Re: Link style sheets [was REL and REV]

Craig Hubley (craig@passport.ca)
Sun, 14 May 95 17:53:04 EDT

> Nope - mailto: specifies an object (a maildrop address), and a
> protocol (usually SMTP). More importantly, if I issue the command
> "lynx mailto:craig@hubley.com" lynx will send mail to you.

Fine. OK, I don't think we want a "linkas" protocol, it's just a
browser behavior.

> What should happen if I issue the command "lynx linkas:html3/next" ?

I jump to the results of this function, which is an HREF, since you did
not give it an argument, the home page can be the default. You go to
your own home page. I admit this is not particularly useful.

I guess you don't see the value of links as first class 'resources' ?
Tim did. R=Resource not R=Destination. Not all URIs are destinations,
necessarily.

> > I really don't see the big issue.
>
> The issue is avoiding things that will lead to user errors.

They can still use the short list of names, if that's all they want.

I just want those names to abbreviate some larger, hierarchical, consistent
namespace of functions, so that when we have 500 of them from 50 vendors
(in the form of new hardwired links, HotJava applets, Tcl programs, ScriptX,
sharable-CGI-scripts, or god knows what else) the author can differentiate,
and communities of usage can safely share, and all these vendors won't need
to invent their own damn tags.

The stylesheet solution is fine. Which function is called to implement a
specific relationship type is definitely a 'style' issue.

All I ask is that there be *some* indirection mechanism so that a more robust
name space than a 'flat' one can be defined, later when it is necessary. I
am also concerned about form and field validation functions, they need the
same treatment eventually.

-- 
Craig Hubley                Business that runs on knowledge
Craig Hubley & Associates   needs software that runs on the net
mailto:craig@hubley.com     416-778-6136    416-778-1965 FAX
Seventy Eaton Avenue, Toronto, Ontario, Canada M4J 2Z5