why can't a web browser mount a server nfs style and then web interactions can be scripted from there as files?
gavino
kodifik
> ??
In order to do so, you probably must start by transforming Farenheit
degrees into radian.
John Passaniti
Donal K. Fellows
> why can't a web browser mount a server nfs style and then web
> interactions can be scripted from there as files?
Because there are a number of file operations that don't map to basic
HTTP operations. The WebDAV extension to HTTP _is_ mountable; your OS
probably supports it already.
Donal.
gavino
couldnt there be a mapping?
then abstract away the interactions?
something like plan9?
is this perhaps what plan9 does with 9p?
hmm
Jordan Henderson
Gavino,
Why would you think this is a subject for discussion on these language
groups?
Having thrown out these kinds of questions for years and receiving
mostly derision and noise in return, don't you think it might be best
to take your questions to a forum where they'll be received better?
I have an idea! Why don't you start a blog and you can post your
questions there. You're as likely to get real discussion of your
questions there as on these groups.
gavino
wrote:
well because this is lisp forum perhaps?
maybe some lisper has tried to do what I said?
no?
no one would read some obscure blog....do you get that the reason I
post here is to perhaps get a response from someone who has created
something near to what I am talking about?
is it hard to figure that out?
obvious to me....
Richard Owlett
>> Gavino,
>>
>> Why would you think this is a subject for discussion on these language
>> groups?
>>
>> Having thrown out these kinds of questions for years and receiving
>> mostly derision and noise in return, don't you think it might be best
>> to take your questions to a forum where they'll be received better?
>>
>> I have an idea! Why don't you start a blog and you can post your
>> questions there. You're as likely to get real discussion of your
>> questions there as on these groups.
>
> well because this is lisp forum perhaps?
Well if this supposedly relavant to lisp, why are you posting to
comp.lang.tcl, comp.lang.forth, comp.lang.scheme also? ?? ???
*FACTOR* your questions. Post to *ONLY* relavant group!
gavino
s/lisp/tcl
Raffael Cavallaro
> s/lisp/tcl
s/gavino/troll
--
Raffael Cavallaro
fortunatus
conditions - remember dial-up lines?
request - serve - request - serve ...
HTTP documents were individual, hyper-linked together. Requests
quick, no context saved between them. Complete documents served each
request. Add cookies if you want context. Good under assumption of
unreliability/unpredicatbility.
NFS style would serve "blocks" rather than "documents". You don't get
a complete segment of information each time, just some blocks. NFS
style would assume a lot about continuity between requests. What if
the remote mounting entity has died and gone away? I'm not saying it
couldn't be worked out for an unreliable environment, but it might be
a lot more awkward.
fortunatus
read HTML documents from it using your browser - this is totally done
every day esp. in corporate network settings. The brower is using
file reads instead of HTTP requests.
gavino
ah yes ideally
can nfs be improved on ?
gavino
the continuations or closures faking them are interesting idea as well
Uwe Klein
it is a bit of a missmatch.
where in http ( or more precisely the server )
dirlisting is a controllable stacked feature
it is an intrinsic property in nfs.
paradigm wrapping in different directions.
uwe
gavino
I wonder about the performance of nfs/gopher vs http dynamic engines.