Re: why is it taking so long?!?! (was: title not centered)

[Phil Holmes]

For the record it's http://code.google.com/p/lilypond/issues/detail?id=1116 
raised June 13th.

--
Phil Holmes
Bug Squad


"Graham Percival" <address@hidden> wrote in message 
news:address@hidden
On Fri, Jun 11, 2010 at 11:45 PM, James Bailey
<address@hidden> wrote:
>

Ok, Bug Squad.  Help me out here.  I'm on the verge of starting to
tear my hair out, which would be a shame because my hair is absolutely
gorgeous.

I'd like to introduce a charming abbreviation often used by one of the
engineering professors here: JFDI.  Just "Frequently" Do It.  (except
that's not grammatical, because the word isn't actually "frequently")


It's been something like 36 hours since this email.  Your target is to
handle issues within 24 hours.  What happened?

The latest policy, which I'll be uploading once 2.13.24 has finished
building (but is already online on kainhofer), has 6 steps on the
checklist.  There's 5 reasons to reject a report, then #6 says "add it
to the tracker".

So as soon as you get this email, start at the beginning.
Can I reject it because 1. it's a question for -user?   nope.
Can I reject it because 2. it has no Tiny example?  nope.
Can I reject it because 3. it's unclear?  nope.  User expected to see
a centered title.  User does not see a centered title with the latest
version.  Move on to the next step.
Can I reject it because 4. that behaviour is expected?  doesn't look
expected to me.
Can I reject it because 5. the issue already exists in the tracker?  nope.

ok, what's left?  6. Accept the report as described in _adding issues
to the tracker_.  So JDFI.  Just BLOODY WELL ADD IT TO THE TRACKER.

You are not programmers.  You are not documentation editors.  ok,
maybe some of you do those tasks as well, but in terms of being
members of the Bug Squad, YOU ARE NOT EXPECTED TO SOLVE ISSUES.  All
you're supposed to do is check for 5 ways to reject a report; if you
cannot find one of those excuses to reject it, YOU BLOODY ADD IT TO
THE BLOODY TRACKER.

Don't stop and investigate.  Don't think.  Do not pass go.  Go
directly to the tracker and jail the issue.  (for non-English
speakers: those are references to the popular game "monopoly")

There are only 5 ways to avoid adding an issue to the tracker.  It
doesn't take concentration. I've done it.  I know what I'm talking
about.  You can check these while you're listening to a lecture.  You
can do this while watching tv or subtitled anime.  You can do it while
having sex with your girlfriend.  ok, I admit that I haven't tried
that last one; anybody living in Scotland or northern England who's
interested in doing some empirical evaluations of quality assurance
procedures in open-source projects can contact me off-list.  Oh, I'll
also be in Bordeaux in a few weeks, so French girls also have a chance
at being "collaborators" in this "research project".

... I lost my train of thought.  What was I ranting about before I
started thinking about those impeccably-dressed French girls?


> Hmm, after looking at it a little longer, it looks like the functions have
> changed.

ah, there we go.

Do you see anything in Changes that could explain this?  I didn't
after 30 seconds.  So that's no excuse not to add it to the tracker.

> \fill-line works here instead of \center-column, so:
> \version "2.12.3"
> \header { title = \markup \fill-line { "How Razorback Jumping Frogs Level
> Six Piqued Gymnasts" } }
> \score { {s1} \layout {} }

Interesting; you can add this note to the issue.

> yields expected output, while
>
> \header { title = \markup \center-column { "How Razorback Jumping Frogs
> Level Six Piqued Gymnasts" } }
> \score { {s1} \layout {} }
>
> yields unexpected output.

Yes.  So JDFI.

> I think this is possibly more a documentation issue.

If that change is deliberate -- which it is NOT your job to
investigate -- then ok.  But we still need an issue for the doc issue.
Oh, and we can EASILY change the Type of an issue after creating one.
Don't stress about what Type of issue it should be.  JFDI.  If you
have absolutely no clue, I suggest Type-defect, Priority-medium for
anything.


This is not a casual rant.  Because of the problems with the Bug
Squad, I've spent a lot of time and energy over the past few weeks
trying to sort stuff out.  As a result, the new website isn't online
yet.

I want to be very clear about this.  I'm aware of approximately 2
hours of technical work that I need to do before the new website can
go online.  Double that for unexpected delays, and double that for
other problems I find later.  Double that again just for fun.  New
website estimate: 16 hours.  Over the past 3-4 months, I've spent more
time than that on Issues.  Taking the most optimistic estimate -- 2
hours, nothing unexpected -- we could have had the new website up
MONTHS ago.


What's worse is that I'm not even going to TRY finishing the website
until I'm confident that the Bug Squad is functioning.  That means:
1.  caught up on the backlog
2.  a new Bug Meister who seems to know what he's doing
3.  at least one week without disasters

Overseeing the Bug Squad, in addition to the new doc editors, and all
the others "daily maintenance" tasks required to keep a project of
this size going, takes up all my available lilypond time.  If I try to
work on new stuff of my own -- the website, GLISS, etc -- then I'll go
way over the amount of time I can afford to spend on lilypond.  (or
rather: I'll be even further above the amount of time I can afford to
spend on lilypond)


I really, really want you guys to get the message: you're trying to do
too much.  If you cannot reject a report, add it to the tracker.
There are only 5 reasons to reject a report.  An estimation of why the
problem occurs is NOT one of those reasons.  JDFI.  If we find a bunch
of bad reports being added, I might add another possible reason to
reject a report, of course.  But for now, it's a 5-item checklist.


Seriously.  If it takes you longer than 5 minutes between reading an
email and either writing a rejection email or adding to the tracker,
then you're doing something wrong.  Don't do something wrong.  You're
holding up the new website.

- Graham