This is the wrong question to be asking. It's all too easy to focus on git's warts and formulate an argument about why subversion is ostensibly better, at least for some use cases. The fact that git was originally designed as a low-level version control construction set and has a baroque linux-developer-oriented interface makes it easier for the holy wars to gain traction and perceived legitimacy. Git proponents bang the drum with millions of workflow advantages, which svn guys proclaim unnecessary. Pretty soon the whole debate is framed as centralized vs distributed, which serves the interests of the enterprise svn tool community. These companies, which typically put out the most convincing articles about subversion's superiority in the enterprise, are dependent on the perceived insecurity of git and the enterprise-readiness of svn for the long-term success of their products.

但问题是:Subversion是架构的死胡同。

Whereas you can take git and build a centralized subversion replacement quite easily, despite being around for more than twice as long svn has never been able to get even basic merge-tracking working anywhere near as well as it does in git. One basic reason for this is the design decision to make branches the same as directories. I don't know why they went this way originally, it certainly makes partial checkouts very simple. Unfortunately it also makes it impossible to track history properly. Now obviously you are supposed to use subversion repository layout conventions to separate branches from regular directories, and svn uses some heuristics to make things work for the daily use cases. But all this is just papering over a very poor and limiting low-level design decision. Being able to a do a repository-wise diff (rather than directory-wise diff) is basic and critical functionality for a version control system, and greatly simplifies the internals, making it possible to build smarter and useful features on top of it. You can see in the amount of effort that has been put into extending subversion, and yet how far behind it is from the current crop of modern VCSes in terms of fundamental operations like merge resolution.

现在，对于那些仍然相信Subversion在可预见的未来足够优秀的人，我有一个发自内心的不可知论的建议:

Subversion永远不会赶上从RCS和CVS的错误中吸取教训的新型vcs;这在技术上是不可能的，除非他们从头开始重新配置存储库模型，但这样就不是真正的SVN了，不是吗?不管你认为自己有多不具备现代VCS的能力，你的无知也无法保护你远离Subversion的陷阱，其中许多情况在其他系统中是不可能或很容易解决的。

It is extremely rare that the technical inferiority of a solution is so clear-cut as it is with svn, certainly I would never state such an opinion about win-vs-linux or emacs-vs-vi, but in this case it is so clearcut, and source control is such a fundamental tool in the developer's arsenal, that I feel it must be stated unequivocally. Regardless of the requirement to use svn for organizational reasons, I implore all svn users not to let their logical mind construct a false belief that more modern VCSes are only useful for large open-source projects. Regardless of the nature of your development work, if you are a programmer, you will be a more effective programmer if you learn how to use better-designed VCSes, whether it be Git, Mercurial, Darcs, or many others.

其他的回答很好地解释了Git的核心特性(这些特性非常棒)。但是还有很多小方法可以让Git表现得更好，并帮助我的生活更加理智。以下是一些小细节:

Git has a 'clean' command. SVN desperately needs this command, considering how frequently it will dump extra files on your disk. Git has the 'bisect' command. It's nice. SVN creates .svn directories in every single folder (Git only creates one .git directory). Every script you write, and every grep you do, will need to be written to ignore these .svn directories. You also need an entire command ("svn export") just to get a sane copy of your files. In SVN, each file & folder can come from a different revision or branch. At first, it sounds nice to have this freedom. But what this actually means is that there is a million different ways for your local checkout to be completely screwed up. (for example, if "svn switch" fails halfway through, or if you enter a command wrong). And the worst part is: if you ever get into a situation where some of your files are coming from one place, and some of them from another, the "svn status" will tell you that everything is normal. You'll need to do "svn info" on each file/directory to discover how weird things are. If "git status" tells you that things are normal, then you can trust that things really are normal. You have to tell SVN whenever you move or delete something. Git will just figure it out. Ignore semantics are easier in Git. If you ignore a pattern (such as *.pyc), it will be ignored for all subdirectories. (But if you really want to ignore something for just one directory, you can). With SVN, it seems that there is no easy way to ignore a pattern across all subdirectories. Another item involving ignore files. Git makes it possible to have "private" ignore settings (using the file .git/info/exclude), which won't affect anyone else.

David Richards关于Subversion / GIT的WANdisco博客

The emergence of GIT has brought with it a breed of DVCS fundamentalists – the ‘Gitterons’ – that think anything other than GIT is crap. The Gitterons seem to think software engineering happens on their own island and often forget that most organizations don’t employ senior software engineers exclusively. That’s ok but it’s not how the rest of the market thinks, and I am happy to prove it: GIT, at the last look had less than three per cent of the market while Subversion has in the region of five million users and about half of the overall market. The problem we saw was that the Gitterons were firing (cheap) shots at Subversion. Tweets like “Subversion is so [slow/crappy/restrictive/doesn't smell good/looks at me in a funny way] and now I have GIT and [everything works in my life/my wife got pregnant/I got a girlfriend after 30 years of trying/I won six times running on the blackjack table]. You get the picture.

http://subversion.wandisco.com/component/content/article/1/40.html

我认为可以很有把握地说，在开发人员中，SVN和Git的争论已经激烈了一段时间，每个人都有自己的观点，哪个更好。这甚至在我们2010年及以后关于颠覆的网络研讨会的问题中被提出。

我们的开源和Subversion公司总统总监Hyrum Wright谈到了Subversion和Git之间的区别，以及其他分布式版本控制系统(DVCS)。

他还谈到了Subversion中即将发生的变化，比如下一代工作拷贝(WC-NG)，他认为这会导致许多Git用户转换回Subversion。

请观看他的视频，并通过在这个博客上评论或在我们的论坛上发帖来告诉我们你的想法。注册很简单，只需要一点时间!

Git和DVCS通常都非常适合于独立编写大量代码的开发人员，因为每个人都有自己的分支。但是，如果您需要从其他人那里进行更改，她必须提交给她的本地回购，然后她必须将该更改集推给您，或者您必须从她那里获取更改集。

我自己的推理也让我认为，如果你做集中发布之类的事情，DVCS会让QA和发布管理变得更加困难。必须有人负责从其他人的存储库中进行推送/拉取，解决任何在最初提交时就可以解决的冲突，然后进行构建，然后让所有其他开发人员重新同步他们的回购。

当然，所有这些都可以通过人工流程来解决;DVCS只是破坏了一些由集中式版本控制修复的东西，以便提供一些新的便利。

它是分布的。基准测试表明，它的速度要快得多(考虑到它的分布式性质，像diffs和log这样的操作都是本地的，所以在这种情况下，它当然要快得多)，而且工作文件夹也更小(这仍然让我大吃一惊)。

当您使用subversion或任何其他客户端/服务器版本控制系统时，您实际上是通过签出版本在您的机器上创建工作副本。这代表了存储库外观的时间快照。通过更新更新工作副本，通过提交更新存储库。

使用分布式版本控制，您没有快照，而是拥有整个代码库。想要一个3个月大的版本吗?没问题，3个月前的版本还在你的电脑上。这不仅意味着速度更快，而且如果您与中央服务器断开连接，您仍然可以执行许多您习惯的操作。换句话说，您不仅拥有给定修订的快照，而且拥有整个代码库。

您可能认为Git会占用大量硬盘空间，但从我看到的几个基准测试来看，它实际上占用的空间更少。不要问我怎么做。我的意思是，它是由莱纳斯构建的，我猜他对文件系统略知一二。