<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#330033">
An off-group response caused me to realize that my original posting
wasn't clear on the environment in which I am working, nor my "real"
problem, but rather one step removed from that. It is easy to fall
into the trap of asking for what you think you want, based on one's
own limited knowledge, ideas, and available solutions. Here is a
bit more information:<br>
<br>
There are two classes of users of my web site: normal users and
queries are by far the bulk of the requests, and the site has been
optimized for their needs, including transforming the data from its
original sources into a different form. This is a reversible
transformation, and lossless.<br>
<br>
Then there are a smaller set of data suppliers. They provide the
data in the original form, and then it is transformed for the site.
So to make their job easier (and understandable), there is a need to
provide them with the original view of their data, in that form.
But storing it in both forms requires too much space... I can keep
enough metadata to provide the appearance of their original data
files, and it is no problem to perform the queries to supply the
data from their original files when they request it, or to accept
new data when they supply it.<br>
<br>
So probably any protocol would work on the wire, FTP, SFTP, SCP,
WebDAV, etc., as long as I get to define to the server implementing
that protocol what the definition of a file is, in code, preferably
Python code, but Perl would be OK, if something similar enough
preexists.<br>
<br>
I went down the path of not finding a server implementation of any
of the well-known protocols that would allow a definition of fake
files at the back end.... except HTTP. And I'm unable to obtain
control of the server ports to even run a non-standard server
software if it does exist, so I'm limited to HTTP as a transport
protocol, although other protocols could perhaps be proxied over
HTTP, but then standard clients probably can't be used either. So
then I started looking for front ends that I could tweak to speak
HTTP to a custom CGI script. That's where I started this thread,
and why I picked this group to ask in, because I'd find it useful
and interesting educationally to tweak a PyQt front end, instead of
one written in some other environment.<br>
<br>
I'm open to suggestions for tackling the problem from other
directions, though.<br>
<br>
</body>
</html>