Re: Requesting a revision of RFC3023

On Wed, 2003-09-17 at 17:34, MURATA Makoto wrote:
> On Wed, 17 Sep 2003 14:10:53 -0400
> Francois Yergeau <FYergeau@alis.com> wrote:
> 
> > But stating that "most XML is not text for casual users" says that there is
> > no loss in deprecating text/*xml (save perhaps transition issues), the text/
> > top-level buys nothing of value.
> 
> Agreed.

I'm sorry to see two people whose opinions I value so highly
agreeing with a position that so troubles me.  One of the 
most important characteristics of XML, as compared with many,
many competing formats for the storage and/or transmission of
data is that it is textual (in the sense of being conceptually
a sequence of characters, and represented on the wire -- at least
so far -- as such).  Since much of the XML which I care about
is also a digital representation of texts (in the sense of
being natural-language utterances with a certain degree of
intra-document linguistic and thematic cohesion), it troubles
me to think that labeling XML as text buys us nothing of
value.  On the contrary, I think: it stresses two important
facts.

I assume that both of you are thinking primarily of browser 
fallback behavior, and reflecting your view that users will
be able to make nothing of XML source if they are confronted
with it -- in that context, I understand, the proposition you
endorse is at least plausible.  Even there, though, I don't
find it compelling:  much XML is quite legible to naive humans --
as legible as anything displayed without much intelligent
formatting will ever be.  And for some humans, almost all XML
is legible without special tools.  (I recognize that the latter
group is a relatively small minority of the human population.)

-C. M. Sperberg-McQueen

Received on Wednesday, 17 September 2003 23:54:29 UTC