Re: [RFC] Git User's Survey 2010 (resend)

2010/7/3 Jakub Narebski [off-list ref]:

I guess it is time for annual (so far) Git User's Survey.  Should
there be one?  When should it start, and how long should it last?

Yes, I think there should definitely be one! IMO one month is enough.

With more that 3000 responses in 2008 and 2009 survey, analysing
free-form data was hard.  That is why in this proposal there is only
single one free-form response (about missing features).

That's a good idea.

== About you ==

NOTES:
^^^^^^
This section gives us a bit of demographical information about survey
responders.  Is it useful?  Should we leave it in survey, or remove it?

Should we for example include 'gender' as one of questions?  Perl Survey
2010 did.

I don't see the point of 'gender'. What does that tells us?

=== 02. How old are you (in years)? ===
(free-form single line)

NOTES:
^^^^^^
Instead of unconstrained free-form response it might be better to have
single choice (or menu) of age ranges.  What do you think?  Of course with
ranges there is question what ranges to use (how to quantize age); goo
solution would be to chose ranges corresponding somewhat to the levels of
education.

What's wrong with a free-form? I think that's easy and it works.

=== 10. What do you use to edit contents under version control with Git? ===
       What kind of editor, IDE or RAD you use working with Git?
(multiple choice, with other)

[...]

NOTES:
^^^^^^
Is this question useful, or should it be removed from survey?

I think this is useful to correlate communities.

=== 15. How do you publish/propagate your changes? ===
(multiple choice, with other)

[...]

NOTES:
^^^^^^
Should it stay, or should it be removed?  I guess it can be
interesting for git hosting sites...  Should we have separate answrs
for different kinds of push (ssh, "dumb" HTTP(S) with WebDAV, "smart"
HTTP - if it is possible, git:// protocol with push enabled)?

I think this question should stay. It would also help projects to
decide how to accept patches based on what most git users are familiar
with.

=== 16. Which of the following features do you use? ===
(multiple choice, with other)

[...]

NOTES:
^^^^^^
The problem is come up not with exhaustive list of features: there are
too many of them to list.  The problem is coming up with list of
important and used enough often features.

So: what features should be included in this list?  What features
should be removed from above list of answers?

I propose to add:
 + git stage/cache/index

We really are not sure how many people are actually aware of it, are we?

And IMO new features should go on the top.

=== 19. Overall, how happy are you with Git? ===
(single choice)

[...]

NOTES:
^^^^^^
I'm not sure if this question is at all useful.

I think it is. Otherwise how do we know that people are happy with it?

=== 20. In your opinion, which areas in Git need improvement? ===
       Please state your preference.
(matrix)

[...]

NOTES:
^^^^^^
Are there any general areas that are missing from this list?
What are they?

How about:
 + communication channels

I think if users have trouble reporting issues, asking questions, we
should catch that.

=== 22. How do you compare the current version with the version from one year ago? ===
(single choice)

[...]

NOTES:
^^^^^^
This question was mainly excuse for providing list of main changes
from the year ago.  I think that this question should be removed, as
it doesn't bring any important information.

Yeah, and I think many people don't even notice the changes as they
come, but learn slowly features that have been there since a long time
ago.

=== 28. How did you hear about this Git User's Survey? ===
(single choice, with other)

[...]

NOTES:
^^^^^^
This list would of course be updated to reflect the list of (planned)
announcement channels.

Should I try to post announcement on mailing list for projects that
use git?  There are entirely too many such projects nowadays, and such
announcement can be considered spamming by some...

I still maintain that we need an official blog (not planet). Last year
the most popular way of finding about the survey was through blog
posts, but you argued that it was because you didn't spam mailing
lists.

http://article.gmane.org/gmane.comp.version-control.git/124609

I still think it's unnecessary to spam mailing lists, but if it helps
us reach considerably more people, we should do it.

Hopefully after this year's result we will know for sure ;)

Thanks for doing this. Cheers.

-- 
Felipe Contreras