What percentage of criticisms are useful and valid and acted upon by the Pidgin team, and how many are too vague to be of use?

Speaking in VERY rough numbers, 75 percent of the complaints are really just people chiming in saying "me too" to the other 25 percent. Some of these remaining complaints, perhaps half, are impossible ones. Our work is horrid (no details), or the graphics need to be replaced (which the user could do on his own if he really cared about it), or similar. The rest are more thoughtful.

Still, by and large, even these more thoughtful ones quickly stop containing new information, new ideas or perspectives. Keep in mind we have a truly massive user base. We estimate about 3 million users. In any field those dissatisfied are always more likely to speak up than those satisfied or happy.

What is the most common complaint about 2.2.0?

2.2.0 did not really add to the list of things some users dislike. There was very little in it about which users could complain.

This is one of the more confusing things about our project. Looking around over the years, we very early on became disillusioned with the idea of having version numbers match some idea of how significant the release will be for users. For us, such an attempt is particularly hard because while a change to the msn code might hugely impact (one person), it has almost no impact on someone else.

For a long time we simply increased version numbers, thus 0.50 followed 0.49, even if there was only one letter changed between them (I did not look precisely what changed between these releases, I picked the numbers randomly to illustrate the point). As we reached the 0.80s though, we realized that people were starting to have expectations of 1.00, the 100th release, as if it would be a "traditional" 1.0 release for our project. We did not like that, it did not match where we were then, it still does not match where we are now.

In parallel to this, we had an ongoing issue. The number of third party plugins was steadily growing, and there was no rhyme or reason to help users, or developers, understand when a plugin would still work or when they would have to look for a new version of a 3rd party plugin to continue using it. The solution, it seemed to us, was to make the version numbers meaningful, but to developers and more particularly plugin authors. Thus the 2.2.0 release was 2.2.0 and not 2.1.2 because we added something to the API, and thus a plugin compiled for 2.2.0 wouldn't work with a 2.1.1 install.

So 2.2.0 itself was really a non-issue from a users' point of view. 2.1.1 caused much more noise, because of the change to the toolbar above the input windows in conversations.

But really, larger than this issue, has been a debate that has been long and protracted ever since the 2.0.0 release, that of protocol icons in the buddy list. This debate is not precisely new, in different forms it has come up again and again since 0.60. In that release, in April of 2003 and after nine months of work, we moved from GTK+ 1.2 to GTK+ 2.x. More pertinently, in that release we made the decision to start looking at things differently, to start to blur the lines between the various accounts you might have signed on in any one instance of Pidgin (Gaim at the time). This vision of how someone with many accounts looks at IM has only grown stronger since then. We feel that most people, most of the time, have multiple accounts because they have some need to, not because they really want to keep track of what friends know what accounts on which protocols.

Obviously some people disagree. Quite vocally. But users have consistently been very happy to be able to collapse a confusion of buddies into a smaller number of contacts each of which represent a single real life person they might try to reach.