Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [epp-dev] A new EPP SOA package

I would like to point out that it hasn't been/is not at all possible for 'anyone' to specify a package and have it created/deployed as a package at EF. As a concrete example: https://bugs.eclipse.org/bugs/show_bug.cgi?id=218829. The 'process' for deciding upon new packages and their contents remains rather mysterious (to me, anyway).

Scott

Ian Skerrett wrote:
IMHO, ideally anyone should be able to specify a package and the EPP
automated build process will build it.  We have the infrastructure in place
so it would be a shame for people to have to re-invent the wheel.

-----Original Message-----
From: Jeff McAffer [mailto:jeff@xxxxxxxxxxxxxxxxx] Sent: Tuesday, September 08, 2009 2:18 PM
To: ian.skerrett@xxxxxxxxxxx; Eclipse Packaging Project
Subject: Re: [epp-dev] A new EPP SOA package

I agree with the agility points 100%. By the same token, EPP is not the only place that "packages" can be put together. Any project, working group, ... is free to bundle up a mess of stuff and call it useful and make it available.

Its just my opinion.  EPP should not be the only conduit.

Jeff

Jeff McAffer | CTO | EclipseSource | +1 613 851 4644
jeff@xxxxxxxxxxxxxxxxx | http://eclipsesource.com




On 8-Sep-09, at 11:25 AM, Ian Skerrett wrote:

I think it is important that we have the SOA package available for SR1. A
couple of points:

- We can't get into a situation where the packages are tightly tied to the release train schedule. In the agile world we live in today, we need to have the ability to respond to new requirements. In this case a new group would like to create a new package, so I believe it is important that we can
respond to their request.

- As for quality, as with any package it is the responsibility of the
package maintainers to ensure the quality. Like all the other packages, Zsolt is responsible for testing and signing off on the quality. I agree it would be nice to have more time to community feedback but I believe that is
Zsolt decision to make.

- In terms of the release review, of course if Swordfish does not pass then they will not be able to release, so a package is moot. However, it is very
easy to remove something.  :-)

- I agree it does seem that SOA package is close to the Java Enterprise package but fortunately/unfortunately (depending on your perspective) the intention of the packages is to appeal to a certain user profiles. In this case the user profile is a SOA developer not a Java developer. The contents are similar but we want to attract developers that are interested in SOA.

In conclusion, I am a +1 on the SOA package. For me it is important EPP continues to be responsive and agile between release trains. We can't live
in a static environment and only change once a year.

Ian


-----Original Message-----
From: epp-dev-bounces@xxxxxxxxxxx [mailto:epp-dev- bounces@xxxxxxxxxxx] On
Behalf Of David M Williams
Sent: Monday, September 07, 2009 9:24 PM
To: Eclipse Packaging Project
Subject: Re: [epp-dev] A new EPP SOA package

Question: Is there anyone out there who is against (1)? Any objections
from other package maintainers or team members?

I'm a little against '1'. I don't want my voice to make the decision, but
I'll at least give a dissenting point of view and see if anyone else
agrees (or disagrees).

My largest concern is that there's no time for any use or feedback from the community or adopters. Part of the purpose of the simultaneous release (and packages, in my view) is not the end-result, but the build-up to that
end-result. Seems like the plan for option '1' focus on only the
end-result ... producing a package ... with out demonstrating the steady
predictability of milestones and release candidates.

A smaller concern is that I think 'Swordfish' is part of the package, and they are not having their Release Review until the very last minute (Sept
14th or so?)
So, there is some risk that it won't make it, and this won't be known
until the very last minute.

In fact, looking at the list of features, it seems like exactly the same thing as the EPP 'Java Package' (that is, the one that has mylin, and xml
in it) except for the addition of
- org.eclipse.wst.ws_ui.feature
- org.eclipse.swordfish.tooling.feature

So ... doesn't seem like anything that should be hard for users to install
or construct.

On the flip side, I see how the "working group" wants the publicity, etc.,
but I just don't like the packages being a "marking tool" without the
normal path of users and adopters trying it out for a few months first.
(Remember, normally, someone must be "in the build" for Simultaneous
Release by M4 or M5, which translates into nearly 6 months of community
use before release).

Would "Galileo SR2" be an option? That'd at least give some time to
publish some versions before it was officially released.

But, like I said, mine is just one view and I don't feel so strongly about it that I'd be insulted if you or others didn't agree. I definitely don't want to be 'negative' to new things, I just felt it important to voice the
counter point of view. I'm fine to go with your decision (as EPP lead)
either way, without further discussion or justification.

Good luck!



_______________________________________________
epp-dev mailing list
epp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/epp-dev

_______________________________________________
epp-dev mailing list
epp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/epp-dev

_______________________________________________
epp-dev mailing list
epp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/epp-dev



Back to the top