Print

Print


After all the META tag discussions at Warwick, I thought I'd do a bit 
of background reading.  Here's what I have so far, in case it's useful
to anyone else.  Comments welcome!

Martin


1. Starting points

There was much interest in defining a minimal set of metadata which 
could be embedded into HTML documents, and establishing a convention 
for doing this using the "META" tag in the document's "HEAD".

I think it was also seen as desirable to have a way of pointing at a 
richer collection of metadata, in whatever format (Somebody Else's 
Problem?) e.g. via UR[LN]s.


2. What we actually have to play with

The following attributes of META are defined by the HTML DTD in RFC 
1866:

<!-- <META HTTP-EQUIV=...>    HTTP response header name     -->
<!-- <META NAME=...>          Meta-information name         -->
<!-- <META CONTENT="...">     Associated information        -->

Note that if a META element has a HTTP-EQUIV attribute, this may be
turned into HTTP headers at the discretion of the WWW server (but
this isn't implemented by anyone? :-)

We also have the following for the LINK tag, which wasn't discussed 
(that I recall) at Warwick, but seems to be very relevant:

<!-- <LINK HREF="...">  Address of link destination         -->
<!-- <LINK URN="...">   Lasting name of destination         -->
<!-- <LINK REL=...>     Relationship to destination         -->
<!-- <LINK REV=...>     Relationship of destination to this -->
<!-- <LINK TITLE="..."> Title of destination (advisory)     -->
<!-- <LINK METHODS="..."> Operations allowed (advisory)     -->

And finally, the "TITLE" element!


3. Related work

These two documents, currently out as Internet Drafts, propose 
formalisations for the values of the META and LINK attributes, 
respectively.

  The META Tag of HTML
  draft-musella-html-metatag-02.txt
  Davide Musella, National Research Council (Italy)
  January 1996

This suggests ...

        keywords:     to indicate the keywords of the document  
        author:       to indicate the author of the document  
        timestamp:    to indicate when the document is authored  
                      (HTTP-date format)
        expire:       to indicate the expire date of the document  
                      (HTTP-date format)
        language:     to indicate the language of the document  
                      (using ISO3316 code or ISO639 code)
        abstract:     to indicate the abstract of the document  
        organization: to indicate the organization of the author  
        revision:     to indicate the revision number of the
                      document

  Hypertext links in HTML
  draft-ietf-html-relrev-00.txt
  Murray Maloney & Liam Quin, SoftQuad Inc.
  December 1995

This suggests (amongst other things!) ...

      REV=MADE        identify the author or "maker" of an
        HTML document
      REL=AUTHOR      hypertext link to an author.
      REL=COPYRIGHT   hypertext link to a copyright notice.
      REL=DISCLAIMER  hypertext link to a legal disclaimer.
      REL=EDITOR      hypertext link to an editor.
      REL=META        hypertext link to a node which contains
        meta-information related to the current document.
      REL=PUBLISHER   hypertext link to a publisher.
      REL=TRADEMARK   hypertext link to a trademark notice.
      REL=TRANSLATION the target is a translation to another
        language.
      REL=LANG        indicates language of the target
        document.
      REL=OBSOLETES   the target document is a later version
        of the current document
      REV=OBSOLETES   the target document is obsoleted by the
        current document.
      REL=UPDATES     the target document contains revisions
        to the current document.
      REL=DERIVED-FROM the target document was derived from
        the current document
      REV=DERIVED-FROM the current document was derived from
        the target document, perhaps by automatic processing
        or by manual editing.

Think I got the RELs and REVs the right way round :-)


4. Current usage (of META)
                
Altavista - <URL:http://www.altavista.digital.com/cgi-bin/query?pg=ah&wh
at=web>

"It is however possible for you to control how your page is indexed by 
using the META tag to specify additional keywords to index, and a short 
abstract."

Understands "keywords" and "description" values for the META NAME 
attribute.  Says it will index both fields as words, and return the 
description along with the URL in any search results.


MOMspider - <URL:http://www.ics.uci.edu/WebSoft/MOMspider/docs/metainfo.
html>

"As shipped, MOMspider only stores the META elements tagged as 
"Expires", "Owner", and "Reply-To".  However, it is very easy to extend 
MOMspider so that it will look for and store other named metainfo.  
Possibilities include IAFA index items for building site description 
files, graphical coordinates for building spacial maps of webspace, 
etc."


ALIWEB cf. <URL:http://www.ai.mit.edu/tools/site-index.html>

Robert S. Thau's WWW site indexing tool generates IAFA templates 
automatically for ingestion into ALIWEB.  It uses the following values 
for the META NAME attribute:

 "description 
     The value of this attribute should be a description of
     the document which makes sense out of context (as it will
     be seen by people who retrieve it from a global index).
  keywords 
     The value of this attribute should be a few keywords
     describing the content areas addressed by the document. 
  resource-type 
     Indicates what sort of object this is. Currently
     recognized values are document and service, the latter
     being appropriate for search engine cover sheets and the
     like. If the document contains no resource-type meta-tag,
     document is the default. 
  distribution 
     Indicates to what groups of users this document is of
     interest. If the script is properly configured (see below),
     then this meta-variable will determine how the document is
     indexed. If you are preparing a single index for global
     distribution, you don't have to worry about this."

"In addition, the <TITLE>s of your documents are used to fill in the 
Title: fields of the IAFA templates; you must give your documents HTML 
titles if they are to appear in your index."

<<any more web crawler type usage examples for META and LINK??
info.webcrawler.com was down when I wrote this, so I couldn't check up 
on the robots list archive>>


5. Problems

META and LINK tags being empty apparently confuses some software which 
has to parse the HTML.

Buffers may be too small for the values being stuffed into them, e.g. 
abstracts and descriptions.

Generating extra HTTP headers via HTTP-EQUIV isn't a widely implemented 
feature.

<<are these serious problems ? any others ??>>


6. Conclusions

Use LINK and REL=META (or was it REV=META? :-) to point at external 
meta info.

Take a grab bag from the above to form the minimal set of embedded meta 
info!





%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%