On Mon, 27 Jan 2020 15:26:23 +0100
Stefan Hundhammer <shundhammer(a)suse.de> wrote:
On 27.01.20 13:42, Josef Reidinger wrote:
I am all for doing
that, but it's a vital piece of infrastructure, so we (!)
need to do this right. Outsourcing this to a GSoC student would be most
Sadly, even if it is vital, we do not assign any resources to libyui in past. Now
situation is better, still I am not sure if we can finish it in reasonable future. But I
agree that this is a bit too much for GSoC and probably result will be only some prototype
how it can work, which can be start of discussion and maybe even not used in final
When we do something that is really important for us, we should get into the
habit of doing it the proper way, not doing something half-baked or, worse,
letting some outsider do something half-baked. We need to make use of our
expert knowledge in the area; otherwise somebody from the outside will repeat
mistakes that we already made in the past. If we do it, we have better
chances of keeping those past mistakes in the back of our head and avoiding
them in a new implementation.
Well, goal of GSOC is to learn that things to our student. So ideally we should catch
early enough that student doing something wrong what does not work in past and explain it
I think we do not necessary do coding, we just need to be sure that it does not repeat
mistakes we do in past and drive that development.
This process starts with collecting requirements, thinking hard (which
requires a free head, i.e. not being interrupted by countless meetings and
unrelated things), coming up with a solid concept, maybe doing a feasibility
prototype, evaluating it, starting a skeleton, fleshing it out.
I am not seeing anything that cannot be done together with student. Last time we have
daily meeting where we discuss stuff with students,
we write ideas down ( which we often do not do ourself ), it is not one man show. So
ideally in GSOC we should just do first part ( requirements needs to be known in advance,
so it can be evaluated ) and probably we wont reach much more then that skeleton depending
on available time and student.
Yes, this can be done in an agile way (if that is what we want).
Agile is useful with student as it is regularly evaluated if it goes properly and in right
direction. So student do not waste month on something that won't work. BTW I think we
do agile ourself for same reason, as often right way is not clear.
Agile does not have to mean doing minimalistic stuff that falls apart when
the first unexpected thing happens. ;-)
It is goal of mentor to ensure that this stuff should not happen. In general GSOC should
be both side win situation. Student learn something new from us, get paid and learn how
open source works ( last few years google promote also social part of contributing ). We
on other hand see newcomers issues, discuss yast with someone with fresh mind, having new
contribution and potentially also new contributor.
To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org
To contact the owner, e-mail: yast-devel+owner(a)opensuse.org