On Friday, January 25, 2019 3:17:55 PM CST Michal Kubecek wrote:
...and now you are building even more questionable conclusions on that unfounded implication.
Just as founded as the benchmarks being absurd.
How about this? Either I want a firewall with connection tracking on that particular box or I don't. In the former case, I'll have to configure and enable it on the distributions which do not provide it out of the box; in the latter, I'll have to disable it on those which do. The important point is that if I'm interested in performance of my box, I'm interested either in performance with the firewall or without. The performance I'm interested in is the performance of the system I'm going to actually use. Whether it's with a firewall or without is my choice independent of what distribution I'm going to use and its defaults.
That is your use-case, that is not the the use-case of the target audience for the benchmarks. They want a system with a network connection that works, perhaps a certain way, out of the box without having to know anything more about it.
Just imagine it. You go to Phoronix, pick a distribution which ranks best in their "benchmarks", install it the "next-next-next" way, do not even choose what packages you want to install, which services you want to run and start using it with the default config? What if you want CIFS and the "fastest" distribution doesn't install Samba by default? Do you just shrug and say "Pity... but what can I do?" Maybe there really are users thinking like this - but I strictly disagree with the idea of tailoring our design decision to them.
Adding things not part of default install is just the same, I want to install samba and configure auth and storage location. I do not want to futs with various settings and related network configuration to get decent performance using it. Obviously, others may want/need to do those things.
If I remember correctly, few years ago there was some discussion about what should be openSUSE's driving idea and target audience. The result was that we want to be the best distribution for software development. Does that sound as if our target audience were people who leave the decision if they want a firewall or not to distribution defaults? I don't think so.
Sure, but then that is not the target audience of the benchmarks necessarily. Although the assumption that a software developer (an extremely broad group) knows how literally everything on the system works or cares to is a bit much. This says nothing of the usefulness of the benchmarks. If I am not the target audience of openSUSE and the benchmarks keep me away then they worked as desired. It is a choice the openSUSE community could make to completely ignore them, but to call them absurd is quite another. -- Jimmy -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org