Hi Paul,
I totally agree with these
code cleanliness, maintainability, community involvement, stability,
production quality, all these and more are criteria on which people
will judge whether this is a better implementation than Sandesha2
However, I do not agree that it they can use to overcome an architecture
that is not sound with respect to the performance.
If the architecture is sound and the performance is good( I am +1 to replace
the Sandesha2 ), the rest is just a matter of few review cycles
Thanks,
Jaliya
----- Original Message -----
From: "Paul Fremantle" <[EMAIL PROTECTED]>
To: "Chamikara Jayalath" <[EMAIL PROTECTED]>
Cc: <[email protected]>
Sent: Tuesday, June 03, 2008 1:08 PM
Subject: Re: Mercury, a new WS-RM implementation
Chamikara
I think we are a long way from a feature complete Mercury :-) I also
think you should be thinking of other criteria than just performance:
code cleanliness, maintainability, community involvement, stability,
production quality, all these and more are criteria on which people
will judge whether this is a better implementation than Sandesha2.
Paul
On Tue, Jun 3, 2008 at 5:50 PM, Chamikara Jayalath
<[EMAIL PROTECTED]> wrote:
Hi All,
As I mentioned earlied, I am really eager to see Mercury being brought
into
Apache (either as a branch in Sandesha2 or as a seperate WS project).
What I
really want to see is a comparison between two implementations. If a
feature
complete Mercury version outperforms Sandesha2 I would be more than happy
to
accept it as a replacement for Sandsha2.
Thanks,
Chamikara
On Tue, Jun 3, 2008 at 10:24 AM, Sanjiva Weerawarana
<[EMAIL PROTECTED]>
wrote:
No objection at all, as long as its not castrated .. i.e., it must be ok
to release. Is that ok with everyone?
Sanjiva.
Paul Fremantle wrote:
Ant
I completely agree. I think the best outcome is to move the code into
ASF SVN, move the discussion over, and simply treat this as a new and
different codebase from Sandesha, with no implication about how this
pans out in the future.
Paul
On Tue, Jun 3, 2008 at 12:23 PM, ant elder <[EMAIL PROTECTED]> wrote:
Just for the record I'd like to point out I've been talking as a WS
PMC
member. I've no idea what IBMs views are on this and haven't spoken to
anyone in IBM about it. Also I'm fine with this code coming into the
apache
svn, that would be better than being developed at wso2 imho. I do
think
for
now until there's consensus on it there should be no implication yet
that
its the "next" version, as outline in the rules for revolutions doc.
...ant
On Tue, Jun 3, 2008 at 12:12 PM, Sanjiva Weerawarana
<[EMAIL PROTECTED]>
wrote:
I'd like to hear what non-IBM Sandesha community members also think
about
this idea. If the entire community doesn't want this code coming
here,
well,
then its ABSOLUTELY fine!
Life will go on .. and if WSO2 wants to it can certainly contrib it
and
start an alternative RM impl in Apache too. Apache does not have a
position
saying "only one of one thing".
Sanjiva.
Paul Fremantle wrote:
1. I agree that this code should not be in Apache with a org.wso2
package namespace.
2. I don't agree that the timing of this is so bad. There is always
a
balance between starting something completely in the open and
pushing
a finished object onto the community. When Amila started this, it
wasn't a conscious decision to replace Sandesha2 - it was simply an
experiment. Amila got it to the point where the experiment proved
that
this particular approach could work and then brought it to the
community. And I don't think this code is that complete - there is a
lot of work to do on it. To be honest I think there are mixed
messages. On the one hand I'm hearing that its too complete and we
should have engaged the community earlier. On the other hand I'm
hearing that because it doesn't implement 1.1. it still hasn't
proven
it can be a cleaner design that Sandesha2. Frankly I don't think
there
is any perfect answer here.
3. Apache has a strong history of allowing multiple implementations
of
the same thing and I don't see any reason why we shouldn't have two
implementations of RM. Whether or not Mercury is a "replacement" for
Sandesha seems to be something that will only will be decided if the
community wants it to be that way. I think its *way* too early to
tell
at this point if Mercury is going to enthuse the community or not.
So
far it only implements about half of the features of Sandesha2.
4. I personally think that the right thing to do here is to engage
the
community. I think the right way to do that is to move the code to
org.apache.something, and move the discussions to sandesha-dev with
a
prefix in the subject like [MERCURY]. This is what we have done time
and again. Either it will engage a wider audience in the design and
implementation or it won't, but unless we do this we can't know.
5. I don't think it really matters whether Sandesha2 has bugs or
not.
The question in my mind is whether this new implementation can be
kept
cleaner, faster, and more maintainable than Sandesha2. If it can be
then it will gain support, and if not it won't. Its as simple as
that.
But I honestly believe in the Apache way, and I don't think it will
end up better without the input of this community, who frankly know
WSRM implementation as well as anyone in the world.
6. I *COMPLETELY* disagree with Ant's point about announcements.
This
is a module that works with Axis2 and there is absolutely nothing
wrong with announcing it to the Axis2 community with an [ANN]
header.
That would be true if this was a commercial extension to Axis2. But
this is an Apache Licensed open source project. Further, a version
of
this code has been donated to Apache. So unless I have somehow an
utter misunderstanding of Apache's mailing lists I cannot see any
reason at all not to post this announcement.
Paul
On Tue, Jun 3, 2008 at 9:18 AM, ant elder <[EMAIL PROTECTED]>
wrote:
On Sat, May 31, 2008 at 4:50 PM, Glen Daniels
<[EMAIL PROTECTED]>
wrote:
<snip>
While that may be true, I'm disappointed that the work moved so
far
forward before being brought to the Sandesha community, and I
would
*really* like to find some navigable path that brings us
eventually
to
a
single implementation of RM-over-Axis2, in Apache.
Agree with that completely, this seems a really sad thing to have
happened.
Is there really no way to get Sandesha working with SecureRM over
SMTP,
there didn't seem to be much discussion about what the issues are
with
doing
that?
While things are as they are I do think things like Mercury
announcements
should be kept off the Apache mailing lists, so no more posts like:
http://apache.markmail.org/message/ounhpi54rx543vqw
...ant
--
Sanjiva Weerawarana, Ph.D.
Founder & Director; Lanka Software Foundation;
http://www.opensource.lk/
Founder, Chairman & CEO; WSO2, Inc.; http://www.wso2.com/
Member; Apache Software Foundation; http://www.apache.org/
Visiting Lecturer; University of Moratuwa; http://www.cse.mrt.ac.lk/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
Sanjiva Weerawarana, Ph.D.
Founder & Director; Lanka Software Foundation; http://www.opensource.lk/
Founder, Chairman & CEO; WSO2, Inc.; http://www.wso2.com/
Member; Apache Software Foundation; http://www.apache.org/
Visiting Lecturer; University of Moratuwa; http://www.cse.mrt.ac.lk/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
Paul Fremantle
Co-Founder and CTO, WSO2
Apache Synapse PMC Chair
OASIS WS-RX TC Co-chair
blog: http://pzf.fremantle.org
[EMAIL PROTECTED]
"Oxygenating the Web Service Platform", www.wso2.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
