You are viewing a plain text version of this content. The canonical link for it is here.
Posted to server-dev@james.apache.org by Danny Angus <da...@apache.org> on 2003/04/29 11:09:30 UTC

RE: [dev-groupware] OOogw server development

Vincent,

You say..

> I've been looking at the Apache James project (http://james.apache.org/),
> but it seems that the current design is not well suited for full groupware
> functionality and I don't think they will change their basic design to
> accomodate these in the near future.

Not quite sure what you base these two statements on, I'd particularly welcome some elaboration of "the current design is not well suited for full groupware functionality". It is an aim, already stated on this list, of the James project to develop James into a groupware server.
I'll enumerate James' current position and plans in response to each of your proposed features..

Of course I'm not trying to denegrate your efforts, I hope your project suceeds, but I wouldn't like to stop the wrong impression of James from being spread on this list. James is a very highly flexible and configurable server proven in commercial operation, James is designed to be extensible, and I would encourage you to look again at working with James to develop the groupware functionality you would like.

 
> I'm planning to implement the following features:
> - JDBC based storage for:
>   - system users
James has this.

>   - directory data and schema(LDAP/X.500)
James has some LDAP, but this could use attention.

>   - messages and other items (single instance!), organised in Folders and
> Queues
James has this.

> - Modular design, allowing quick intergration of new protocols
James has this. The Avalon Phoenix platform which James is built upon allows new protocols to be added at something closely aproaching the least possible delta.

> - Dynamic class loading for:
>   - SocketListeners (listening servers)
>   - MaintenanceTasks (integrity checks, delivery retry etc.)
>   - and possibly BEEPProfiles (like iCAP)
>   This will allow extentions to be loaded without having to recompile the
> code.
James has a high degree of this, in fact most of the domain specific functionailty is assembled at runtime through the use of configuration files to specify what classes should be used to provide which services.

> - HTTP based server Administration
James doesn't have this but does have an RMI administration interface which can be used by a web app.

>   - server status
James RMI doesn't currently offer this.

>   - stopping (parts of) the server
James RMI can only manage full shutdown

>   - (re)starting parts of the server
This is not applicable because of the above.

>   - user management
This is provided.

>   - JDBC store access (read-only for debugging purpose)
We assume that 3rd party SQL tools can be used for this, but obviously a web administration application could provide this for completeness.

> - support for the following protocols:
>   - IMAP4rev1
Under development, currently stuck while we re-design the repository interfaces around JavaMail.

>   - POP3
James provides this

>   - (E)SMTP
And this

>   - LDAP v3
And this

>   - iCAP (with underlying BEEP protocol)
James has no plans in this direction at present.


I mean this in the best spirit, and no disrespect to you but I think that James *does* in fact provide a good platform for extension to provide a groupware server, and I hope that not too many people will be put off James by your unsubstantiated dismissal of it.


d.


---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


Re: [dev-groupware] OOogw server development (James)

Posted by Serge Knystautas <se...@lokitech.com>.
William L. Terry wrote:
> My questions for the James Project group are:
> 
> 1. OK so you have a server, what about the client end?  What use is
>    replacing MS Exchange, if you still end up using Outlook or
>    Outlook express?  What are the folks that are using this server
>    doing for the client end?

No idea.  I use Mozilla, but that's just because I couldn't stand to run 
Exchange.  Basically I've forgone groupware features for now. :(

Ideally I would say we mimic and integrate with Exchange, then improve 
on it.  Standard software strategy stuff... Microsoft is one of the best 
at this.

> 2. Is there a value add that the OpenOffice.org group can provide?
>    How can we better cooperate, or should we just leave it at providing
>    a link to your site?

I would say you have a lot of groupware expertise that could really 
benefit James.  We would like to get James there, but people with clear 
understanding of what a groupware server should do, a sense of the big 
picture, and the ability to quickly point out planning flaws like "What 
are we doing about the client", are very useful.

Most of the James developers have recently signed up to OOogw listserv, 
and maybe you could come check us out more...the people, the code... I 
think the vision is relatively in-line.  I'm not against link swapping 
as well if we think that's useful.

-- 
Serge Knystautas
President
Lokitech >> software . strategy . design >> http://www.lokitech.com
p. 301.656.5501
e. sergek@lokitech.com


---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


RE: [dev-groupware] OOogw server development (James)

Posted by Danny Angus <da...@apache.org>.
William,

My personal answers..

> My questions for the James Project group are:
> 
> 1. OK so you have a server, what about the client end?  What use is
>     replacing MS Exchange, if you still end up using Outlook or
>     Outlook express?  What are the folks that are using this server
>     doing for the client end?

I'm still using outlook. But would love to be use OOGW :-)

> 2. Is there a value add that the OpenOffice.org group can provide?

Yes I think so, if we co-operate we can create a feedback loop and ensure that the client has an open source server and vice versa, and that they interoperate smoothly, a definite value add in the "marketing" stakes.

>     How can we better cooperate, or should we just leave it at providing
>     a link to your site?

I think a link is enough for now, the real value is for us to lurk on each others lists and cross fertilize.
I'm not about to suggest we should join forces, first of all its not for me to say, and secondly I think we can achieve just as much rubbing along as we already are, with genuine goodwill on both sides.

d.
 


---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


RE: [dev-groupware] OOogw server development (James)

Posted by Vincent de Lau <op...@delau.nl>.
> -----Original Message-----
> From: William L. Terry [mailto:bill@sweye.com]
>
> I have to play the devils advocate here and ask...  How does this
> project fit into the OpenOffice.org framework?  I think some very
> powerful points have been made about the James project, and I for
> one am going to give it a *much* closer look, but back to the
> question, how does this relate to OOo.  I am not trying to be the
> voice of dissent here, rather I would like to open this up for
> a broader discussion.

I think the only requirement we have for a OOogw server is to provide the
services the OOogw client needs. I think consesus is reached that we should
use open standards, so that the client can use any server and the server can
handle any client. However, OOo should be packed with a server that operates
with the OOogw client (probably GLOW) and serves all its needs.

> We also still have the much bigger problem of the client portion.
> My high hopes for the GLOW project has dimmed somewhat as the time
> stretches out, but it does at this point still seem to be the best
> hope.  I have also looked somewhat at the Chandler Project, and at
> this point I am not jumping through hoops with excitement.  I do
> realize it is in very early alpha, but IMHO I am not excited about
> the direction they are taking.  I would love other folks feedback
> on this client.  I will go into more detail when I feel that I
> understand the project better.
>
> My questions for the James Project group are:
>
> 1. OK so you have a server, what about the client end?  What use is
>     replacing MS Exchange, if you still end up using Outlook or
>     Outlook express?  What are the folks that are using this server
>     doing for the client end?

I think for OOo, providing a standards complying client is basically enough.
It is up to the administrator to choose a server that fits his or her needs.
Providing a reference server might however be usefull and I think that if
the server performs well and provide all the client supported protocols, 90%
of the administrators will stick to it.

On the other hand, the server should not be limited to handle only the OOogw
client. Especially during migration and in cross-platform environments,
other clients are being used.

> 2. Is there a value add that the OpenOffice.org group can provide?
>     How can we better cooperate, or should we just leave it at providing
>     a link to your site?

I think it is the responsibility of the OOogw community to:
- select an existing product and help to develop it to make sure it fits in
with OOo, or;
- create a a server product themselves

In my current opinion (changed a bit since the last few days), James is a
good place to start. If however James can not become what we want it to be,
I'm willing to continue working on a new product. However, the emails I've
read the last few days, give me some confidense that James will become what
we wan't it to be and I'll see in what extend I can contribute to that
effort.

> One thing I must mention to the OOogw group: it seems we keep waiting
> for the perfect project to roll in.  Something looks promising
> down the road (i.e., Chandler, GLOW, etc) and we wait for it to arrive.
> This is why I gave a kudos to Vincent, at least he said screw it, I
> doing something (figuratively, not in so many words ;-) ).   On a
> personal side I feel that I have enough skills to contribute to
> a project, but not enough skills to *start* one. I am sure there are
> others in the same position.  So where do we need to go from here?

I think it is decission making time for the OOogw community, although we
have a history of not being good at making decissions. Sun kind of made one
decission be assigning developers to a client project, but we have to make
one ourselfs for the server part.

My questions to the OOo(gw) community:
- Do you agree that we have to provide a reference groupware server with
OOo?
- Do you feel convident that the Apache James project is a good starting
point for such a server?

If both answers are yes, I think we should concentrate on working with the
Apache James project to implement all the features we need.

Vincent de Lau
 vincent@delau.nl


---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


[dev-groupware] OOogw server development (James)

Posted by "William L. Terry" <bi...@sweye.com>.
Noel J. Bergman wrote:
> Following up on Danny's reply ...
> 
> The goal of the Apache James project is to evolve James into a full
> groupware server.  The purpose of the James v2 branch is to preserve that
> stable MTA while we evolve James v3.  The vast majority of your list are
> already present in James, and those that aren't are already in the plan for
> James v3.  We'd be happy to work on them with you.
> 

I have to play the devils advocate here and ask...  How does this
project fit into the OpenOffice.org framework?  I think some very
powerful points have been made about the James project, and I for
one am going to give it a *much* closer look, but back to the
question, how does this relate to OOo.  I am not trying to be the
voice of dissent here, rather I would like to open this up for
a broader discussion.

We also still have the much bigger problem of the client portion.
My high hopes for the GLOW project has dimmed somewhat as the time
stretches out, but it does at this point still seem to be the best
hope.  I have also looked somewhat at the Chandler Project, and at
this point I am not jumping through hoops with excitement.  I do
realize it is in very early alpha, but IMHO I am not excited about
the direction they are taking.  I would love other folks feedback
on this client.  I will go into more detail when I feel that I
understand the project better.

My questions for the James Project group are:

1. OK so you have a server, what about the client end?  What use is
    replacing MS Exchange, if you still end up using Outlook or
    Outlook express?  What are the folks that are using this server
    doing for the client end?

2. Is there a value add that the OpenOffice.org group can provide?
    How can we better cooperate, or should we just leave it at providing
    a link to your site?


One thing I must mention to the OOogw group: it seems we keep waiting
for the perfect project to roll in.  Something looks promising
down the road (i.e., Chandler, GLOW, etc) and we wait for it to arrive.
This is why I gave a kudos to Vincent, at least he said screw it, I
doing something (figuratively, not in so many words ;-) ).   On a
personal side I feel that I have enough skills to contribute to
a project, but not enough skills to *start* one. I am sure there are
others in the same position.  So where do we need to go from here?

> 
> I hope that this helps, and we invite you (and anyone else who is
> interested) to participate on the James project.
> 
> 	--- Noel
> 



-- 
William L. Terry               (bill@sweye.com)



---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


RE: [dev-groupware] OOogw server development

Posted by "Noel J. Bergman" <no...@devtech.com>.
Following up on Danny's reply ...

The goal of the Apache James project is to evolve James into a full
groupware server.  The purpose of the James v2 branch is to preserve that
stable MTA while we evolve James v3.  The vast majority of your list are
already present in James, and those that aren't are already in the plan for
James v3.  We'd be happy to work on them with you.

Plans exist for James to migrate from its proprietary user repository to
JNDI, thus providing a standard system for locating, accessing, and
extending user data.  Likewise, plans exist for James to migrate from its
proprietary message stores to JavaMail, with appropriate Store providers.
Those issues were discussed earlier, and are recorded in the archives.  The
administrative interface hasn't had as much discussion, in part because
Avalon is adding JMX support, and we should leverage that new ability.
Tomcat is likewise moving to a JMX solution.

You say that "I don't think THEY will change their basic design to
accomodate these in the near future" [emphasis mine].  That statement is
quite wrong, and I really don't know how you arrived at it.  James is an
Open Source project, driven by the people who choose to work on it.  WE have
plans to change the design, and WE invite you to become part of the
solution.

> I already posted some of my ideas on James/groupware on james-dev,
although
> nobody seems to have seen it or didn't find it interesting enough to
reply:

Your message last week was a followup to related questions by Sid Stuary
that were about topics that had been discussed, but were not recorded other
than in the message archives.  It is expected that newcomers will ask
questions without searching the mailing list archives (we all do the same
thing), but it really does take time to go back over the same territory, so
I summarized some of the state in the Wiki, and posted a general note about
it.  Using the Wiki allows ideas to be preserved and evolved without having
to start each discussion from scratch.

I hope that this helps, and we invite you (and anyone else who is
interested) to participate on the James project.

	--- Noel


---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


RE: [dev-groupware] OOogw server development

Posted by Vincent de Lau <vi...@delau.nl>.
(If this doesn't appear in james-dev I'll repost it there. I'm using
different e-mail addresses for the two lists.)
> -----Original Message-----
> From: Danny Angus [mailto:danny@apache.org]
> Sent: Tuesday, April 29, 2003 11:10 AM
> To: dev@groupware.openoffice.org
> Cc: James Developer List
> Subject: RE: [dev-groupware] OOogw server development
>
>
> Vincent,
>
> You say..
>
> > I've been looking at the Apache James project
> (http://james.apache.org/),
> > but it seems that the current design is not well suited for
> full groupware
> > functionality and I don't think they will change their basic design to
> > accomodate these in the near future.
>
> Not quite sure what you base these two statements on, I'd
> particularly welcome some elaboration of "the current design is
> not well suited for full groupware functionality". It is an aim,
> already stated on this list, of the James project to develop
> James into a groupware server.

First of all, let me make it perfectly clear that I find James a good
product, but IMHO it is more designed as a MTA then as a groupware server. I
should have stated it as my (humble) opinion, not as a fact. Also the
language barrier might have sprung up a bit (natively Dutch).

I already posted some of my ideas on James/groupware on james-dev, although
nobody seems to have seen it or didn't find it interesting enough to reply:
- Storage architecture (was: Serializing Users in FilesForwarding mail to
NNTP) (21-4 0:43 GMT)
- JavaMail usage in future James and users vs. recipients (21-4 1:43 GMT)

I think that a basic groupware server should at least provide:
- mail transport (SMTP)
- mail storage and access (POP3 and IMAP4)
- shared folders (NNTP and IMAP4)
- shared directory (LDAPv3)
- shared calendars (iCAP, still in RFC draft status)
- distribution list (mailing list) support

You could of course combine separate servers to make a complete groupware
server, but I think only an integrated solution can unleash the full power
of groupware. With integrated I mean:
- working on the same stores
- working with the same user accounts
- providing a single administration interface

With this philosophy in mind, I think that the real core of a groupware
system is its storage facility. As long as that part is operating, all other
parts of the system can run independently of each other. You could for
instance stop you NNTP and SMTP server (for configuration updates or to stop
a virus from spreading) while still providing POP3/IMAP4, iCAP and LDAP.
This adds up into another part of my core system: server management. This
part of the system is responsible for starting and stopping services and
starting maintenance tasks (integrity checks, cleanup, archiving). It should
also be able to detect failures and act upon them.

If I look at the current (2.1 line) state of the system the parts that are
implemented and reached some level of maturity are MTA based: recieving,
filtering, sending and storing messages and simple acces through POP3. IMAP4
is hard to implement since it requires a different perspective on message
stores, NNTP is a completely seperate subsystem (correct me if I'm wrong,
but I think it is only using the same user accounts) and I couldn't find any
part of a LDAP server.
JavaMail might solve some of the problems at hand by providing a single
interface to the message storage subsystem, but there are no plans AFAIK at
the moment to implement something similar for user/recipient data
(directory).

> I'll enumerate James' current position and plans in response to
> each of your proposed features..
>
> Of course I'm not trying to denegrate your efforts, I hope your
> project suceeds, but I wouldn't like to stop the wrong impression
> of James from being spread on this list. James is a very highly
> flexible and configurable server proven in commercial operation,
> James is designed to be extensible, and I would encourage you to
> look again at working with James to develop the groupware
> functionality you would like.

If you think James can be the product I described above, I'd be glad to help
out. That's why I started two topics on the james-dev list.

Some comments question on your comments:

> > I'm planning to implement the following features:
> > - JDBC based storage for:
> >   - system users
> James has this.
> >   - directory data and schema(LDAP/X.500)
> James has some LDAP, but this could use attention.

The only thing I could find in the source is LDAP support for user stores,
not a LDAP server implementation.

> >   - messages and other items (single instance!), organised in
> >     Folders and Queues
> James has this.

Not single instance storage (covering folders and spools), if I'm correct.

> > - Modular design, allowing quick intergration of new protocols
> James has this. The Avalon Phoenix platform which James is built
> upon allows new protocols to be added at something closely
> aproaching the least possible delta.

<Blow below the belt:> not modular enough to allow IMAP4 to be implemented
without the major store overhaul

> > - Dynamic class loading for:
> >   - SocketListeners (listening servers)
> >   - MaintenanceTasks (integrity checks, delivery retry etc.)
> >   - and possibly BEEPProfiles (like iCAP)
> >   This will allow extentions to be loaded without having to
> recompile the
> > code.
> James has a high degree of this, in fact most of the domain
> specific functionailty is assembled at runtime through the use of
> configuration files to specify what classes should be used to
> provide which services.

This feature is not high on my spec list, but I thougt it would be nice if
you could drop a jar into the classpath and trough an admin interface:
:> load net.delau.geserver.MyNewProto.MyNewProtoListner
:> config mynewproto
...
:> start mynewproto

> > - HTTP based server Administration
> James doesn't have this but does have an RMI administration
> interface which can be used by a web app.
>
> >   - server status
> James RMI doesn't currently offer this.
>
> >   - stopping (parts of) the server
> James RMI can only manage full shutdown
>
> >   - (re)starting parts of the server
> This is not applicable because of the above.
>
> >   - user management
> This is provided.
>
> >   - JDBC store access (read-only for debugging purpose)
> We assume that 3rd party SQL tools can be used for this, but
> obviously a web administration application could provide this for
> completeness.

I should have left JDBC out. I meant a way to look at your store trough the
store logic (not raw table access).

> > - support for the following protocols:
> >   - IMAP4rev1
> Under development, currently stuck while we re-design the
> repository interfaces around JavaMail.
>
> >   - POP3
> James provides this
>
> >   - (E)SMTP
> And this
>
> >   - LDAP v3
> And this
>
> >   - iCAP (with underlying BEEP protocol)
> James has no plans in this direction at present.

I think this would be a requirement for a groupware server.

> I mean this in the best spirit, and no disrespect to you but I
> think that James *does* in fact provide a good platform for
> extension to provide a groupware server, and I hope that not too
> many people will be put off James by your unsubstantiated dismissal of it.

Again, I have no problem with James, but I wouldn't call it a groupware
solution yet. I needs more funcionality than it currently provides and if we
think of a solid way to enhance the James core to be able to provide them,
James would be a far better product then the thing I'm developing.

Another quick ntoe to keep in mind: I've been an MS Exchange administrator
for a couple of years and I have had numerous discussion about Exchange vs.
open source solutions. A lot of interesting things have been proposed to me
for replacing Exchange, but they always failed one of these criterea:
- easy administration (ever seen Exchange Administrator?)
- shared folders (mailbox sharing or 'public folders')
- calendar support

I hope that I've answered your questions. Back to you... or anybody else who
has to say anything usefull on the topic.

Vincent de Lau
 vincent@delau.nl



---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org