I agree that it is very easy to go wrong with usage data collection. I’ve seen it happen several times. The key is to start with questions rather than what data should be collected. In this case, we have clear questions and concrete actions that could be taken based on the answers, so I don’t see lack of interested to be an issue. Lack of resources to implement such a system is a different matter entirely…
From: cross-project-issues-dev-bounces@xxxxxxxxxxx [mailto:cross-project-issues-dev-bounces@xxxxxxxxxxx] On Behalf Of Pascal Rapicault
Sent: Monday, July 15, 2013 9:58 AM
To: Cross project issues
Subject: Re: [cross-project-issues-dev] FW: [Doug on the Eclipse CDT] New comment on "Eclipse smells kind of dead"???.
I agree that this could be useful but we have to remember that the UDC has been retired because supposedly everybody cared about such data but in reality nobody ever did… So who is going to sign up to maintain this, and analyze the data.
On 07/15/2013 06:44 PM, Konstantin Komissarchik wrote:
Regarding default preferences, the “best” thing to do is to capture what preferences the users are tweaking using an automated usage collection system. Then one can produce a report showing, for instance, how many users prefer line numbers in the margin. If the stats really do favor line numbers, then the default setting should be revisited regardless of history. If the stats show the opposite, then the reports can be published, which should partly help to pacify those who incorrectly assume that their preference is representative of the majority.
+1, that's a good idea.
In any case, such reports should be made public. It shows users that they have some control on the deliveries. And their content will be very nice input to the most famous trolls of code style (tab vs margins, 80 columns...).
Should I open a bug to suggest this "Preference Collector" mechanism?