Published
Thursday, July 9, 2015 - 16:37
Written by

tldr

  • July: Open development. Major initiatives may still be accepted for 4.7 if significant implementation work has been done.
  • August: Feature freeze.  Existing 4.7 initiatives should be mostly done, but there may still be some work on development/maturation/gaps/loose-ends. Bug fixes encouraged. New (minor) enhancements may be accepted, but this will grow stricter as the month progresses.
  • September: Stabilization. 2 or 3 alphas. Mostly bug fixes. Split off 4.7 from master.

Discussion

CiviCRM is community open-source, heavily influenced by the participation of partners and users. Twice each year, we gather for week long sprints to plan and develop. The April sprint in Colorado produced a number of great initiatives, and a major factor was the alignment with the release: 4.6.0 was out; the roadmap for 4.7 was as clear as the Rocky Mountain sky; and we could focus on new initiatives without loose-ends to distract us. It was fun. To create the same dynamic for the October sprints, we'll need to wrap up the current initiatives beforehand.

Looking through the on-going intiatives being developed by active contributors and the core team, there's still a lot of work to be done for 4.7. Here's an incomplete list of the larger initiatives:

  • Payment processing refactor
  • ACLs for financial types
  • Status messages / reports
  • Dynamic settings UI; settings cleanup
  • Payment and Order APIs
  • JS upgrades
  • Webform / IPN
  • Scheduled reminder - hookable models
  • Simplified API connections
  • Security audit
  • Migration of User/Admin docs
  • Website redesign

Most of these initiatives have made substantial progress since April, but we need to finish them.

Cuts and Edits

A couple initiatives were conceived for the 4.7 cycle but don’t appear feasible. A few words about them:

  • Installer Rewrite: This will get pushed to the next cycle. Hopefully, we can renew this one during the fall sprints.
  • Form Builder POC: Coleman and I had hoped to spend 150-200 hours on a broad proof-of-concept form-builder (encompassing multiple components), but it won’t fit in this cycle. So we’re attacking this problem from another angle - the dynamic settings UI has a lot of overlap with the form-builder (using APIs and JSON metadata to construct the UI), and we’ll spend an extra 50-80 hours making components from the dynamic-settings UI more robust. In the next cycle, we can build on them to make a broad POC.

(Update) Joe Murray pointed out over IRC that the list of on-going initiatives was missing a few cool ones, so I've added them. There appears to be some technical difficulty with the spam filtering on the blog's comments.  If there are other issues/comments, please ping totten on IRC or email (@civicrm.org).

(Update) Comments should be working again.

Comments

If it's finished the ability to customise relative search parameters will be a really appealing feature

I discussed this with Jon Goldberg and he is happy to work on it but won't have it ready in time for 4.7. Thanks Jon for keeping it going!