New releases, priorities

I've just come back from DrupalCon in DC and participated there and previously in some CiviCRM consultant discussions. There are some useful notes here: Notes from Drupalcon DC Birds of a Feather Discussion, and here: http://civicrm.org/node/499.

One issue I'd like to highlight for Canadian organizations using CiviCRM is that of priorities in the development of the CiviCRM code base. In the past there has been considerable input from Canadian sources (Joe Murray, Green Party of Canada spring to mind), but there hasn't been much recently as far as I know. From the discussions noted above, as well as a conversation with Mike Gifford of OpenConcept, it seems there's a concern about too many new features being developed to the detriment of full development of the existing features. There's also some unhappiness about the dependency of CiviCRM 2.1+ on Drupal 6.

For myself - I'm not too concerned about the Drupal 6 dependency. It takes some time to adjust, but if you're doing Drupal development for any length of time, you have to come to grips with the fact that it's a moving target. My rough rule is that any medium-sized website with any degree of customization should probably set aside at least $3000/yr for upgrades. Larger and/or more complex/customized sites will need to do more. Once you've had to upgrade some old sites, you start really appreciating a discipline for following the 'drupal way' [minimal number of custom modules, only ones that have been around for a while and look like they're be upgraded, using a zen theme (or other standard) with customizations in css as much as possible, etc.]. I realise that small organizations often can't, or don't think they can, afford that much - in which case you really have to be upfront about the long-term costs of those little fiddly changes they think they want.

On the issue of priorities - I concur with the general feeling that there needs to be more priority put on development of the existing features than new ones. I also think that it's up to us, the user base [particularly the site developers amongst us] to provide more feedback in this department. CiviCRM has many fewer code developers so detailed bug reports are extremely important for the improvement of the existing features. But because the code is more complex, it's harder to debug in my experience, and also harder to even report bugs because many of them seem to be transitory and hard to nail down. In our discussions at DrupalCon, two good suggestions were:

1. More helpful error messages from CiviCRM [e.g. details, perhaps with backtraces, emailed to an administrator?]
2. Use of the forums for sharing issues that aren't quite definite enough to generate a bug report.

Any responses?