Re: [Bug-gnupedia] Re: Classification difficulty and incompletene ss

[Mike Warren]

"Thomas E. Vaughan" <address@hidden> writes:
> On Thu, Jan 18, 2001 at 02:38:37PM -0000, Duncan Lock wrote:

> > You could have as many content section as you like, depends on how
> > you write the DTD. You ought to have some metadat for each
> > explaining what it is (ie the HTML version, the latex2html
> > version, with pictures, without pictures, version 1, version 2,
> > etc...)

> That seems reasonable.

I disagree. The DTD should be designed such that all the information
is encoded in one well-known DTD, rather than a bunch of different
ones embedded in meta-data. There could be different content sections
for translations to other languages, but there's no reason why
pictures can't just be embedded in the <content> and -- when
translated -- removed, if this is desired. Similarly, sounds and
movies can all be a well-defined tag in the content. For example:

<content language="English">
Hello. This is a picture of a dog.
<mime type="image/x-png">***</mime>
</content>

where the *** is the actual PNG data for the dog picture. Different
schemes might be derived, like having the pictures in separate
top-level tags altogether and referring to them by ID (this would
allow pictures used multiple times to easily appear only once):

<mime id="unique-id-1234" type="image/x-png">
[data]
</mime>

<article>
...
<content>
Hello. Here is a picture. <mime id="unique-id-1234">
</content>
...
</article>

-- 
address@hidden
<URL:http://www.mike-warren.com>
GPG: 0x579911BD :: 87F2 4D98 BDB0 0E90 EE2A  0CF9 1087 0884 5799 11BD