ThinkingRock Forum

Based on the discussion on the release date for the next version, and many people telling they are developers and/or engineers, it might be useful to start a dedicated discussion on the development process of TR. From my point of view (I'm a Software Engineer myself btw) there are many ideas floating around that might help TR development, especially because they are planning to go Open Source.

There has been lots of discussion on the release date being pushed back several times. Several people mentioned that there are many valid reasons for this. However, I think there are a few options to make the development process more transparent for users and also to work towards a release date that is a lot less subject to change.

Claire has stated that they currently manage outstanding issues in a bugtracking system (I think they use Mantis). Opening this system up for users, if only for review, might silence a lot of questions about the current development state.

The other problem with developing software is that new ideas come up during the development process, new bugs are being detected during testing, etc. Developing software is indeed a volatile process. From the developer's standpoint a program is never finished and releasing your results to the public is always coupled with the notion that users will find things that you didn't see. Because of this developers tend to push back release dates because they want to implement 'that last important feature' or fix 'that last annoying bug'. But let's face it, there will always be more features and bugs.

I would suggest using a system like Trac (http://trac.edgewall.org/) to manage the project. The important functionality in Trac is the option to set release milestones and bind issues to them. This helps to move certain issues to future releases and set a closed target (features and bug-repairs) for the next upcoming release. Also, users can help by replicating reported bugs on their system and add that information to a issue, something that is incredibly helpful for a developer.

So that's it for now. If you have other suggestions please post them here as well. This thread might function as a list of ideas that Claire and Jeremy can use to streamline their process and even more when they are going Open Source with TR

Hans Peter.

Thanks for your thoughts. Just want to add more background.

In Mantis, we have created projects per release (which corresponds to your milestones idea). You are right in saying that we always add to the initial list thinking 'this will so nice', 'easy to do' (sometimes not), 'annoying bug', etc.

Just so everyone knows the scale of our issues database, this is a summary:
- in 2.0, the total number of issues is 136: 62 change requests and 74 bugs (most functionality had to be re-implemented and re-tested due to the platform change);
- some issues (there are 48 open ones) have been moved to 2.0.1.
- then we have 35 issues for the 2.1 release and as Future Functionalities, there are 108...

And there are many topics on the forum which have not been entered into Mantis yet.

So you will understand that I am not very keen to transfer everything on a different tracking tool (and Mantis is pretty good).

_________________
Claire
ThinkingRock Analyst and Tester

Claire,

I completely understand that moving to something else is certainly not something to look forward to. I still suggest you take a good look at Trac as it has serious advantages over Mantis (I like Mantis as well btw). Better to make the move now (or soon) if this is useful then wait for a few other releases and find Mantis doesn't cope anymore and you have to move several times the amount of issues you have now.

There is of course a way to manage such a transaction a bit. Your Mantis is currently not public. So you could setup Trac (after the 2.0 release) and start managing outstanding issues for the release after 2.0 and further. You don't move issues that are currently in Mantis and set for the next release after 2.0, just enter new surfacing issues into Trac. Any issues in Mantis that will move to the subsequent release AFTER the next release after 2.0 can be moved to Trac over time.

Still, it's just a suggestion

I do not know mantis other than by name but I wanted to mention that there are conversion tools:
http://trac.edgewall.org/wiki/TracImport#Mantis


There is also the possibility to use an open project infrastructure hosting service (like sourceforge etc.)

CANONICAL's https://launchpad.net/ (also used for UBUNTU projects) has a neat feature called https://translations.launchpad.net/ where anybody with an launchpad account can help with i18n and l10n of apps using gettext (can java use this?).

Big +: launchpad.net doesn't want to sell you something (yet), no ads.

Are we talking about GtD or GtW (Getting things worse)?

are there really enough people out there in the TR universe who are really interested in such information. Maybe it's time for another poll. Currently I do not see any need for the big majority of TR users, but I may stand corrected.

Just my 02 cents

Volker

Volker, did you see the topic of this discussion, or the section where this discussion is placed?

In this section (Development Process) some people (like me) that have a lot (or a little) development experience are talking about ideas and information that Jeremy and Claire can help and use to develop the program further. This discussion is one of the first steps of going 'Open Source' where other people chip in with help, development ideas and knowledge to support the original developers.

So yes, there are several people here on the forum that are interested in such information. And that includes Jeremy and Claire as they are the ones that manage this forum and hence created this discussion area