On Mon, Mar 15, 2004 at 08:42:58PM +0100, Martin Pitt wrote:
> IMHO the templates can be regarded as an inofficial standard; tests
> should not be weaker than them; however, I don't think that a little
> freedom on the AM's side hurts. E.g. Martin Michlmayr said that he
> won't put my additional tasks (package correction, RC bug fixing,
> writing actual manpage for a program that does not already have one,
> and applicant-specific ones) into the official templates; however, he
> appreciates them.
I'd say a few of the template questions are harder than is necessary
to ensure that an NM candidate is competent. For example, the big long
question on linking:
W. Fundamental runtime linker knowledge:
W0. What are library sonames, and what are they used for? What is the
"ELF" format?
W1. How does the utility "fakeroot" work? How is that tied to
LD_PRELOAD, and the runtime linker?
W2. What is a symbol-versioned library? Why are libdb2, libdb3 and libc6
compiled using symbol versioning? What problems does symbol-versioning
solve?
W3. What is the -Bsymbolic ld flag, exactly what does it do, and how
that differs from library symbol versioning? What problems do
-Bsymbolic linking solve? Why is libc6 not compiled with
-Bsymbolic?
I'd wager that many DDs can't answer that without looking it
up. Especially, if they, like me, don't really maintain any C programs
(which isn't a requirement of course...).
> > I somewhat ashamed to say that I don't know if I could answer some
> > of the questions on the template without research.
> Me neither. Although I did e. g. the bad licences stuff a handful of
> times now, I always forget some clumsy details. Also I don't know by
> heart the nifty details of Emacs mode creation. But personally I'm not
> ashamed of that. I would be ashamed if I could not answer the
> questions/solve the excercises at all, but life is about "knowing
> where it is written", isn't it?
I agree that it is fine to ask questions that might require the
applicants to look something up; however, I'd prefer that the
knowledge gained from this is useful. The intricacies of the runtime
linker (or "fundamental runtime linker knowledge", as the question
writer calls it) are not particularly useful, IMHO.
As far as a "standard" difficulty goes, I really agree it's up to the
individual AM. The tougher the questions, the better the applicant
looks to the FD and DAM. If it's not broken, don't fix it, I guess.
--
Duncan Findlay
Attachment:
signature.asc
Description: Digital signature