Re: [Results] Priorities for 1.2
On Oct 18, 2006, at 8:37 AM, Alan D. Cabrera wrote:On Oct 18, 2006, at 12:24 AM, David Jencks wrote:On Oct 17, 2006, at 11:02 PM, Alan D. Cabrera wrote:On Oct 10, 2006, at 8:28 AM, Matt Hogstrom wrote:Thanks Alan ... I figured that was the plan but thought I should ask than assume :)Jencks is encouraging me to fix some of the TranQL problems in TCK so I'm updating my copy to 1.2. I simply can't wait to start running the TCK :-PGreat. I'm thinking that we can try to cut a real release by the end of this month and run it through the TCK wringer. What does everyone think?I was thinking more in terms of getting the tck to pass and then cutting a release. Otherwise we run the risk of having months of work on 2 codebases to get the tck to pass. Experience has shown me that we can't apply fixes to 2 codebases very well.Good point. Unfortunately the TCK licensing prevents us from providing an estimate on when this will happen.To me the main open question about 1.2 is whether we can certify on j2ee 1.4 with jee5 spec libraries. If so it is fairly simple to include jee5 preview features. If not we are going to have to jump through a lot of hoops to demonstrate any jee5 functionality. I haven't heard back from Geir with any news from sun on this question.Can you explain your comment about "we are going to have to jump through a lot of hoops to demonstrate any jee5 functionality"? It sounds scary.Yup, I'm worried. Lets take a simple example -- the tx manager. To do jpa according to spec, we need a jta 1.1 spec jar. It's possible to swap tx manager impls using the module/config aliasing, but I don't see a way to make this easy for users other than shipping 2 separate servers, one j2ee 1.4 w/no jee5 stuff and one uncertified jee5 server. Shipping twice as many servers would strain our resources as far as keeping track of so many zips.thanksdavid jencksRegards,Alan --Apache Geronimo - http://geronimo.apache.orgApache Yoko - http://incubator.apache.org/yokoLiveTribe - http://www.livetribe.org
Re: [Results] Priorities for 1.2
On Oct 18, 2006, at 2:02 AM, Alan D. Cabrera wrote:Great. I'm thinking that we can try to cut a real release by the end of this month and run it through the TCK wringer. What does everyone think?I think it would be excellent to get past TCK and branch trunk for the 1.2 release. I am hearing more and more folks interested in JEE 5.0 (the JUG last night was yet another confirmation). I'll volunteer to start the march to JEE 5.0. Based on the meeting with the JBoss guys it would be so great for us to beat them...if not beat them...be really, really close :-) Matt Hogstrom[EMAIL PROTECTED]
Re: [Results] Priorities for 1.2
On Oct 18, 2006, at 12:24 AM, David Jencks wrote:On Oct 17, 2006, at 11:02 PM, Alan D. Cabrera wrote:On Oct 10, 2006, at 8:28 AM, Matt Hogstrom wrote:Thanks Alan ... I figured that was the plan but thought I should ask than assume :)Jencks is encouraging me to fix some of the TranQL problems in TCK so I'm updating my copy to 1.2. I simply can't wait to start running the TCK :-PGreat. I'm thinking that we can try to cut a real release by the end of this month and run it through the TCK wringer. What does everyone think?I was thinking more in terms of getting the tck to pass and then cutting a release. Otherwise we run the risk of having months of work on 2 codebases to get the tck to pass. Experience has shown me that we can't apply fixes to 2 codebases very well.Good point. Unfortunately the TCK licensing prevents us from providing an estimate on when this will happen.To me the main open question about 1.2 is whether we can certify on j2ee 1.4 with jee5 spec libraries. If so it is fairly simple to include jee5 preview features. If not we are going to have to jump through a lot of hoops to demonstrate any jee5 functionality. I haven't heard back from Geir with any news from sun on this question.Can you explain your comment about "we are going to have to jump through a lot of hoops to demonstrate any jee5 functionality"? It sounds scary.Regards,Alan --Apache Geronimo - http://geronimo.apache.orgApache Yoko - http://incubator.apache.org/yokoLiveTribe - http://www.livetribe.org
Re: [Results] Priorities for 1.2
On Oct 17, 2006, at 11:02 PM, Alan D. Cabrera wrote:On Oct 10, 2006, at 8:28 AM, Matt Hogstrom wrote:Thanks Alan ... I figured that was the plan but thought I should ask than assume :)Jencks is encouraging me to fix some of the TranQL problems in TCK so I'm updating my copy to 1.2. I simply can't wait to start running the TCK :-PGreat. I'm thinking that we can try to cut a real release by the end of this month and run it through the TCK wringer. What does everyone think?I was thinking more in terms of getting the tck to pass and then cutting a release. Otherwise we run the risk of having months of work on 2 codebases to get the tck to pass. Experience has shown me that we can't apply fixes to 2 codebases very well.To me the main open question about 1.2 is whether we can certify on j2ee 1.4 with jee5 spec libraries. If so it is fairly simple to include jee5 preview features. If not we are going to have to jump through a lot of hoops to demonstrate any jee5 functionality. I haven't heard back from Geir with any news from sun on this question.thanksdavid jencksRegards,Alan --LiveTribe - http://www.livetribe.org
Re: [Results] Priorities for 1.2
On Oct 10, 2006, at 8:28 AM, Matt Hogstrom wrote:Thanks Alan ... I figured that was the plan but thought I should ask than assume :)Jencks is encouraging me to fix some of the TranQL problems in TCK so I'm updating my copy to 1.2. I simply can't wait to start running the TCK :-PGreat. I'm thinking that we can try to cut a real release by the end of this month and run it through the TCK wringer. What does everyone think?Regards,Alan --LiveTribe - http://www.livetribe.org
Re: [Results] Priorities for 1.2
On Oct 16, 2006, at 11:28 AM, Joe Bohn wrote: Dain, Could you clarify the "Full Java 5 support" item some more? I was under the impression when I first read this (esp. the statement about certification) that this was the work necessary to get to JEE5. However, others have indicated that this is really certification (TCK at the 1.4 spec level) but now on a 1.5 JVM. Certification on a Java 5 VM. -dain
Re: [Results] Priorities for 1.2
Dain, Could you clarify the "Full Java 5 support" item some more? I was under the impression when I first read this (esp. the statement about certification) that this was the work necessary to get to JEE5. However, others have indicated that this is really certification (TCK at the 1.4 spec level) but now on a 1.5 JVM. Thanks, Joe Dain Sundstrom wrote: The results are in! We had a good number of respondents, and the results are quite clear. Last night I spent a few hours tallying the results and analyzing them. There was a very clean ordering from the most popular to the least. Additionally, the features clearly fell into categories. Top --- These features were easily visually separated from the others. They tended to have a lot of #1 ranking and almost no low rankings. Additionally, these are the most clearly defined, have several people working on them, and are already mostly complete. This the feature set that will define the 1.2 release. Full Java 5 support (really is certification on Java 5) - Dain, Jencks and Rick OpenJPA integration - Blevins and Jencks Yoko ORB support - Rick and Jencks Global JNDI - Dain and Jencks Some support These features had a few very few 1-2 rankings, but many 3-5 rankings. Additionally, the also had a a few lower half rankings. Additionally, "Console usability improvements" and "More server modularization via plugins" are improvements and not features. I do not believe that anyone is working in "Console extensibility" or "OpenEJB 3.0 integration with @Stateless" with an eye for inclusion in 1.2 and these features are likely to be quite destabilizing. Therefor, they will not be included in the 1.2 release. Console usability improvements Console extensibility More server modularization via plugins OpenEJB 3.0 integration with @Stateless Little Support -- These features had a small amount of support noted buy a single vote for a top 3 ranking, and normally a supporting vote for a top 5 ranking (except CMP which only had a single #2 vote). Therefor, I think these features are best demonstrated via a plugin or branch to gather more support before they are included in any release. More out of the box samples Geronimo OSGi bundle GShell integration CMP improvements No Support -- These features had no rankings in the top half (1-7). JAF 1.1 Jetspeed integration -dain.
Re: [Results] Priorities for 1.2
Thanks Alan ... I figured that was the plan but thought I should ask than assume :) Jencks is encouraging me to fix some of the TranQL problems in TCK so I'm updating my copy to 1.2. I simply can't wait to start running the TCK :-P Matt Hogstrom [EMAIL PROTECTED]
Re: [Results] Priorities for 1.2
On Oct 10, 2006, at 6:33 AM, Matt Hogstrom wrote: Just as a point of reference on process. I'm just about done with the CMP improvements. These changes are really to tranQL and OpenEJB. I've had users asking for them and based on the note below I'm not sure what your expectations are. I plan on getting them tested and integrated this week. There are no changes to Geronimo framework but they are new features for the Assembly we call J2EE certified. Outstanding. I think the prioritization on the items below is great feedback but if someone is working on an area that is not in the mainstream of community priority but the code is being delivered what then? Not trying to be argumentative but I'm not understanding the tangible results of the discussion below apart from getting a community pulse on development. In the past people worked on what they were interested in. I thought the point of the exercise was to identify what the community wanted in the release before we cut it and to provide a roadmap for interested parties but not to exclude development in other areas. This is different than what we've done in the past. You bring up some great points. Your question indicates that we need to add some "color" to Dain's post. Hernan's roadmap has been out there for quite a while. As I've mentioned a couple of times with my discussion with Hernan, this is a great place to get the bigger picture and for people to see what features are in the pipeline and who's working on them. What Dain has posted is a tactical set of priorities for the v1.2 release. If someone wants to work on something for v1.2 that isn't on this list they are more than welcome to commit to adding it to that release. I hope that that list was inclusive of all the features that people have committed to completing for v1.2. If not, then we need to get them on this list. I think that the idea was to feature items that allow us to concretely plan. We cannot plan about features that have no resources assigned to them. Regards, Alan
Re: [Results] Priorities for 1.2
Just as a point of reference on process. I'm just about done with the CMP improvements. These changes are really to tranQL and OpenEJB. I've had users asking for them and based on the note below I'm not sure what your expectations are. I plan on getting them tested and integrated this week. There are no changes to Geronimo framework but they are new features for the Assembly we call J2EE certified. I think the prioritization on the items below is great feedback but if someone is working on an area that is not in the mainstream of community priority but the code is being delivered what then? Not trying to be argumentative but I'm not understanding the tangible results of the discussion below apart from getting a community pulse on development. In the past people worked on what they were interested in. I thought the point of the exercise was to identify what the community wanted in the release before we cut it and to provide a roadmap for interested parties but not to exclude development in other areas. This is different than what we've done in the past. Thanks ! Matt Hogstrom [EMAIL PROTECTED]
Re: [Results] Priorities for 1.2
For those that are interested, here is the raw data: http://cwiki.apache.org/GMOxDEV/priorities-12-results.html -dain On 10/9/06, Dain Sundstrom <[EMAIL PROTECTED]> wrote: The results are in! We had a good number of respondents, and the results are quite clear. Last night I spent a few hours tallying the results and analyzing them. There was a very clean ordering from the most popular to the least. Additionally, the features clearly fell into categories. Top --- These features were easily visually separated from the others. They tended to have a lot of #1 ranking and almost no low rankings. Additionally, these are the most clearly defined, have several people working on them, and are already mostly complete. This the feature set that will define the 1.2 release. Full Java 5 support (really is certification on Java 5) - Dain, Jencks and Rick OpenJPA integration - Blevins and Jencks Yoko ORB support - Rick and Jencks Global JNDI - Dain and Jencks Some support These features had a few very few 1-2 rankings, but many 3-5 rankings. Additionally, the also had a a few lower half rankings. Additionally, "Console usability improvements" and "More server modularization via plugins" are improvements and not features. I do not believe that anyone is working in "Console extensibility" or "OpenEJB 3.0 integration with @Stateless" with an eye for inclusion in 1.2 and these features are likely to be quite destabilizing. Therefor, they will not be included in the 1.2 release. Console usability improvements Console extensibility More server modularization via plugins OpenEJB 3.0 integration with @Stateless Little Support -- These features had a small amount of support noted buy a single vote for a top 3 ranking, and normally a supporting vote for a top 5 ranking (except CMP which only had a single #2 vote). Therefor, I think these features are best demonstrated via a plugin or branch to gather more support before they are included in any release. More out of the box samples Geronimo OSGi bundle GShell integration CMP improvements No Support -- These features had no rankings in the top half (1-7). JAF 1.1 Jetspeed integration -dain.
[Results] Priorities for 1.2
The results are in! We had a good number of respondents, and the results are quite clear. Last night I spent a few hours tallying the results and analyzing them. There was a very clean ordering from the most popular to the least. Additionally, the features clearly fell into categories. Top --- These features were easily visually separated from the others. They tended to have a lot of #1 ranking and almost no low rankings. Additionally, these are the most clearly defined, have several people working on them, and are already mostly complete. This the feature set that will define the 1.2 release. Full Java 5 support (really is certification on Java 5) - Dain, Jencks and Rick OpenJPA integration - Blevins and Jencks Yoko ORB support - Rick and Jencks Global JNDI - Dain and Jencks Some support These features had a few very few 1-2 rankings, but many 3-5 rankings. Additionally, the also had a a few lower half rankings. Additionally, "Console usability improvements" and "More server modularization via plugins" are improvements and not features. I do not believe that anyone is working in "Console extensibility" or "OpenEJB 3.0 integration with @Stateless" with an eye for inclusion in 1.2 and these features are likely to be quite destabilizing. Therefor, they will not be included in the 1.2 release. Console usability improvements Console extensibility More server modularization via plugins OpenEJB 3.0 integration with @Stateless Little Support -- These features had a small amount of support noted buy a single vote for a top 3 ranking, and normally a supporting vote for a top 5 ranking (except CMP which only had a single #2 vote). Therefor, I think these features are best demonstrated via a plugin or branch to gather more support before they are included in any release. More out of the box samples Geronimo OSGi bundle GShell integration CMP improvements No Support -- These features had no rankings in the top half (1-7). JAF 1.1 Jetspeed integration -dain.