On Thu, Sep 18, 2025 at 04:05:58PM -0400, Jeffrey Walton wrote: > On Thu, Sep 18, 2025 at 3:14 PM Bill Allombert <[email protected]> wrote: > > > > On Thu, Sep 18, 2025 at 02:50:56PM -0400, Jeffrey Walton wrote: > > > Popularity Contest is used to collect information on packages > > > installed by users. However, there is currently a need to determine > > > whether a particular architecture should be supported, but > > > popularity-contest does not collect the information. See [1] and [2]. > > > > popularity-contest purpose is to collect packages usage, not hardware usage. > > However some hardware usage can be inferred from software usage, by looking > > at > > hardware specific packages, like kernels and formware packages. > > > > > I think it would be useful to have popularity-contest collect > > > information on the underlying hardware to guide arch decisions. > > > > See bug #238687. > > Thank you sir!
Independently of popularity-contest, the difficulty is -- to define what informations are collected and how. -- to define how the informations are exploited. Then we can see how to proceed. However a popularity-contest-like survey will not be useful when targetting very small set of users. For powerpcpse I can offer this graph: https://popcon.debian.org/stat/sub-powerpcspe.png which show zero submissions since approximately 2020. Cheers, -- Bill. <[email protected]> Imagine a large red swirl here.
