[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Various things


>> I think it's mainly people who know, or can search and then produce
>> the answers, but feel that they are wasting their time on that.

> If you do not know the answer it will of course take so much time to
> learn first.

Well, in my case I spent more time writing down the messages that actually
researching and learning.  And that was the point of my previous message: when
a test becomes a test on how well you do a write-up, on how nicely you can
rephrase something that is already online, it becomes a tedius essay instead of
a good test.

>> Completing the 4 templates takes many many hours, hours that could be
>> spent bugfixing, packaging, writing code, etc.

> Yes, ANY checks will take time from other things.

Well, yes of course, but if instead of explaining how the BTS works you
actually go and fix some bugs, you are both fulfilling the test and fixing bugs
at the same time.

>> In my case, my P&P questions came in the middle of a BSP, I promptly
>> stopped fixing bugs to answer these questions.  In the end, I stole
>> something like 12 hours of the BSP to answer P&P questions...  Isn't
>> that like a big big waste?

> No it is not. If you had known the answers you had answerd this mail in
> less than say 30 minutes. As you did not (or was unsure enough to check)
> you had to learn. Do not take this personally. Maybe you knew but wanted
> to double-check. It is fully acceptable but will of course take some time.

Know the answers?  Let's have a look at a couple of questions.

In the first P&P template:

5. At http://people.debian.org/~joerg/bad.licenses.tar.bz2 you can
    find a tarball of bad licenses. Please compare the graphviz and three
    other (your choice) licenses with the first nine points of the DFSG
    and show what changes would be needed to make them DFSG-free.
    There's no need to compare word for word (which would be impossible
    for some licenses anyway), but you should spot the biggest mistakes.

This forces the person in NM to write a VERY long essay on the differences of
the licences, how they relate to the DFSG, their bugs, etc.  This is no easy
writing, you can't "know the answer", since you have to _compose_ whatever you
are going to write. I mean, even if you are a frequent poster of debian-legal,
you'd still need to take some time to read the licences that you are going to
speak about and then write your opinions about them.

In the second template, you first have:

 2. Please tell me 3 different methods to close a bug in the BTS, the
    difference between them, and when to use which method.

Which is fine, even if it's a bit tedious.  And then, near the end:

24. Again, something BTS related: Consider you have a package, with a
    set of open bugs. Some of them fixed by the new upstream version
    (which you are about to upload), some of them aren't really bugs, or
    are not relevant anymore (because they were fixed ages ago, for
    example). How would you handle them?

Which is almost repeating the answer in 2, but rephrasing to fit 
this question.  Which is really bothering.

Also, if you count both P&P1 and P&P2 (I'm not counting the T&S templates), you
have around 40 questions.  There's no way you could answer them in half an
hour.  All questions take time, no matter how perfectly you know the answers.
It takes time to phrase the answer, it takes time to check for typos, it takes
time to recheck the whole mail once it's finished.

Finally, just as a sample, of how much my p&p mails weight:

$ cat pyp1 pyp2 | grep -v '>' | wc
 683    5184   30033

This is to say, while writing my P&P messages, I wrote 30kb of information.
Would you agree that it's sort of difficult to produce this much _useful_ 
information in half an hour?

> I can agree on this that it is more interesting to "do do actual things"
> but it is very important that all Debian developers know the basic
> policy and procedures. What would happen if we accept people after
> making some checks that the person can patch packages or write manual
> pages without checking if they know anything about Debian policy or
> what the procedures are?

No one is saying that we should start accepting anybody that wants to enter
into Debian.  What we are saying is that actually _doing_ some of the stuff
that is asked about in the NM templates might be better, both for the Applicant
and for Debian as a whole.

 Bezos,     (o.
     Marga. (/)_

Attachment: signature.asc
Description: Digital signature

Reply to: