[Prev][Next][Index][Thread]
Re: FRAMESET & NOFRAMES & WWW-CL
At 4:57 PM -0500 1997-02-04, Richard Lynch wrote:
>As near as I can figure...
>
>WWW-CL is screwing up the <NOFRAMES></NOFRAMES> thing.
This is a pretty vague report. Can you be more specific?
Given your vagueness, I'll explain what we do in the CL-HTTP examples.
If people are using the url /cl-http/docs.html, there
is supposed to be a computed redirect that sends them to the
frames version only if they are known to have a frames capable
browser. See the code in
http://wilson.ai.mit.edu/cl-http/sources/examples/export.lisp
The general problem of browser capabilities is supposed to be addressed by
the features extension to HTTP, for which there is a draft specification.
(Anybody wnat to implement it?) At present, cl-http uses the variable
http:*user-agent-capabilities* to record this information. A web site
needs to keep it current (big pain, hence the new standard).
>
>People with old browsers are not seeing anything, or getting a "Document
>contains no data" message.
Can you provide browser version info that don't work?
We would prefer not to use frames/noframes in static html because
it requires duplication of html, ergo, creates a maintenance problem.
>
>This may be strictly an MCL thing, but I haven't delved into it enough to
>even begin to guess. Sorry.
>
>Meanwhile, a guy I maintain a web page for is rather upset... since he
>thinks nobody can see it, when the reality is that anybody who doesn't have
>a really old browser sees it just fine. :-^
>
>Any ideas how to "fix" it. Telling him to install a new browser is
>probably extremely computationally intensive, if ya know what I mean. :-)
If this is your own frames code or html, then I think it is your bug amd not
cl-http's. The macro ns2.0:without-frame-capability can be used to generate
html for browsers that don't do frames. Be careful to get the scoping right on no frames.
Alternatively, the computed redirect discussed is more elegant.
References: