For those paying attention, the TJDO project is undergoing an administration change. Mike and the other developers who have worked on the project since it's been open-sourced have not worked at TriActive for a while now and they're mostly moving on to other things, although Mike is willing to supply some limited advice going forward. With this change, it's a good time to discuss where TJDO is now, and where it's going in the short term.
Where we are now:
1) we've got a relatively solid 1.0.1 JDO implementation that only does datastore identitity. It has a few significant bugs, does not support all the DBs that I would like it to (hsqldb and Derby) and is missing support for a number of significant optional features (lists, etc.). We also have a minor issue with compatability between the Sun reference enhancer and Java 1.5
2) one truly active developer (me), with some good advisors in the wings
3) very limited support from TriActive; I get to work on TJDO issues that directly effect TriActive business on TriActive company time, otherwise it's my volunteer time...
4) some good competing open source JDO implementations, either working or in progress: JPOX, Apache OJB
a lot of good commercial JDO implementations
a lot of other good persistance technologies available or apparently coming (e.g. Hibernate and EJB3)...
In other words, the arena is very very different now that it was when TJDO initially came out...
While my employer (TriActive) has choices if TJDO no longer meets its needs, so far that's not the case, and the thought of migrating 100GB of customer data to a new JDO implementation is a LOT of incentive to make the current implementation work better. Not that migration is impossible or even that hard, but that it's a lot mind-numbing effort that could be better applied to benefiting everybody...
Given the realities, I'm going to have to pick and choose where I spend my time. Here's where I'm thinking of taking the project over the next year or so, in no particular order:
1) recruit some more help, either developers or contributers
2) fix the bugs, get a solution for the enhancer issue (in progress now), keep the project up with current technologies...
3) implement the top few major missing optional features (lists, etc), integrate the remaining submissions, add support for more RDBMSs
4) performance improvements, such as speed and reduced memory footprint on large queries. A number of these will be specifically targeted to Oracle deployments (e.g index-organized tables for certain situations), since that's what TriActive uses and what they'll pay me to do...
5) Improve the documentation and tutorial
6) improve logging, auditing/monitoring and general usability
7) complete the 1.0.1 TCK
8) release some TJDO-based utility libraries and demos (possibly in another project or three)
Some things that I know I'm NOT going to have time for in the short term: application identity, and JDO 2.0 compliance (unless TriActive decides they need 2.0 features, which is unlikely at the moment).
Now, I'm open to additional suggestions or discussions, especially from those willing to contribute. What do YOU think should be the focus and priorties in the near term?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'd rather stop working with TJDO, there are several items to solve and few developers who want to help. Getting much money is not enough, we need more help.
By the way I really recommend "3) implement the top few major missing optional features (lists, etc), integrate the remaining submissions, add support for more RDBMSs", because without it the TJDO wil die by itself.
Ive realized this discussion would have more attention.
Regards.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
From my (customer's) point of view, the best way is to integrate your efforts with JPOX team. JPOX is very similiar to TJDO (it is based on this).
JPOX has only two very good developers who really need your help.By joining JPOX, you have many missing features from start and you can influence on quality of the product.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
For those paying attention, the TJDO project is undergoing an administration change. Mike and the other developers who have worked on the project since it's been open-sourced have not worked at TriActive for a while now and they're mostly moving on to other things, although Mike is willing to supply some limited advice going forward. With this change, it's a good time to discuss where TJDO is now, and where it's going in the short term.
Where we are now:
1) we've got a relatively solid 1.0.1 JDO implementation that only does datastore identitity. It has a few significant bugs, does not support all the DBs that I would like it to (hsqldb and Derby) and is missing support for a number of significant optional features (lists, etc.). We also have a minor issue with compatability between the Sun reference enhancer and Java 1.5
2) one truly active developer (me), with some good advisors in the wings
3) very limited support from TriActive; I get to work on TJDO issues that directly effect TriActive business on TriActive company time, otherwise it's my volunteer time...
4) some good competing open source JDO implementations, either working or in progress: JPOX, Apache OJB
a lot of good commercial JDO implementations
a lot of other good persistance technologies available or apparently coming (e.g. Hibernate and EJB3)...
In other words, the arena is very very different now that it was when TJDO initially came out...
While my employer (TriActive) has choices if TJDO no longer meets its needs, so far that's not the case, and the thought of migrating 100GB of customer data to a new JDO implementation is a LOT of incentive to make the current implementation work better. Not that migration is impossible or even that hard, but that it's a lot mind-numbing effort that could be better applied to benefiting everybody...
Given the realities, I'm going to have to pick and choose where I spend my time. Here's where I'm thinking of taking the project over the next year or so, in no particular order:
1) recruit some more help, either developers or contributers
2) fix the bugs, get a solution for the enhancer issue (in progress now), keep the project up with current technologies...
3) implement the top few major missing optional features (lists, etc), integrate the remaining submissions, add support for more RDBMSs
4) performance improvements, such as speed and reduced memory footprint on large queries. A number of these will be specifically targeted to Oracle deployments (e.g index-organized tables for certain situations), since that's what TriActive uses and what they'll pay me to do...
5) Improve the documentation and tutorial
6) improve logging, auditing/monitoring and general usability
7) complete the 1.0.1 TCK
8) release some TJDO-based utility libraries and demos (possibly in another project or three)
Some things that I know I'm NOT going to have time for in the short term: application identity, and JDO 2.0 compliance (unless TriActive decides they need 2.0 features, which is unlikely at the moment).
Now, I'm open to additional suggestions or discussions, especially from those willing to contribute. What do YOU think should be the focus and priorties in the near term?
Well my friend, we're here again.
I'd rather stop working with TJDO, there are several items to solve and few developers who want to help. Getting much money is not enough, we need more help.
By the way I really recommend "3) implement the top few major missing optional features (lists, etc), integrate the remaining submissions, add support for more RDBMSs", because without it the TJDO wil die by itself.
Ive realized this discussion would have more attention.
Regards.
From my (customer's) point of view, the best way is to integrate your efforts with JPOX team. JPOX is very similiar to TJDO (it is based on this).
JPOX has only two very good developers who really need your help.By joining JPOX, you have many missing features from start and you can influence on quality of the product.