Re: Tag proposal

Benjamin C. W. Sittler (bsittler@mailhost.nmt.edu)
Fri, 8 Mar 1996 11:42:28 -0700 (MST)


Date: Fri, 8 Mar 1996 11:42:28 -0700 (MST)
From: "Benjamin C. W. Sittler" <bsittler@mailhost.nmt.edu>
To: Ka-Ping Yee <kryee@novice.uwaterloo.ca>
Cc: Zenon Panoussis <oracle@stockholm.mail.telia.com>, www-html@w3.org
Subject: Re: Tag proposal
In-Reply-To: <Pine.3.87.9603081133.B32176-0100000@novice.uwaterloo.ca>
Message-Id: <Pine.SUN.3.91.960308112945.24219A-100000@gray>

On Fri, 8 Mar 1996, Ka-Ping Yee wrote:

> On Fri, 8 Mar 1996, Zenon Panoussis wrote:
> 
> > New HTML link tag proposal.
> [...]
> > The trouble with links is that they can take you anywhere, and you never 
> > know beforehand where that will be. What hides behind a link can be the 
> > final answer to the main question, or a silly connection to somebody's 
> > silly home page. You never know beforehand.
> 
> I agree that knowing the role of a given link is important.  However,
> there has been a way to do this in HTML for a very long time -- it's
> just that browser implementors have, for the most part, ignored it.
> 
> > we also interpret "Back" and "Next" 
> > relatively when they come from the keyboards of different web authors.
> > 
> > Three more directions are useful. "Up", which is also used in documents 
> > today in the sense of "towards the more general", "Down", indicating a 
> > dive in a particular issue that is deeper than the general scope of the 
> > actual document and "Sideways", which could indicate either a side 
> > issue within the scope of the document or an interesting but totally 
> > different issue. "Sideways" could be split in its two possible meanings 
> > or not. Appropriate tags would then be:
> > 
> > ANG=BK  ANG=FW  ANG=DN  ANG=UP  ANG=SD  or
> > ANG=BK  ANG=FW  ANG=DN  ANG=UP  ANG=SD  ANG=OU
> > 
> > meaning respectively back, forward, down, up, sideways and out. Embedding
> > these tags in HREF-statements should present no problems.
> 
> What you are proposing is not a tag, but in fact an attribute to the anchor
> tags (A and LINK).  But as it turns out, such an attribute already exists.  
> Actually, there are two of them: REL and REV.  REL specifies the 
> relationship of the target to the anchor, and REV specifies the reverse 
> relationship (of the anchor to the target).  Common practice for 
> identifying the author of a page, for instance, is to use
> 
> <link rev="MADE" href="mailto:the-author">
> 
> in the document's HEAD, thus indicating that "the-author" MADE the 
> containing document.  There are several useful suggested values for REL 
> and REV in a discussion paper on the topic by Murray Maloney, which you 
> can find at
> 
> http://ogopogo.nttc.edu/tools/html/mmaloney_links.html
> 
> So, in short, the mechanism to do what you want is certainly there.
> All we have to do is convince a browser manufacturer to implement
> this mechanism.

emacs-w3 [1] (works on any machine that runs emacs) and UdiWWW [2] (for
several varieties of Windoze) both make use of the LINK tag.  emacs-w3
builds a menu of documents referenced by LINKs. UdiWWW's dynamic button
bar is also a nice example implementation. If you want to test a browser
on a page with these links (although I can't say that they are used very
well) see my home page. [3]

[1] The Emacs World Wide Web Browser
    http://www.cs.indiana.edu/elisp/w3/docs.html

[2] UdiWWW - HTML3 Browser
    http://www.uni-ulm.de/~richter/udiwww/index.htm

[3] Benjamin Sittler's Home Page
    http://www.nmt.edu/~bsittler/homepage.html