{"thread":{"id":"25508","subject":"Git User's Survey 2010 very short summary","startedAt":"2010-10-21T21:24:56Z","lastAt":"2010-10-21T21:24:56Z","messageCount":1,"participants":["Jakub Narebski"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"154032","messageId":"201010212324.58048.jnareb@gmail.com","threadId":"25508","inReplyTo":null,"subject":"Git User's Survey 2010 very short summary","fromName":"Jakub Narebski","fromEmail":"jnareb@gmail.com","sentAt":"2010-10-21T21:24:56Z","receivedAt":"2010-10-21T21:24:56Z","isPatch":false,"sender":{"key":"jnareb@gmail.com","avatar":"https://avatars.githubusercontent.com/u/2706?v=4"},"body":"You can see the results of \"Git User's Survey 2010\" at Git Wiki:\n\n  https://git.wiki.kernel.org/index.php/GitSurvey2010\n\n\nYou can also view them and analyse using provided filters at\nSurvs.com:\n\n  https://www.survs.com/results/33Q0OZZE/MV653KSPI2\n  http://tinyurl.com/GitSurvey2010Analysis\n \n\n..................................................................\nBelow there is very short summary of some of the results of\n\"Git User's Survey 2010\"\n\nTotal responders: 8841 (compared to less that 4000 in 2009)\n\nAbout you\n^^^^^^^^^\nMost responders are from USA among other coutries (31%), followed by\nGermany, UK and France (10% to 6%).  Most responders are from Europe\namong other continents (50%), with North America second (36%).\n\nMost responders are between 26 and 30 years old; most common age (mode)\nis 25 years old.  Youngest user who answered this question is 10 years \nold (next oldest is 12 years old), oldest is 87 (next to oldest is 81).\n\n\nGetting started with Git\n^^^^^^^^^^^^^^^^^^^^^^^^\nResponders say that Git is reasonably easy to learn, and easy to \nreasonably easy to use; it is easier to use than to learn.\n\nMore than 80% use git version 1.7.x at least at one place.  There are 12 \nunfortunate stragglers who use pre 1.3.x version.\n\nResponders fall somewhere between 'everyday use' (39%) and 'can offer \nadvice' (32%) proficiency.\n\n\nHow you use Git\n^^^^^^^^^^^^^^^\nMost people (74%) obtain/install Git at least at one place using ready \nbinary packages.\n\nGNU/Linux dominates among operating systems used, then there is MacOS X,\nthen MS Windows (where msysGit is preferred to Cygwin).\n\nAlmost 9% of responders use JGit.\n\ngitk dominates among graphical tools (58%), GitX is second (28%), \nclosely followed by git-gui (25%).\n\nThe leader among git hosting sites is (like in previous years) GitHub \nwith 77% of responders (up from 62% in 2009); \"self hosted\" is second \nwith 44%, but it is quite lower than 57% in 2009.  The high position of \nGitHub migh be at least partially caused by the fact that GitHub has \ngood announcement system for new features, which was used to announce \n\"Git User's Survey 2010\" to GitHub users.\n\ngit:// protocol and SSH are almost the same often used to fetch changes: \ngit protocol with 78%, SSH with 71%.  Only 41% responders use HTTP for \nfetching (somewhere).\n\n\nMost popular feature, by quite a large margin, is *stash*, with two \nthirds (66.4%) of responses. \"git stash\" first appeared in git version \n1.5.3, from September 2007.\n\nSecond most common used feature is \"shell completion of commands\", with \nmore than half of responses (52.0%).\n\nThen there are, with percentage of responses from 40% to 45%, comitting \nwith dirty tree and interactive / partial commit, interactive rebase, \nand aliases (git aliases, shell aliases, one's own git scripts).\n\nMore than third of responses have also git-aware shell prompt with 35%, \nand submodules with 33%.\n\nLeast used, with only 0.5% of responses, is \"git cvsserver\". \n\n\nOne of the most requested features is native support for tracking empty \ndirectories, with 35.2%, followed closely by request for '-n' like \noption for each command, which would print what would happen, but do \nnot actually execute command, with 33.5%.\n\n\nWhat do you think of Git\n^^^^^^^^^^^^^^^^^^^^^^^^\nMost responders are \"very happy\" with Git, and this answer seems to be a \ncenter of responses.\n\nUser interface, tools and documentation need improvement; performance \nrather doesn't need to be improved; people mostly don't care about \nimproving communities or localization (but please remember that this \nsurvey was in English, and announced on English-speaking sites / \nlists).\n\n\nDocumentation. Getting help.\n^^^^^^^^^^^^^^^^^^^^^^^^^^^^\nIt looks like we have good community surrounding Git, if 62% of Git \nquestions are resolved quickly and accurately, and only 7% couldn't get \nquick response and one to their liking. Those results are very similar \nto those from 2009 survey.\n\nMost used way to request help about Git (help 'channel') is to simply \nask somebody better versed in Git personally, with almost 60%; it is \nthe same result as in previous surveys. This year StackOverflow \n(a Q&A site) came as second with 37%, moving up from 17% last year when \nit was a new thing (!!!).\n\nThe #git IRC channel with usually fast real-time response is now third \nat 18%, down from 31% in 2009 and 37% in 2008. Together with other IRC \nchannels (git/SCM related with 7%, other with 11%) it comes up to 34%, \nclose to StackOverflow 37%. One wonders how disruptive \nhttp://chat.stackoverflow.com would be...\n\nClose fourth, with almost 18% of responses is instant messaging \n(XMPP/Jabber, Google Talk, AIM, QQ, ICQ, etc.), also with real-time \nresponse; 2009 survey had similarly 19% of IM users.\n\n-- \nJakub Narebski\nPoland\n"}]}