httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From gord...@base.com (Gordon Irlam)
Subject Re: speaker for W3C Distributed Authoring Symposium?
Date Sat, 12 Oct 1996 01:49:11 GMT
> My officemate, Jim Whitehead (ejw@ics.uci.edu), is organizing a symposium
> in the Bay Area about distributed authoring on the Web.  He is looking
> for someone who can speak about what is needed for a server like
> Apache to support PUT and version control capabilities, or someone
> who would like to speak about "what the Apache Group plans to do to
> support such capabilities in the future."  Is anyone strongly interested
> in this topic or would like to speak at a W3C symposium on behalf
> of Apache (assuming we could come up with an actual plan for support
> of those features)?

I'm interested in this topic, and had been planning to attend this
symposium.

I wrote a WikiWikiWeb / WebEdit style cgi script that we use
internally at Cygnus for editing web pages.  Context diffs are
generated whenever someone edits a page and depending on the domain
of the submitter the changes either immediately applied to the page
or sent to the owner of the page for approval.  This was all just
a quick hack.  I had been hoping someone would combine PUT and RCS
professionally and Cygnus could throw my system out (Xerox might
have done this, but last I heard their work wasn't free).

If Cygnus had the resources, this is something I would like us to
be working on.  We currently don't though so I will just be attending
as an interested observer.  If pushed I could possibly give a 10
minute talk on the status of the free s/w that already exists and
needs integrating to create a complete solution (Apache, Amaya, RCS,
CVS, Perl, ...), but I couldn't speak of any plans to do this work.

If there is an Apache person interested in attending this talk let
me know.  Cygnus is allowed to send 2 people to W3C events, and so
we could possibly get you in free.

                                                   gordoni

PS: Here is a blurb I wrote recently on where I imagine this technology
might be heading...

A brief note on the likely merger between version control systems and the web.

Since it's inception HTTP has included a PUT intended to allow people
to edit and put back changed versions of documents retrieved from a
web server.  It is just now that people are starting to take advantage
of this capability.  Microsoft Frontpage, Netscape Gold, and W3C Amaya,
are all browser/editors that now allow people to perform WYSIWYG editing
of web content.  Apache also recently added support for the PUT method.
Both AOL and Microsoft are working hard on standards that will nail down
how PUT is meant to work making it easy for end users to update web content.

>From a technical point of view this work is now all as good as happened.
The next technical steps now being worked on are looking at issues of
version control, and extending GET and PUT to include specifiers for
dealing with version control.

    http://www.ics.uci.edu/~ejw/versioning

provides a good overview of the current state of this technology.

It seems obvious that the eventual technical outcome of this work as
far as source code management is concerned is going to involve web
servers acting as final source code repositories, and tools like cvs
turning into web clients.

If done right, source code version control via the web could have
some major benefits.  The tremendous power of the web is really just
the result of redesigning the features of NFS to make files globally
accessible.  Similarly large benefits are likely to occur in the world
of source code management.  Cygnus s/w is built from 20 different components
from 20 different sources.  If I was able to name these sources and have
my version management system notify me whenever it saw these sources
change and ask me automatically if I wanted to automatically upgrade
these components, as well as automatically attempt to push changes back
for approval, the time taken trying to keep in sync with the net would
drop significantly.

Mime
View raw message