:42 +0100 (P. M. Hallam-Baker) wrote:
> In article <> you write:
> |> >>>>> "gtn" == Gavin Nicol <> writes:
> |> Phil's comment about SGML being spec'd without implementation isn't strict
> |> true. In fact, the situation is pretty much the reverse... SGML grew out
> |> of IBM's GML- A lot of the peculiarities are the result of implementation
> |> features of GML that slid their way in to the standard. All sorts of whack
> |> restrictions in the standard make a strange sort of sense if you think abo
> |> them in the context of one particular implementation of a parser.

> As for `most people with sense avoid the worse features of this standard'. Le
> ts think
> about that for a moment eh? Basically you are saying that the standard works
> because
> people don't follow it.

I think what [lost attribution] was talking about
are things like, DATATAG, #CURRENT attribute declared
values, CONCUR, empty start tags, and so on. These are
optional features: people either don't use them (empty
start tags, #CURRENT) or don't implement them (DATATAG, CONCUR)
because of their limited utility. No SGML applications or
software that I know of willingly *violate* the standard (except for
WWW documents and browsers, and that's more from ignorance than

[ earlier ]

> I was thinking specifically of the HYTIME extensions that recently were passe
> 'Unanimously' despite there being no implementation. And of course without bo
> thering
> to look at what was already in the field.

Ah, well, HyTime, um...

No comment there.

> As for not reading the TEI spec. I follow the ISO standard, 50 quid it cost m
> e. Are
> you saying that I have the wrong standards body and that TEI is really the de
> finitive
> version? If so we should get this cleared up as soon as possible.

No, TEI is just an example of a very well-thought-out
SGML application. It's completely ISO 8879 compliant.
You shouldn't read it *instead of ISO 8879*, you should
read it *instead of reinventing the wheel*. (I would
suggest checking out IBM ID DOC and InfoMaster, too;
there's a lot of good ideas in there as well.)

> If TEI are willing to send me a copy of their specification without cost I wi
> ll
> read it.

The TEI standard is available on the Web: see

> If anyone is using their standard within the Web it would of course
> have
> greater weight.

TEI is being used on the Web. (Sorry, no URL.
Official Slogan of the WWW: "I saw it on the
Web someplace, but I just can't remember where...")

> But if its simply an unconnected commercial implementation

It's not a commercial endeavor; all the TEI
material is publically available.

> then I am sure the IETF will give it due weight in their considerations.

Since it's one of the more significant advances
in computing technology for the Humanities,
I'm sure the IETF *will* give it some weight..