August's DUG meeting wrap-up

[copied from my WorkHabit blog post]

We had a great time at the S.D. Drupal User Group meetup this month. there were almost 20 people in attendance this time, including Myself, Michael Murdoch, and Harry Slaughter, the latter of whom you would expect to be there because he was the impetus and organizer! I even saw some of the Achieve Internet gang there, including Alex Villeux and Karl Scheirer, my former supervisors. Awesome! A few of the gang from Achieve's Costa Rican development office were up for the week, and I got to hang out with them for a little bit too. <href="http://drupal.org/user/208711">Nestor Mata is a Central American codehead who I had the pleasure of working with on the http://namm.org project (he built the Rotor module, among others) and it was great to finally meet him face to face. Cheers, Nestor!

We all pulled some tables together at the Cream coffee house here in San Diego, and despite the background XM Radio and the polite yet curious stares of the other patrons, we proceeded to discuss old business, new meetup locations, and All Things Drupal (That's ATD, for short. Yes, we know it's similar to ADD. I wonder why?).

I had mentioned in an earlier comment on the meetup thread that I might be willing to discuss a rather uncommon but not unique problem that we'd run into on a client's site not too long ago dealing with the publishing and unpublishing of a single content type for administrators who didn't have the sitewide 'administer nodes' permission. I had not received any responses or replies to that post, so I didn't think there was any interest in it. After all, it's like I said. This is a rather rare case of the Drupal publishing world, and it's user case stems from a lack of trust and a need for control in such a way that certain users just could not be granted the administer nodes perms.

Boy was I wrong about the interest of this topic!

After we'd gotten our old business out of the way and addressed the need to find a location for the next meetup that didn't have constant music playing and espresso machines lamenting their own efforts in the background, Harry leaned over the tables and asked me if I would be willing to talk about Domenic's WorkHabit client request which I'd posted that g.d.o. forum comment about. So I did.

The basic premise of Domenic's dilemma was that the client's user case was established as content maintainers that needed to be allowed to publish or unpublish a certain content type without being allowed to delete or edit that content type. If those maintainers are granted the Administer Nodes perms, they can publish or unpublish certain nodes, but then they could also edit the content of those nodes, or even delete them. The Override Node Options module looked like a great fix for this use case until we realized that users can't even see unpublished nodes unless they have that dreaded Administer Nodes permission.

Our DUG talked through the approaches to the problem from a Drupal standpoint, why it even existed in the first place, whether it was core's fault, a contrib module's fault, or just an oversight that nobody had actually run into before. We talked about different ways to solve it, and even whether that solution should be a patch to core, a patch to an existing contrib module, or a standalone contrib.

All in all, we'd like to see this functionality become part of an existing contrib module rather than being yet another module out there in hyperspace. WorkHabit just doesn't know what to patch the new code against. Ultimately, I think the DUG decided that this really was a Drupal core deficiency but that it couldn't be solved within the current system of roles and permissions without adding even more clutter to the stack. It's really a non-mainstream workflow to begin with, and if core can't easily be adapted to alleviate this pitfall, then core needs to become more generically flexible in it's permissions structures. Luckily there are some fabulous minds at work on this very idea for D7, and thus, we all decided that the fix-it code can stay as a contrib module called View Unpublished and just hope that people can find it when they need it.

The DUG discussion then devolved into a focus on change management, why's it' so hard to keep different environments in sync, content identification vs. configuration information, and the recent developer's mailing list discussion about universally unique identifiers.

What we discovered during these fragmented discussions (fragmented due to high ambient noise and an extremely long table) is that change management is still a hot issue for the community at large, but it's progressing toward solution(s) at an encouraging pace. It was a great meetup, we didn't solve the UUID problem, but we did have a few beers, a lotta coffees, and some marvelously entertaining and informative mind-swapping.

I'd call that a good day. See you all next month!