critera proposal/discussion: boot.fedoraproject.org

[Chris Murphy]

On Jan 11, 2013, at 5:06 AM, Kamil Paral <kparal at redhat.com> wrote:

> I tried BFO a few times and I've found it heavily outdated every time (including today) so I wasn't particularly thrilled about it.

That's a problem.

> Do you think this is something we should block our release on (i.e. creating a release criteria for it)?

What percent of hardware does ipxe work/fail on? Even if it works on a majority, a significant minority failure could make testing for a release blocking standard difficult. And at the moment it's BIOS only, not for UEFI.

> Also, I'm not fully sure who should benefit from it. If you need to install Fedora, you do it once in a long time, and there's no problem to download ISO and boot it. If you install Fedora often (e.g. you're QA), booting from the Internet every time is a no-go, not until it's as fast as your USB flash drive.

Fedora LiveCD is ~700MB, and soon after release it's ~400MB of updates. BFO gets you that one time install downloaded and installed, up to date, faster than ISO.

The thing is, LiveCD or DVD's inefficiencies still meet a wide range of requirements. BFO is a much narrower use case, that doesn't meet many needs. Therefore it requires qualification to explain who should use it. And finding out about it is yet another issue, it's totally non-obvious to find BFO.

Chris Murphy