On Thu, Feb 6, 2020 at 12:01 AM Kevin Kofler <kevin.kofler@chello.at> wrote:
So you propose to optimize Fedora for your own internal use at the expense
of thousands of users?

Optimizing for automated QA and CI will increase the quality for everyone (perhaps not immediately, but definitely in the long run, at least that's my belief), meaning for both your purported thousands of near-zero-bandwidth users and also for the one million other users. So yes, if we can improve the experience for everyone in one area, at the expense of a few percent worsened experience for a few percent of our user base in another area (who might not even exist, we simply don't know), that sounds like a pretty good deal to me.

No matter what you do, there will always be some people positively affected and some people negatively affected. The point is to strike the best balance, and also knowing your target audience. Unless that changed again, the last time I remember, we were targeting developers and creative people. They are going to be a minority in your already mentioned minority, because near-zero bandwidth will simply not allow them to do this job well. If you don't look for the best balance but instead speak in absolute statements ("size must never go up, whatever the benefits might be" -- not a direct quote, just how you sound), then you've simply picked your own target audience that gets 100% attention and don't care about the rest at all. Which of course you absolutely can do for your personal goals, but I don't believe Fedora-wide decisions should be done this way. It might also be tempting to simply not act on a change because you don't want to make life a bit harder for some group of people. But not acting is also acting, because by not implementing the change, you're making the future life harder for some other group of people, compared to what it could be. There is nothing like not acting, just forks on the road.