dotgnu-general
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [DotGNU]DotGNU task list


From: S11001001
Subject: Re: [DotGNU]DotGNU task list
Date: Fri, 01 Feb 2002 17:45:03 -0600
User-agent: Mozilla/5.0 (X11; U; Linux i586; en-US; rv:0.9.7+) Gecko/20020106

Bill Lance wrote:

--- "Gopal.V" <address@hidden> wrote:
:
:

  <li>program for setting up an internet

presence, coordinating free softwar
        A group development app is essential, the starting
point might
be the virtual p2p server.

:
:
:

        * Provide a P2P development environment
                - develop a *SCALABLE* loosely coupled cluster for

                  data serving for load sharing and failover.
                        - this looks a lot like the P2P dataserver



Assuming your talking about the Virtual Remote Server
project that's recently surfaced, I think making that
possible depends on these items:



        * Rewrite savannah using phpgwapi so that it
supports XMLRPC

:
:

                - extend it to work using webservices (after
XMLRPC)



But even more specificly, those webservices would need
to run as callable components, and they would need to
be packaged into 'Datasets'.  A 'Dataset' is a
self-contained object, holding both data and
executable methods.  It's stored and retrieved as a
single file.

I mentioned earlier that the system will not do
dynamic, script-based services.  Chris Smith has
pointed out a possible way of doing this where a VRS
component acts as a proxy to a traditional server
running in the normal manner.  This may serve to hide
the location of the host 'script server' if that's an
issue.  But other things being equal, you'd probably
be better off just going directly to the host server
in the first place.

Assuming that Savannah is transmorgified into an RPC
based thing, in what manner would a user access it? RPC components are for program-to-program interface,
not human-to-program.  Some kind of client like GUI
front-end would be needed.  That's the one nifty thing
the browser does for the WWW, the universal User
Interface.


I agree that it simplifies things, but it is intended to be a document

format, not an application user interface. A full-blown program could

provide not only a better UI, but also the communication between programs
I called for in the Development Environment proposal. (i.e. you fix things
in the emacs major mode, and can instantly apply them to Savannah.)


--
Linus?  Whose that?
        -- clueless newbie on #Linux



reply via email to

[Prev in Thread] Current Thread [Next in Thread]