Apache Subversion

Retain a history of changes to a set of files

Allow multiple people to collaborate on a set of files and merge changes together

Apache Subversion works great when all users have access to the central Subversion server. This might mean users need to be on the same network.

Apache Subversion isn't great for "offline" work as it requires access to the central Subversion server.

Easy install

Integrates into agile project standards

Revision regressions tracking in projects

Requirements more straight forward (CVS editing options)

Some packaged dependencies are unnecessary

Potential for vendor packaged security issues

Track and control concurrent versions of the same files.

Has good support from many different software, including visualization, DevOps toolchain.

Well documented and understood by developers as it has been around for a long time.

Does not support distributed environment unlike Git.

Merging can be painful.

Does not natively support advanced features such as pull requests.

Revision control done properly - you have end to end visibility of all changes in the project.

Conflict resolution - visually highlighting the differences helps to track down the problem.

Being open source and very popular.

We are using SVN hosted in our network - it is very stable, we had almost zero downtime in 4 years.

Rollbacks are made simple and easy to use.

It is missing the pull request feature which Git has. You can still do it in SVN but more work is needed.

It is centralized. Nowadays software developers and teams need more flexibility and will choose Git for that.

Performance is not a strength of SVN pulls and commits.

The disk space use by working copies is almost double due to the way SVN organizes its working files.

Less support for .NET developers since it comes from the open source world.

Code reviews could be made simpler to help the reviewer more.

Can be used from many locations, like a cloud-based system but with more custom control--and its free!

Multiple projects easily stored in a single repository, which aids in maintenance of common code, yet also easily allows for separate repositories where no sharing is desired.

Very stable, with lots of additional tools to help maintain and examine repositories (e.g. websvn).

Much easier to understand when coming from more traditional SCM systems like CVS and Perforce (as opposed to Git, which is a bit of a paradigm shift).

Refactoring the layout of a respoitory--or a part of a repository--can be a bit painful, especially for users with workspaces associated with the affected part of the repository. Not sure what could be done to make that better, but it would be nice if something was possible.

Folks coming from Git can have problems using Subversion. Again, not sure anything can (or should) be done to address that, but it is occasionally an issue.

Software versioning

Very stable product

Easy to use

The installation requires some initial configuration

Improved interface

Inconvenient update management

Version control - it's what it's designed for.

Modifiable - It only takes a little bit of knowledge of a scripting language (I used Windows BAT files calling Perl scripts) to extend capabilities, like the aforementioned integration with Redmine.

Back end administration- It's a breeze. There's very little work involved in terms of administering it once you've got it installed on a server, and even setting that up isn't bad.

Distributed development - I've never worked in an environment where distributed development (developers widely scattered geographically) was a factor, but that's why Git exists.

Merging - Merging of code from one branch to another can be painful, especially if it's not done frequently. (On the other hand, doing merges is one of the reasons I get a nice salary, so I can't complain too much!)

Acceptance - Let's face it, Git is what "all the cool kids are using." If you've got a bunch of developers fresh out of school, they'll probably know Git and not Subversion.

Old is gold. Apache Subversion has existed before many other version control systems, including Git. It's old, stable, and easy to use with not many complications.

Excellent versioning system. You can jump between any particular version of your project to any other version just by reverting or updating, you can also create patches of your own changes and then apply the patch on your own system on a different check out or on someone else's computer who has a checkout copy.

You can search for any older commit by using words used in the comment log or by using an exact commit number or anything in between. You can also check the log of each and every individual file instead of the whole checkout.

Tortoise SVN is a client for Apache Subversion. It has one of the best UIs I have ever seen for a version control system.

Merge conflicts is one area where I think that Apace Subversion can improve a lot in. Where there is a single file being edited by two different people and the person who tries to update after someone before him commits with changes on the same file then Subversion tries to merge the changes and create an ideal file but fails miserably.

Any file renames or deletion or additions have to be specifically made through Apache Subversion or has to be notified to Apache Subversion in a round about fashion or it will be disregarded while committing the changes. While this is reasonable, it is quite annoying until a user gets used to it.

Faster on Linux and slower on Windows. Apache Subversion can be improved by increasing its checkout, update and commit speeds on Windows.

User friendly and easy to learn and understand

Free software

Easy to keep track of code versions

Not distributed like Git

Git more user friendly and advanced. Git has lot more features .

The default conflict resolution option, to merge locally, has led to a much more efficient work environment when working with large teams on large codebases. The traditional single-person file locking can really get in the way of team work, as you have to wait for your team member to finish their changes before you can start working on the same file, even if they called in sick for work that day. While sometimes this requires manually figuring out what to do when two changes affect the same line of code, most of the time the changes are on the same lines of the file, and merging can happen transparently.

I have enjoyed the branching process in subversion. Branches and tags are not strict features of the product, which allows for fudgibility, but when you use the recommended trunk/tags/branches folder layout, it behaves as if it was built it. Implemented simply as copy/branch and merge functions, I have found them to work just as well as a built it system would work, and it does a good job pointing out issues with a change's ancestry.

Subversion also have a rich ecosystem of third-party tools and service providers. I personally have used TortoiseSVN for years, but there are several plugins that integrate directly into Visual Studio or Eclipse. Also, I have found hosting services like CVSDude (now called CloudForge) to be a big time-saver over hosting a repository on your own servers, while providing peace of mind that your code-base is in a different physical location, in case say, your server farm burned down. (I'd call that a serious edge condition, but my job involves edge conditions!)

At times, locking problems can be difficult to solve. This normally happens you make a mistake, like attempting to update a folder that has running executables within it. Often times, you can just unlock a parent folder and you' are all set, but on more than one occasion I have had to recheckout a folder because I could not resolve the lock, even after rebooting.

Looking up the history of a file can be very slow, taking several minutes, especially when looking at the history of an entire folder.

Novice users of subversion often make the mistake of dragging folders in Windows that are managed by subversion. This does not cause the folder to move in the svn repository, which can lead to serious confusion why a user's folder structure does not seem the synchronize correctly with other users. To an advanced user, this can be a very useful feature, but absolutely hell for the novice.

I have not found any third-party tools yet that let me visualize the commits for a codebase, particularly across branches. This may be partially because of the slow history I already pointed out, but I believe it's because I have not been willing to even look at the multi-thousand dollar code management solutions I have seen advertisements for in the past.