Re: MCF's new implementation (via XML)

Andrew Daviel (
Mon, 16 Jun 1997 12:55:51 -0700 (PDT)

Date: Mon, 16 Jun 1997 12:55:51 -0700 (PDT)
From: Andrew Daviel <>
Subject: Re: MCF's new implementation (via XML)
In-Reply-To: <>
Message-ID: <>

On Mon, 16 Jun 1997, Nick Arnett wrote:

> At 06:47 AM 6/16/97 -0700, Benjamin Franz wrote:
> >Examining Netscape's proposal, it is MUCH too complex for successful
> >widespread use without automation tools completely hiding it from the
> >end user: Its a "programmer's language."


Yes, one would need automation tools, at least to start with (editing
existing data fields should be relatively simple). A CGI script, Java
program etc. would suffice, I think, for many cases.

I was less impressed with the original MCF at Yahoo, etc. Given the rich
metadata possible in MCF, the display on the plugin was interesting but
functionally limited - you could just say "Wow! These guys have a lot of
pages!" I think, though, that this data was determined in a simple-minded
way from page titles and URLs, and doesn't fairly represent the

> Given that approach, I'd be interested in hearing other critiques of the
> proposal.

Will it fly ? Well ... maybe, if some major search engine uses it to
achieve  significantly better matches than fulltext. Given that some 5%
or more can't manage to use <TITLE> in HTML, I don't suppose it will
acheive more than 15% penetration. Academics and others who are used to
think in terms of abstracts, references, publication dates etc. may
be more enthusiastic.

Re. linking MCF data - XML: wait and see. HTML: "presumably the HTML LINK
element would be used ... "
I proposed using <LINK REV=META SCHEME=MCF in HTML (or Link: ..
;scheme=MCF; rev=META in HTTP for non-text objects) as per, where the META modifier indicates that
the link is to/from metadata (and not an index, related document, etc.),
and the SCHEME modifier indicates it's MCF (and not FGDC, plaintext, or
some other scheme).

What's the prognosis for XML, anyway ? Is it going to gradually take over
from HTML as a repository of cool features ? PDF seems to have achieved
a niche for product datasheets, but I see it as a hiccup in the smooth
flow of HTML (start another process, crack another window, wait ...)

Andrew Daviel
TRIUMF & Vancouver Webpages