GitSurvey2009

From Git SCM Wiki
Revision as of 09:59, 14 January 2010 by KorgWikiSysop (Talk | contribs)

Jump to: navigation, search

Contents

Git User's Survey 2009 summary

The Git User's Survey 2009 has been closed on September 15, 2009.

The survey could be found here:

This survey was open from 15 July to 15 September 2009 (for 2 months).

You can get raw data (individual responses) for the survey here:

Results of the survey:

There were 3868 individual responses in this survey (the 30 responses in 'test' channel were removed).<
> There were 3236 individual responses in GitSurvey2008 (including 21 responses in 'test' channel).<
> There were 683 individual responses in GitSurvey2007.<
> There were around 117 responses in GitSurvey2006.


<b>Table of contents: <<(3)>>


Information about survey completion

Number of questions answered per responder

align="right"

<tablestyle="font-size: smaller;" rowstyle="font-weight: bold; background-color: #ffffcc;"> Answered Resp.[n] Resp.[%]
30 29 0.7%
29 129 3.3%
28 247 6.4%
27 439 11.3%
26 606 15.7%
25 681 17.6%
24 504 13.0%
23 418 10.8%
22 210 5.4%
21 116 3.0%
20 84 2.2%
19 53 1.4%
18 31 0.8%
17 30 0.8%
16 22 0.6%
15 32 0.8%
14 26 0.7%
13 13 0.3%
12 6 0.2%
11 5 0.1%
10 3 0.1%
9 2 0.1%
8 3 0.1%
7 0 0.0%
6 0 0.0%
5 0 0.0%
4 4 0.1%
3 2 0.1%
2 4 0.1%
1 7 0.2%
0 162 4.2%

The table above shows histogram of the number of answered questions by responder from the 30 questions in the survey. Most people answered from 23 to 27 questions from 30. Average number of skipped (not answered) questions is 6.7 +/- 5.9.

Note that in one of questions, "22. Did you participate in previous Git User's Surveys?" lack of answer didn't necessarily mean that somebody skipped this question; lack of answer was there answer in itself. There are also optional (well, more optional) 3 free-form essay questions. That makes 26 questions to answer.

Most answered question was "07. I use Git for (check all that apply)" multiple-choice question.

Number of replies per question

GitSurvey2009-resp hist.png

The graph above shows histogram of number of replies per question. As you can see the least replied questions were:

  • "12. What tool (or kind of tool) would you like to have Git support in?" (free-form essay question),
  • "19. What features would you like implemented in Git? What features are you missing?" (also free-form essay),
  • "22. Did you participate in previous Git User's Surveys?" (where skipping means simply not participating in previous surveys), and
  • "30. What other comments or suggestions do you have that are not covered by the questions above?" (again free-form essay), and also
  • "28. Which communication channel(s) do you use?" (multiple choice question, without 'other' answer)

Date of response

GitSurvey2009-date hist.png

The graph above shows histogram of number of responders by date (in CET / GMT+1 / +0100 timezone).

There are a few spikes spikes on above Completion Rate Graph; from those the following four spikes are most visible:

  • 15 Jul | 573 respondents / day
  • 21 Jul | 622 respondents / day
  • 27 Jul | 155 respondents / day
  • 13 Aug | 103 respondents / day

(note that those results are slightly different from the ones presented in [summary, part 1' thread] mentioned above; this might be possibly the matter of different timezones used).

This survey was announced at the following blogs (note that dates below for blog posts are in local timezone of blog, whatever it is, while Survs.com dates are in survey admin timezone, which is GMT+01:00 / +0100):

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Local Date Date URL
July 14th, 2009, 06:34 pm 14 Jul http://gitster.livejournal.com/36139.html
July 15th, 2009 at 4:23 am 15 Jul http://blogs.gnome.org/newren/2009/07/15/git-users-survey-2009
July 15, 2009 - 12:47 pm 15 Jul http://gitlog.wordpress.com/2009/07/15/git-users-survey-2009
July 17, 2009 at 10:40 am 17 Jul http://blog.gitorious.org/2009/07/17/the-git-users-survey
20 Jul 2009 20 Jul http://github.com/blog/460-git-user-s-survey-2009
20 Aug 2009, 10:08 am 20 Aug http://www.ceondo.com/ecte/2009/08/git-survey-2009 (http://www.indefero.net/blog/)
Sep 10, 2009 @ 10:11 AM 10 Sep http://blog.assembla.com/assemblablog/tabid/12618/bid/10431/Bloggers-compare-Git-notes.aspx

"Git User's Survey 2009" was also announced on the following blogs (found via blogsearch):

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Local Date Date URL
17th Jul 2009, 8:47 am 17 Jul http://marilyn.frields.org:8080/~paul/wordpress/?p=2632
2009-07-27 27 Jul http://blog.hartwork.org/?p=451
4 Aug 2009, 10:13 04 Aug http://virtualgenius.spaces.live.com/blog/cns!CC3C966FA733667C!811.entry
2009/08/18 18 Aug http://www.debian-news.net/2009/08/18/misc-developer-news-17/

Another quite important channel for finding about this survey was, according to question 19, microblogging platforms such as [[2]], [[3]], [[4]], etc.

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Local Date Date URL
5:41 PM, Jul 15th 2009 15 Jul http://twitter.com/jhelwig/statuses/2652743883
11:56 PM, Jul 19th 2009 19 Jul http://twitter.com/kevmoo/status/2728027845
12:24 AM, Jul 21st 2009 21 Jul http://twitter.com/github/status/2746745923
6:02 PM, Jul 21st 2009 21 Jul http://twitter.com/howdous/status/2760055856
8:06 PM, Aug 23rd 2009 23 Aug http://twitter.com/scyldinga/status/3495621307
<-3>
24-Jul-09 00:27:46 UTC 24 Jul http://identi.ca/notice/6906889
24-Jul-09 10:47:50 UTC 24 Jul http://identi.ca/notice/6927573
03-Aug-09 05:36:49 UTC 03 Aug http://identi.ca/notice/7461438

As you can see the announcement on !GitHub Blog (and via Twitter: [[5]]) correlate with the second spike in responses on 27 July 2009, when we take difference in timezone into account.

It looks like the announcement on Hartwork Blog, part of Planet Gentoo, might be responsible for small spike in responses at 27 July. But it might have been cased by announcement arriving via other channels, see answers to question "29. How did you hear about this Git User's Survey?" below.

Survey statistics

Total respondents: 3868
First response: Jul 15, 2009
Last response: Sep 16, 2009
Open during: 72 days
Average time: 49 minutes

At most often time between responses was below 1 second (the precision of Survs.com response timestamp), while maximum time between responses was slightly more than 12 hours (half a day).

There were 25 responses in the non-default "GitSurvey2009: no cookies, anonymous" channel that required no cookies, and was totally anonymous (no logging of IP address). This was at the cost of not being able to go back to one's responses to correct or complete it.

About you

01. What country do you live in?

(tabularized free-form single line)

This table is sorted alphabetically.

<tablestyle="font-size: smaller;" rowstyle="font-weight: bold; background-color: #ffffcc;"> Country Count Perc.
America (possibly USA) 4 0.1%
Argentina 21 0.6%
Armenia 1 0.0%
Australia 121 3.3%
Austria 49 1.3%
Bangladesh 1 0.0%
Belarus 3 0.1%
Belgium 39 1.1%
Brazil 58 1.6%
Bulgaria 6 0.2%
Canada 158 4.4%
Chile 8 0.2%
China 37 1.0%
Colombia 6 0.2%
Costa Rica 4 0.1%
Croatia 7 0.2%
Cuba 2 0.1%
Czech Republic 47 1.3%
Denmark 30 0.8%
Egypt 4 0.1%
El Salvador 1 0.0%
Estonia 9 0.2%
Faroe Islands 1 0.0%
Finland 52 1.4%
France 157 4.3%
Germany 400 11.0%
Gibraltar 1 0.0%
Greece 11 0.3%
Hong Kong 2 0.1%
Hungary 11 0.3%
Iceland 2 0.1%
India 51 1.4%
Indonesia 4 0.1%
Iran 2 0.1%
Ireland 18 0.5%
Israel 22 0.6%
Italy 65 1.8%
Japan 27 0.7%
Kazakhstan 1 0.0%
Korea, Republic of 4 0.1%
Korea (which one?) 1 0.0%
Kyrgyzstan 1 0.0%
Latvia 6 0.2%
Lithuania 10 0.3%
Luxembourg 2 0.1%
Macedonia (former Yugoslav Republic) 3 0.1%
Malaysia 7 0.2%
Malta 1 0.0%
Mexico 13 0.4%
Moldova 4 0.1%
Nepal 3 0.1%
Netherlands 76 2.1%
New Zealand 39 1.1%
Nicaragua 2 0.1%
Norway 63 1.7%
Pakistan 1 0.0%
Panama 1 0.0%
Peru 1 0.0%
Philippines 3 0.1%
Poland 68 1.9%
Portugal 11 0.3%
Romania 22 0.6%
Russian Federation 75 2.1%
Serbia 5 0.1%
Singapore 3 0.1%
Slovakia 4 0.1%
Slovenia 8 0.2%
South Africa 14 0.4%
Spain 59 1.6%
Sri Lanka 2 0.1%
Sweden 73 2.0%
Switzerland 56 1.5%
Taiwan, Province of China 11 0.3%
Thailand 4 0.1%
Trinidad and Tobago 1 0.0%
Tunisia 1 0.0%
Turkey 12 0.3%
Ukraine 22 0.6%
United Kingdom 293 8.1%
United States 1189 32.7%
Uruguay 4 0.1%
Vietnam 3 0.1%
<rowstyle="border-top: dotted;"> Invalid answer 7 0.2%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3631 / 3868

As one can easily see, slightly less than third of Git users (32.7%) are in the USA (those who answered this survey, and this question). Next of countries is Germany with around 11.0% responses. Third is United Kingdom of Great Britain (including England, Scotland, Wales and Northern Ireland) with 8.1%.

The names of the countries were normalized using [[6]] Perl module version 2.07 (with the fix that Serbia and Montenegro split into Republic of Serbia and Montenegro on 5 June 2006).

Top 12 Countries:

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Country Count Perc.
United States 1189 32.7%
Germany 400 11.0%
United Kingdom 293 8.1%
Canada 158 4.4%
France 157 4.3%
Australia 121 3.3%
Netherlands 76 2.1%
Russian Federation 75 2.1%
Sweden 73 2.0%
Poland 68 1.9%
Italy 65 1.8%
Norway 63 1.7%

You can also take a look at [Activity Map] at [[7]], where you can see on (large) Google Map locations of selected people who have Git in a [[8]], or locations of Git contributors (those with known location).

It is a bit pity that this survey did not include question "Does git.git repository include code produced by you?" (or equivalent), so we would be able to see where the developers are.

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Continent Count Perc.
Africa 19 0.5%
Asia 201 5.5%
Europe 1769 48.7%
North America 1371 37.8%
South America 98 2.7%
Oceania 161 4.4%
<rowstyle="border-top: dotted;"> Unknown 12 0.3%

In the table above "Oceania" includes (among others) Australia and New Zealand. As you can see most responders are from Europe (49%, slightly less than half), followed by North America (38% responders).

The conversion from country name to continent was done using [[9]] Perl module from CPAN, which in turn uses data from http://www.worldatlas.com/cntycont.htm

02. How old are you (in years)?

(tabularized free-form single line)

align="right"

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Years Count Perc.
< 18 59 1.6%
18-21 339 9.4%
22-25 862 23.8%
26-30 1169 32.3%
31-40 962 26.6%
41-50 168 4.6%
51-75 45 1.2%
76+ 6 0.2%
<rowstyle="border-top: dotted;"> Not a number 5 0.1%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3615 / 3868

Discounting joke responses (or errors in parsing), youngest user who answered this question is 12 years old, oldest is 90 (next to oldest is 68). That is quite a span.

Most Git users (around one third, 32.3%) fall in the 26-30 range. The age of 27 got most responses (in 2007 it was 25, as in 2008).

Getting started with Git

03. Have you found Git easy to learn?

(single choice)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
Very easy 155 4.3%
Easy 750 20.6%
Reasonably easy 1984 54.6%
Hard 675 18.6%
Very hard 71 2.0%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3635 / 3868

Nice Gaussian curve. Most users find it reasonably easy to learn, slightly biased towards easy.

04. Have you found Git easy to use?

(single choice)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
Very easy 356 9.8%
Easy 1348 36.9%
Reasonably easy 1606 44.0%
Hard 301 8.2%
Very hard 40 1.1%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3651 / 3868

Most users find it reasonably easy to easy to use Git.

When analyzing this data (and the data for the question before) you should take into account that somebody considering or finding Git too hard to learn wouldn't be probably Git user, and thus wouldn't fill this survey. So because it is *Git User's* Survey we should consider that results can be skewed towards lower value (easier).

What's interesting is comparing (percentage) results for questions 3. and 4.; how hard is git to learn versus how hard is to use. It seems like Git is reasonably easy to learn, and reasonably easy to easy to use. So it looks like Git just have somewhat steep learning curve, and the difficulty to learn pays in being more powerful to use.

05. Which Git version(s) are you using?

(multiple choice, with other)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Version Count Perc.
pre 1.3 9 0.2%
1.3.x 14 0.4%
1.4.x 46 1.3%
1.5.x 698 19.0%
1.6.x 3271 89.2%
<-3>
minor (maintenance) release 1.x.y.z 978 26.7%
'master' branch of official git repository 220 6.0%
'next' branch of official git repository 43 1.2%
other, please specify 68 1.9%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3666 / 3868

~- Description:<
> One can find git version by using "`git --version`" or "`git version`".

"Minor release" is additional specification, so if one for example use git version 1.6.3.3, one should check both "1.6.x" and "minor release". -~

Analysis:<
> As you can see from above results great majority (83%) run latest 1.6.x series of git (note: because it is multiple-choice question it means 'at least on one machine'). Only a few use 1.3.x or pre 1.3 versions somewhere.

Only a bit more than one fourth (26.7%) of responders use (somewhere) maintenance releases. That, or the question was,'t formulated clear enough.

06. Rate your own proficiency with Git:

(single choice)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Proficiency Count Perc.
1. novice 156 4.2%
2. casual, needs advice 675 18.3%
3. everyday use 1418 38.5%
4. can offer advice 1204 32.7%
5. know it very well 228 6.2%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3681 / 3868

~- Description:<
> One can think of it as 1-5 numerical grade of your proficiency in Git. -~

Analysis:<
> As you can see most responders know Git enough for everyday use, or can even offer some Git advice. If we treat possible answers as a proficiency grade, the average proficiency is around 3.2.

Either Git users don't stay novices long, or survey announcement(s) didn't reach novice users.

This data could be useful later to be able to check how it differs git usage for novices and for gurus (what commands they run, what features they use at different levels of proficiency).

How you use Git

07. I use Git for (check all that apply):

(multiple choice, with other)

Note that above choices are neither orthogonal nor exclusive. One might want to check multiple answers even for a single repository.

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
work projects 2908 78.9%
unpaid projects 2887 78.4%
<-3>
proprietary projects 1332 36.2%
OSS development 2422 65.7%
private (unpublished) code 2834 76.9%
<-3>
code (programming) 3173 86.1%
personal data 1156 31.4%
documents 1145 31.1%
static website 1037 28.1%
web app 1527 41.4%
sharing data or sync 777 21.1%
backup 893 24.2%
backend for wiki, blog, or other web app 318 8.6%
managing configuration files 1202 32.6%
frontend to other SCM (e.g. git-svn) 1145 31.1%
<-3>
other (please specify) 97 2.6%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3684 / 3868

~- Description:<
> Note that above choices are neither orthogonal nor exclusive. You might want to check multiple answers even for a single repository. -~

08. How do/did you obtain Git (install and/or upgrade)?

(multiple choice, with other)

Note that this question is multiple choices question because one can install Git in different ways on different machines or on different operating systems.

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
binary package 2624 71.5%
source package or script 853 23.3%
source tarball 741 20.2%
pull from (main) repository 632 17.2%
preinstalled / sysadmin job 236 6.4%
<-3>
other - please specify 134 3.7%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3668 / 3868

~- Description:<
> Explanation: "binary package" covers pre-compiled binary (e.g. from rpm or deb binary packages); "source package" covers things like deb-src and SRPMS / *.src.rpm; "source script" is meant to cover installation in source-based distributions, like 'emerge' in Gentoo.

Automatic update (apt, yum, etc.) in most cases means binary package install; unless one uses source-based distribution like Gentoo, CRUX, or !SourceMage, where automatic update means using source package (or source script).

The option named "preinstalled / sysadmin job" means that either you didn't need to install git because it was preinstalled (and you didn't upgrade); or that you have to ask system administrator to have git installed or upgraded.

Note that this question is multiple choices question because one can install Git in different ways on different machines or on different operating systems. -~

Analysis:<
> Most people (71.5%) use ready binary packages, which was kind of expected; that is the easiest way.

09. On which operating system(s) do you use Git?

(multiple choice, with other)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Operating System Count Perc.
Linux 3223 87.5%
MacOS X (Darwin) 1547 42.0%
MS Windows/msysGit (MINGW) 826 22.4%
MS Windows/Cygwin 367 10.0%
FreeBSD, OpenBSD, NetBSD, etc. 271 7.4%
 !OpenSolaris 120 3.3%
other Unix 77 2.1%
<rowstyle="border-top: dotted;"> MS Windows (any) 1193 32.4%
<-3>
Other, please specify 45 1.2%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3682 / 3868

~- Description:<
> On Unices you can get the name of operating system by running "`uname`". -~

Analysis:<
> Most common used operating system is Linux, next is MacOS X, and then MS Windows (on MS Window dominates msysGit version). This is quite understandable, as Git was created on Linux and for Linux, and it works best there.

The "other" answer include Solaris, AIX, HP-UX (other Unix), GNU/Linux (Linux), Debian, NixOS (Linux distributions), none (?); iPod, OpenVMS, QNX, Hurd and "my toaster" (?). And some commentaries and clarifications.

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)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
simple text editor 942 25.6%
programmers editor 3221 87.6%
IDE or RAD 1208 32.8%
WYSIWYG tool 205 5.6%
<-3>
other kind 108 2.9%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3678 / 3868

~- Description:<
>

  • "simple text editor" option includes editors such as pico, nano, joe, Notepad,
  • "programmets editor" option includes editors such as Emacs/XEmacs, Vim, !TextMate, SciTE (syntax highlighting, autoindentation, integration with other programmers tools, etc.)
  • "IDE (Integrated Development Environment) and RAD (Rapid Application Development)" option includes tools such as Eclipse, !NetBeans IDE, IntelliJ IDE, MS Visual Studio, KDevelop, Anjuta, Xcode, Code::Blocks but also tools such as Quanta+, !BlueFish or Screem (for editing HTML, CSS, PHP etc.), and Kile or LEd for !LaTeX.
  • "WYSIWYG tools" option includes word processors such as MS Office or !OpenOffice.org, but also tools such as Adobe Acrobat (for PDF) or GIMP (for images), or WYSIWYG DTP tools such as !QuarkXPress,  !PageMaker or Scribus, or WYSIWYG HTML editors such as !FrontPage, Dreamweaver or !KompoZer.

-~

Analysis:<
> Most popular kind of tool is programmer's editor with very strong 88% lead. Next in turn are IDE or RAD with 33%, followed by simple text editor, with 26%. The WYSIWYG tools are last, with very small 6% of replies; also people who use WYSIWYG tools also use other kind of tools to edit (presumably to edit commit message at least).

From browsing through some of "other tool" responses it seem like there should be included 'web interface' among possible tools (written as: wiki frontend, !GitHub editor, <textarea>). Other responses include Xournal (note taking application for graphic tablet, using stylus), Tinderbox (also note taking application), Git GUI tool: GitX (?), custom git-aware CMS (probably also 'web interface'), scripts and oneliners, command line and 'On occasion, Unix text manipulation tools'. Also there is "audio editors" answer (which probably should fall under "WYSIWYG tool)...

Many answers in 'other tool' falls in one of ready categories; perhaps they are clarification of answer?

11. What Git interfaces, implementations, frontends and tools do you use?

(multiple choice, with other)

<tablestyle="font-size: smaller;" rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
git (core) commandline 3557 97.3%
JGit (Java implementation) 164 4.5%
library / language binding 134 3.7%
<-3>
Cogito (DEPRECATED) 13 0.4%
Easy Git 28 0.8%
Pyrite 7 0.2%
 !StGIT 61 1.7%
Guilt 15 0.4%
 !TopGit 38 1.0%
pg aka Patchy Git (DEPRECATED) 2 0.1%
gitk 1673 45.8%
git gui 742 20.3%
QGit 231 6.3%
 !GitView 22 0.6%
Giggle 156 4.3%
 !GitNub 152 4.2%
 !GitX 685 18.7%
git-cola 58 1.6%
tig 209 5.7%
 !TortoiseGit 201 5.5%
Git Extensions 79 2.2%
git-cheetah 7 0.2%
git-instaweb 101 2.8%
git-sh 54 1.5%
Gitosis (as admin) 398 10.9%
repo (to manage multiple repositories) 65 1.8%
<-3>
editor/IDE VC integration 417 11.4%
filemanager / shell extension (any) 98 2.7%
graphical history viewer/browser (any) 575 15.7%
graphical commit tool (any) 223 6.1%
graphical diff tool 594 16.3%
graphical merge tool 541 14.8%
graphical blame or pickaxe tool 84 2.3%
my own scripts (for daily use) 318 8.7%
my own scripts (for special tasks) 375 10.3%
<-3>
Other (please specify) 284 7.8%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3655 / 3868

~- Description:<
> Here "graphics diff tool" means tools such as Kompare, and graphical merge tool means tools such as Meld and !KDiff3. Those answers include graphical merge and diff tools used by programmers editors and IDEs.

"graphical history browser (any)" covers tools such as gitk, QGit, Giggle, tig etc., but also built-in git commands such as "`git log --graph`" and "`git show-branch`". If you use one of mentioned tools as history browser, mark both a tool and "graphical history browser (any)"; if you use some graphical history viewer not listed here, please both mark this answer and specify it in the "other tool" answer.

Similarly for other answers marked "(any)". -~

Analysis:<
> Currently the above table is only roughly divided into categories and subcategories.

12. What tool (or kind of tool) would you like to have Git support in?

For example IDE, RAD, editors, continuous integration, software hosting, bugtracker, merge tool...; this includes language bindings and Git (re)implementations

(free-form essay)

1308 / 3868 non-empty responses

13. Which git hosting site(s) do you use for your project(s)?

Please check only hosting sites where you publish/push to

(multiple choice, with other)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Hosting site Count Perc.
repo.or.cz 236 6.9%
 !GitHub 2110 61.9%
Gitorious 345 10.1%
Savannah 46 1.4%
 !SourceForge 138 4.1%
Assembla 44 1.3%
Unfuddle 222 6.5%
kernel.org 58 1.7%
freedesktop.org 56 1.6%
Alioth 78 2.3%
Fedora Hosted 52 1.5%
<-3>
git hosting site for related projects 102 3.0%
generic site without git support 127 3.7%
self hosted 1929 56.6%
<-3>
Other (please specify) 187 5.5%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3407 / 3868

Note that replies might be affected by the fact that announcement about this survey was shown at some git hosting sites, but not at other (additionally users of some git hosting sites may skip welcome page more often than for other hosting site; not all hosting sites have blog or a news / announcements section).

Note also that there probably are other git hosting sites that were not present in the ready list of answers, but are hiding in the "other (please specify)", like for example Codebase.

The leader here is (again) !GitHub (62%); only self hosted has close amount of replies (57%).

14. How do you fetch/get changes from upstream repositories?

(multiple choice, with other)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
git protocol 2818 78.5%
ssh 2562 71.4%
http 1164 32.4%
rsync (DEPRECATED) 19 0.5%
filesystem 868 24.2%
via git-bundle 66 1.8%
foreign SCM (e.g. git-svn) 799 22.3%
<-3>
Other, please specify 48 1.3%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3588 / 3868

~- Description:<
> This question asks about how do you get changes (updates) from projects you follow into your local repository. It is not about how do you get latest version of Git.

Fetching (or rather cloning) via bundle could mean that project publishes ready for download bundles to reduce traffic and load on server (HTTP download [of bundle] can be resumed, git-clone currently cannot; one can also distribute bundle using P2P). -~

15. How do you publish/propagate your changes?

(multiple choice, with other)

<rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer Count Perc.
push 3285 92.7%
pull request (+ any form of announcement / notification) 1003 28.3%
format-patch + email 789 22.3%
format-patch + other (e.g. reviewboard, issue tracker or forum) 342 9.7%
git bundle 78 2.2%
git-svn (to Subversion repository) 879 24.8%
git-p4 (to Perforce repository) 34 1.0%
foreign SCM interface (other than mentioned above) 71 2.0%
<-3>
other - please specify 94 2.7%
<rowstyle="font-weight: bold; font-style: italic; background-color: #ccffff;"> Base 3543 / 3868

~- Description:<
> Publishing via bundle could mean sending bundle via email, or posting it on review board (or forum). -~

16. How often do you use the following forms of git commands or extra git tools?

(matrix)

<tablestyle="font-size: smaller;" rowstyle="font-weight: bold; background-color: #ffffcc;"> Answer never rarely sometimes often Avg. / Count
git add -i / -p 1259 37.4% 586 17.4% 669 19.9% 849 25.2% 2.3 / 3363
git add -u / -A 1744 51.9% 565 16.8% 453 13.5% 588 17.5% 2.0 / 3350
git am 2128 63.3% 522 15.5% 402 12.0% 273 8.1% 1.6 / 3325
git am -i 2567 76.3% 382 11.4% 188 5.6% 162 4.8% 1.4 / 3299
git apply 1908 56.7% 717 21.3% 441 13.1% 215 6.4% 1.7 / 3281
git apply --whitespace=fix 2503 74.4% 397 11.8% 201 6.0% 146 4.3% 1.4 / 3247
git archive 2246 66.8% 460 13.7% 346 10.3% 170 5.1% 1.5 / 3222
git bisect 1704 50.7% 763 22.7% 541 16.1% 191 5.7% 1.8 / 3199
git bisect run <cmd> 2352 69.9% 474 14.1% 249 7.4% 105 3.1% 1.4 / 3180
git annotate 2174 64.6% 509 15.1% 318 9.5% 155 4.6% 1.5 / 3156
git gui blame 2477 73.7% 340 10.1% 220 6.5% 99 2.9% 1.3 / 3136
git blame 1473 43.8% 724 21.5% 682 20.3% 240 7.1% 1.9 / 3119
git blame -L <start>,<end> etc. 2539 75.5% 342 10.2% 139 4.1% 83 2.5% 1.3 / 3103
git bundle 2680 79.7% 248 7.4% 100 3.0% 60 1.8% 1.2 / 3088
git cherry 2302 68.5% 417 12.4% 251 7.5% 107 3.2% 1.4 / 3077
git cherry-pick 1455 43.3% 604 18.0% 649 19.3% 364 10.8% 2.0 / 3072
git cherry-pick -n / --no-commit 2182 64.9% 476 14.2% 258 7.7% 150 4.5% 1.5 / 3066
git citool 2601 77.3% 186 5.5% 111 3.3% 159 4.7% 1.3 / 3057
git clean 1483 44.1% 602 17.9% 417 12.4% 547 16.3% 2.0 / 3049
git add + git commit 278 8.3% 183 5.4% 393 11.7% 2189 65.1% 3.5 / 3043
git commit -a 431 12.8% 363 10.8% 535 15.9% 1711 50.9% 3.2 / 3040
git commit <file>... 999 29.7% 468 13.9% 529 15.7% 1039 30.9% 2.5 / 3035
git commit -i <file>... 1849 55.0% 508 15.1% 338 10.1% 337 10.0% 1.7 / 3032
git commit --amend 1294 38.5% 470 14.0% 620 18.4% 646 19.2% 2.2 / 3030
git cvsexportcommit 2851 84.8% 103 3.1% 40 1.2% 35 1.0% 1.1 / 3029
git cvsserver 2819 83.8% 122 3.6% 54 1.6% 30 0.9% 1.1 / 3025
git daemon (pushing enabled) 2522 75.0% 300 8.9% 112 3.3% 89 2.6% 1.3 / 3023
git difftool 2253 67.0% 377 11.2% 245 7.3% 146 4.3% 1.4 / 3021
git ... --dirstat 2348 69.8% 339 10.1% 199 5.9% 134 4.0% 1.4 / 3020
git fetch [<options>] 1156 34.4% 656 19.5% 604 18.0% 602 17.9% 2.2 / 3018
git filter-branch 2215 65.9% 560 16.7% 176 5.2% 64 1.9% 1.4 / 3015
git format-patch 1778 52.9% 521 15.5% 421 12.5% 292 8.7% 1.7 / 3012
git grep 1949 58.0% 467 13.9% 254 7.6% 339 10.1% 1.7 / 3009
git imap-send 2748 81.7% 142 4.2% 70 2.1% 47 1.4% 1.1 / 3007
git instaweb 2375 70.6% 332 9.9% 183 5.4% 116 3.4% 1.3 / 3006
git log --grep/--author/... 1408 41.9% 619 18.4% 569 16.9% 407 12.1% 2.0 / 3003
git log -S<string> (pickaxe search) 1987 59.1% 515 15.3% 326 9.7% 172 5.1% 1.6 / 3000
git log --graph 1692 50.3% 662 19.7% 361 10.7% 278 8.3% 1.7 / 2993
git merge 809 24.1% 581 17.3% 820 24.4% 781 23.2% 2.5 / 2991
git merge with strategy 2151 64.0% 517 15.4% 207 6.2% 115 3.4% 1.4 / 2990
git merge --squash 2008 59.7% 490 14.6% 278 8.3% 208 6.2% 1.6 / 2984
git mergetool 1839 54.7% 400 11.9% 304 9.0% 431 12.8% 1.8 / 2974
git pull (no remote) 732 21.8% 510 15.2% 575 17.1% 1147 34.1% 2.7 / 2964
git pull --rebase [<options>] 1251 37.2% 603 17.9% 504 15.0% 580 17.2% 2.1 / 2938
git pull <remote> 486 14.5% 550 16.4% 738 21.9% 1090 32.4% 2.8 / 2864
git pull <URL> [1]

Cite error: <ref> tags exist, but no <references/> tag was found
Personal tools