[Dev] p2p stuff
nitin at borwankar.com
Fri Oct 25 10:33:09 PDT 2002
See also my post on the "design" list about similar matters.
I am the owner of the JXTA Python project on www.jxta.org and also the
P2P-email project there.
Recently we have booted a JXTA peer using Jython and are (very slowly)
working on a Python implementation of JXTA.
While IRIS and other projects have power technology, JXTA is a flexible
and powerful infrastructure platform that is here today and is
incorporating the best parts of these, such as Distributed Hash Tables,
into future reference implementations.
While jabber is excellent for the IM component of Chandler, I would
strongly recommend taking a look at JXTA as well.
As I said in my "design" post, the Chandler project will otherwise spend
significant effort in re-inventing what has been evolved by the JXTA
community over the past 18+ months.
At last count the members on the JXTA lists number 11,000+.
JXTA embodies a number of powerful P2P primitives expressed as XML-based
protocols and the current project web site has reference implementations
on Java and C ( the latter running on a number of PDA's).
While there is a real information "firehose" problem while first
approaching JXTA, the recent O'Reilly JXTA in a Nutshell is an excellent
way to get started.
Also as i said in my design list post - the Chandler project can get
tremendous leverage from re-using successful open source efforts in
and by focusing on the application architecture for the small group
messaging and collaboration platform that it is meant to be.
Moreover attempting to extend Jabber to do P2P stuff will hit a wall
fast for many architectural reasons which may simply be summarised by
saying it was not designed to be a P2P layer.
It does what it was designed to do excellently and again it would be
good leverage to re-use that.
Overall one suggestion would be to spend some effort in expressing
Chandler's semantics in XML-message based protocols, so that current and
future applications can communicate with it without needing to commit to
part or all of its implementation. The ideal scenario would be a set of
libraries (== API's, interfaces, ....) which represent Chandler
reference implementation and the first release would be AN
implementation of AN application set built on these libraries.
This would be orders of magnitude more effective in opening up this
application area to innovation as it would enable developers from many
adjacent open source efforts
to integrate with the Chandler effort. As expressed currently the
Chandler effort appears to be part application part platform and IMHO
the way to leverage that is to be both - a weel defined API set (the
Chandler toolkit - re-usable by developers in a plug-and-play manner )
and a single powerful set of applications - the Chandler Suite - usable
by end users in a plug-and-play manner.
Elankath, Tarun (Cognizant) wrote:
>Hi to Mitch and all,
>This is a wonderful product idea that you have conceived and I really hope
>that it works out. The world needs such an open-source killer app. Chandler
>+ Mozilla + OpenOffice [PIM + Browser + Office Suite] would complete the
>holy software trio that the world awaits.. :-)
>-->Could you investigate and consider JXTA (www.jxta.org) for P2P as well ?
>A lot of p2p stuff has already been written for JXTA and its architecture
>and design is definitely good.
>-->Incase you (and the community) decide to design your own application p2p
>protocol, you could check out BEEP (which is on the internet standards
>track) at www.beepcore.org
>-->"Interoperability" is a keyword if Chandler needs to ease its way into
>becoming the preffered choice of PIM in peoples lives. Right now, I am
>typing this mail in Outlook, which is _not_ my favourite client, but I need
>to use it at work because our organization uses MS Exchange Server (and its
>communication protocol). If I (and many others) want to use Chandler, it
>would need to work with an Exchange Server.
>I do understand that Chandler is not looking towards this now...but please
>consider providing a powerful plugin architecture/framework, so that people
>can extend it to incorporate such features.
>Your foundation mentions a target of 2002 end for the initial alpha release.
>With no fixed design plans or source code to be seen so far, this target
>looks extremely doubtful.
>You mention that work started on this effort since last year. Even if no
>coding has been done (which sounds diffcult to believe.. all open-source
>freaks love to code...), how about releasing some design-documentation ? I
>mean _something_ concrete should have got done in the period of a year,
>apart from the feature list... :-)
>Chandler is quite exciting. During a project of this scale in Python would
>mean tremendous productivity. Such a project would also really test Python,
>wxWindows and other technologies used...if Chandler succeeds all of the
>technologies used in its development would get a tremendous boost.
>I would love to contribute as a coder to a project like this. As a user of
>open-source I have always felt guilty of not effectively contributing back
>to its development. This project has struck a key with me and I can't wait
>to get started...
>When are you guys getting started ? When you guys thrasing your design and
>dev plans on the web ? More information NEEDED!
>PLUS: Is the dev list dead ???
>All the best,
>This e-mail and any files transmitted with it are for the sole use of the intended recipient(s) and may contain confidential and privileged information.
>If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.
>Any unauthorised review, use, disclosure, dissemination, forwarding, printing or copying of this email or any action taken in reliance on this e-mail is strictly
>prohibited and may be unlawful.
> Visit us at http://www.cognizant.com
More information about the Dev