RE: typing

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Not exactly a troll.  Just sort of. :-)

System ids should always retrieve 
an entity from whereever it is as long 
as the resource is available.  No 
it shouldn't travel with every instance. 
Using a protocol morph in a public 
name is a terrible idea, but we've 
been through that before.  

I don't think the W3C servers can stand 
the punishment of millions of pages 
hitting them at once.  Entity resolution 
of centralized resources doesn't look 
to me like a scalable idea.  Local 
copies do.  

I don't know if it is a bug because 
by this point, surely, they would 
have seen it.  It appears somone 
is objecting to the design and voting 
in code.   That is what I am wondering 
about.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Jonathan Borden [mailto:jborden@mediaone.net]
Sent: Tuesday, March 13, 2001 10:05 AM
To: Bullard, Claude L (Len); XML DEV
Subject: Re: typing


Len

>
> Folks, the external entity idea
> based on a common URL identifier
> doesn't work worth beans.  Why not?
> MS does know and it has been tested
> to death.  Why was it not fixed?
> Maybe URL/URI management of a common
> resources is not such a good idea.
>

Is this a troll?

Suppose I simply want to put the DTDs at an intranet site served by a local
HTTP server what would be the problem with this? Are you really suggesting a
DTD needs to travel with every document instance? Or for that matter every
schema.

Are you defending a clearcut bug on the basis that _you_ don't like the way
XML 1.0 defines a system identifier, or the W3C's XHTML Basic
recommendation?

IMHO documents with DOCTYPES referencing DTDs written either by the W3C or
according to W3C specs ought not outwrite hang IE5. To suggest otherwise is
tatamount to suggesting standards only ought be followed when you personally
agree with them.