Now that’s just bad.

On a more positive note, at the recent OZ-IA conference I learned that Atlassian have hired an experienced User Experience designer to tackle the customisation and extensibility issues.

You are spot on with your description. Confluence is a nightmare to extend. The documentation is deplorable, even embarassing for a product that costs more than a few dollars, or even one that doesn’t. (Witness the usually effective docs at the jakarta commons.) In addition to this it is monstrously slow, as compared with other wikis. I have wandered through the source (in an effort to understand why the product does not function as advertised,) and indeed it is quite complex. In fact, I would say they bit off far more than they could chew.

However, my biggest beef with this “customizable” app is that is seems to be written for designers, not programmers; not at all. Performing even the simplest of things in a user macro is arduous. (Try manipulating the DOM; it is either impossible or so deeply hidden as to be unavailable.) One of my first goals was to write a reasonable editor for the macros since the provided text box is pre-alpha work. But without any docs this became impossible to do in any reasonable time. All but the most trivial of macros are simply too expensive to undertake. We have been able to build plug-ins to perform things like adding email notifications when documents change, adding interfaces to other databases and performing actions exposed by other apps (through the DB and http requests,) but like macros they are poorly documented with no real way to test them, and with error messages that are essentially meaningless. Of course our first deployed plug-in was to add a library to extend the paltry number of objects found in the velocity context, some of which simply do not exist as documented! (What ever possessed them to use velocity?)

To be sure, for writing and sharing pages it is very effective, and my organization has definitely benefited from that. But that would have been possible without a purchase. Of course, it has definite holes like editing a page whose macros cause exceptions, returning to the previous page after removing a parent or deleting a page and a treeview that remembers its state. (I built something like that 3 years ago with simple javascript!) The ironic bit is that we bought it because it seemed to offer the best customizability; I rue that day. Thankfully the original projects we had in mind, which seemed very cool at the time, have been abandoned in favor of easier things like integrating our webapp with remote manufacturing software. Well, that’s not really fair as we know our stuff pretty well. Ok, scratch that last point.

Sorry for the rant. The upshot is that this is pretty good as a wiki but it is not customizable any more than myspace.

Best of luck
-J

Best luck (to us all!),
-jd.

With regard to whether I’m happier now, well yes :) I haven’t had to do anything more with Confluence for some time. And my situation was not typical, I was trying to customise the look and feel of Confluence for a client who was using it for a purpose other than what it’s usually used. Sounds like you’re going to be approaching it from the systems admin point of view. So my troubles were atypical, I think. That said, I still think it would be good if Atlassian made the presentation engine easier to customise.

I’ve been “volunteered” to help out with our Confluence / Jira implementation. Can you point me to some good resources or books that could help in the maintenance and upgrading, etc.? Any help would be appreciated. How have you managed so far? Are you happier now?

Thanks for the help!

Eric

I asked why this was the case and was told that most people run Confluence on Windows and that it would be too expensive to write doco for all platforms. Hmmm

Confluence support groups are popping up around the place…which indicates a failure rather than a success for Atlassian.