Contents
Summary
To maintain some sort of statistics of packages installed by Mageia users, to justify their inclusion in the repositories and the need to maintain them.
Owner
- Name: Lewis Smith
- Email: lewyssmith@onetel.com
Resources
List here people who will be implied in this feature (packagers, QA, doc, ...)
Current status
- Targeted release: Mageia 6 or 7
- Last updated: 2026/03/25
- Percentage of completion: <XX>%
Detailed Description
Mageia offers perhaps 10k packages, each of which requires (I think) a packager/maintainer, and is subject to periodic updates which can be difficult to test. We often do not know whether anyone out there is interested, uses the package. We might be toiling for nothing in the case of unused packages. We have no idea.
I propose some sort of centralised, unique, anonymous register which attempts to record the number of instances of packages actually [un]installed by users. This can never be accurate, but better some idea than none at all. The register would have to recognise all packages in installation ISOs as such, for it is impossible to count them; the number of ISO downloads is no clue; and counting those is a separate issue. Conversely, it is impossible to cater for simple removal of installed Mageia systems with all their packages; the result of which would be falsely +ve counts for any packages explicitly installed post-installation.
I see the thing being done at the Install/Remove Software, urpmi/urpme level, reporting anonymously packages involved. It should not even be necessary to make it opt-outable, since it is anodine & anonymous and in users' interests.
There was hot debate on this subject on the QA mailList. Plenty of support for the principle, but strong opposition on the basis that Mageia would be "spying on" its users like M$ or Google. This is simply not true with an anonymous system which simply 'counts'. The complaint was that any automatic user -> Mageia contact was spying; but this overlooks that we (like all distributions) already do this with the automatic software update process - which is not contested as 'spying'.
Why it would be good for Mageia to include it
Most importantly, Mageia might be able to identify a lot of packages which never seem to get installed by anyone - and ditch them. This would rid the packaging and maintenance workloads of 'dead wood'. We could eliminate some work for nothing.
OTOH It must be easy for users to request packages they want but we do not offer; I think there is already a mechanism for this.
Test case
Software / Packages Dependencies
Install/Remove Softwrae, urpmi/urpme
