Re: entity header field

>The TITLE entity header field, like the LINK field have an explicit correlation 
>with the same HTML tag; but MUST a server process an HTML file to extract these
>values or a Server can ignore them even if it receives an explicit request of
>these fields ????

I think the title field is poorly motivated wrt a HTML document. I see little 
advantage in getting the Title of the document up a few packets earlier. There 
is an issue with the HEAD method however, if one does a head then one wishes to 
obtain meta-information relating to the entity without obtainig a body and hence 
Title does have some use. 

Title and link are rather more usefull when dealing with content types which do 
not provide for header or link attributes. It is usefull to add a title to a 
JPEG or GIF, it would have been nice if Link could be used to add links into 
such objects - ie client side image maps.

Overall I'm not too keen on the link tag/header because I don't think that 
anyone has proposed a comprehensive and coherent semanitcs for them.

My view of what link should be is annotations in the Hyper-G model. Ie they 
should be used to inform a client about links which relate to an entity which 
are not a part of it. Annotations are one example.

This model would indicate a need to incorporate a position indicator into the 
tag:

Link: <http://www.cern.ch/TheBook/chapter2>; rel="Previous";
    pos=1654-1754

Where pos in this case is a simple character offset. I suspect that we really 
need a comprehensive notation for describing positions within a hypertext 
structure.


	Phill

Received on Friday, 15 March 1996 18:07:01 UTC