Alphas, betas, and releases, oh my! [Was: VOTE]

[Sam Ruby <ru...@us.ibm.com>]

Sanjiva Weerawarana wrote:
>
> I am personally of the opinion that this project is losing focus on
> the number one priority: getting a completed SOAP implementation done
> that busts the crap out of Apache SOAP in terms of performance and
> usability etc. etc. and competes with the best of them out there. Yes,
> WSDL generation / stub generation all that is useful and important,
> but if the underlying SOAP engine ain't done, who cares? Apache SOAP
> never had integrated stub gen stuff (there was only the WSDL Toolkit
> via alphaWorks AFAIK), but that didn't prevent it from being used for
> some useful (and certainly some useless) things. To me the priority
> is getting SOAP v1.1 suppport and SOAP with Attachments support
> completed ASAP in the form of a quality v1.0 distribution of Apache
> Axis. None of this alpha/beta/gamma crap - its time to buckle down
> and get the basics completed.

Very little work is being done on the underlying SOAP engine at this point.
It essentially is done, modulo minor tweaks, bug fixes, and performance
work.  You may not see the value in alphas and betas, but it does encourage
more people to bang on the the core engine while the rest is coming online.

> Flame away ..

Got any patches to contribute?

> In a related note, I think I'm going to advocate doing an Apache
> SOAP v2.3 before the end of the year to make the various small
> functional and interop fixes that have gone into Apache SOAP come
> into a stable release. The time 2.2 was done in May, the idea was
> that Axis would be done in the summer and that 2.2 was it. Its now
> October and Axis is still in alpha2 with no visible progress on
> the attachments stuff (which IMO is a performance killer). So I
> think its justified in putting out (hopefully) one final, final
> version of Apache SOAP.

+1 on a 2.3 in principle; but please, please, please, spend a moment or
three and write down a plan.  And if it is a release that you want to be
enduring, I would encourage the plan to include posting a release candidate
a week or so before it is official, and locking down changes between that
point and the official release.

- Sam Ruby

AW: Alphas, betas, and releases, oh my! [Was: VOTE]

[Matthew Langham <ml...@sundn.de>]

I would like to jump in here and please don't get me wrong - I don't mean to
tread on anyone's toes.

Axis is currently the "Soap engine" of choice in the Cocoon project an we
have now started to use this duo in a real-world application and came
accross several issues that prevented us from getting something up and
running (proxy, Apache Soap 2.2 Vector). These problems we sent to the list
but received little (or no) reply.

So in the end we fixed the things ourselves and Carsten submitted these
patches (httpProxy, VectorDeserializer) - but they seem to take long to make
it into the distribution (the VectorDeserializer is still not there) - why
is this? Over on the Cocoon project we are used to short submittal cycles
and find the Axis project be rather slow when it comes to this.

I think that in order to get Axis into real world scenarios (if that is the
goal) the project needs to put more emphasis on these types of problems in
addition to what Sanjiva writes (and I agree on his points).

So - what is the answer? Should Cocoon switch back to using Apache Soap 2.x?

Just my 0.02$

Matthew

--
Open Source Group               sunShine - Lighting up e:Business
=================================================================
Matthew Langham, S&N AG, Klingenderstrasse 5, D-33100 Paderborn
Tel: +49-5251-1581-30   [mlangham@sundn.de - http://www.sundn.de]
=================================================================


-----Ursprungliche Nachricht-----
Von: Sam Ruby [mailto:rubys@us.ibm.com]
Gesendet: Donnerstag, 4. Oktober 2001 13:08
An: axis-dev@xml.apache.org
Betreff: Alphas, betas, and releases, oh my! [Was: VOTE]


Sanjiva Weerawarana wrote:
>
> I am personally of the opinion that this project is losing focus on
> the number one priority: getting a completed SOAP implementation done
> that busts the crap out of Apache SOAP in terms of performance and
> usability etc. etc. and competes with the best of them out there. Yes,
> WSDL generation / stub generation all that is useful and important,
> but if the underlying SOAP engine ain't done, who cares? Apache SOAP
> never had integrated stub gen stuff (there was only the WSDL Toolkit
> via alphaWorks AFAIK), but that didn't prevent it from being used for
> some useful (and certainly some useless) things. To me the priority
> is getting SOAP v1.1 suppport and SOAP with Attachments support
> completed ASAP in the form of a quality v1.0 distribution of Apache
> Axis. None of this alpha/beta/gamma crap - its time to buckle down
> and get the basics completed.

Very little work is being done on the underlying SOAP engine at this point.
It essentially is done, modulo minor tweaks, bug fixes, and performance
work.  You may not see the value in alphas and betas, but it does encourage
more people to bang on the the core engine while the rest is coming online.

> Flame away ..

Got any patches to contribute?

> In a related note, I think I'm going to advocate doing an Apache
> SOAP v2.3 before the end of the year to make the various small
> functional and interop fixes that have gone into Apache SOAP come
> into a stable release. The time 2.2 was done in May, the idea was
> that Axis would be done in the summer and that 2.2 was it. Its now
> October and Axis is still in alpha2 with no visible progress on
> the attachments stuff (which IMO is a performance killer). So I
> think its justified in putting out (hopefully) one final, final
> version of Apache SOAP.

+1 on a 2.3 in principle; but please, please, please, spend a moment or
three and write down a plan.  And if it is a release that you want to be
enduring, I would encourage the plan to include posting a release candidate
a week or so before it is official, and locking down changes between that
point and the official release.

- Sam Ruby