Git vs SVN

If you are a programmer, you may have or will eventually come across the need to use a version control system (VCS) in order to make sure that your users are all able to get all the latest updates and changes of the software you are developing. So, what is exactly a VCS? A VCS, also known as version control, revision control, or source control, is a part of software configuration management that is responsible for the task of management of programs, documents, files, large websites, or other collections of information. Revision control systems become a very important part in software development, where a group of people may have access to read and change the same files. Today, revision control systems are usually run as independent, stand-alone applications, but some types of software, such as those content management systems, have their revision control systems embedded.

Two version control systems very popular right now are Git and SVN. There are many who prefer to use Git, but many others choose SVN instead. However, Git and SVN are very different from each other. If you are not sure yet about which one to choose, continue reading below! Here, we are going to break-down the differences and comparisons between Git and SVN.

About Git

Originally made for the development of the Linux kernel, Git was created by Linus Torvalds in 2005. The current maintainer now is Junio Hamano. Just like the Linux kernel, Git is freely distributed under the GNU General Public License version 2. Unlike most client-server systems, every Git directory on any computer is a full-fledged repository, featuring full version tracking abilities and complete history independent of any central server or network access. Git is usually used for source code management, but can also be used to keep track of changes on any set of files. Git is aimed to provide data integrity, speed, and support for distributed non-linear workflows. It now works with Windows, Linux, and macOS.

About SVN

SVN, which is an abbreviation of Subversion, is developed and distributed as open-source software under the Apache License. It was originally created by CollabNet Inc. in 2000, and is currently a top-level Apache project that is continually built and used by a global contributor community. SVN is used to maintain current and historical versions of files, and has been widely implemented in various projects such as FreeBSD, Free Pascal, SourceForge, and Apache Software Foundation. SVN is a cross-platform version control system.

Git vs. SVN: Distributed vs. Centralized

The main point in the whole debate between Git and SVN comes down to whether you prefer a distributed version control system (DVCS) or a centralized version control system. Git is a DVCS; every developer has their own local copy of the complete version history. SVN is a centralized version control system; the version history is stored only in a server-side repository.

The difference is significant. When you are about to make changes, your system will have to check the files to be modified first. Git, as a DVCS, offers much faster local operations as well as an enhanced collaboration. You don’t need a constant connection to the server because most of the operations are performed locally; thus, it can be a lot faster. And since every peer of the same project has each own full version history, the peers can share changes between each other before submitting these changes to the central repository, allowing for a more effective collaboration and ensured code quality.

As a centralized VCS, SVN is easier to understand, especially for non-technical people. SVN works in a client-server manner, and requires a connection to the server whenever you make a commit. Still, in some situations, SVN may perform better than Git. We’ll see this in the next section.

Git vs. SVN: Working with Large Binary Files

Working with large binary files on Git can be quite problematic. Every developer has a full version history on their computer, and the size of this repository grows with every change. So, every change in a large binary file will cause the repository to grow by the size of the said file. You need to prepare the storage capacity.

Meanwhile, SVN only checks for the working tree (the latest versions) on individual client machines. As the effect, changes to large binary files will not bloat the client’s repository as easily. The storage capacity needed on the client side can be notably smaller. Thus, checkout times can be quite significantly faster. This is the very reason why many game developers and web designers, as well as those who manage graphical assets, prefer SVN.

Git vs. SVN: Branching

Branching is the act of making a copy of an object, such as a source code file, in order to work on multiple different versions of the same file simultaneously. You can merge the changes back into one file later. Branching is performed to allow multiple collaborators work together simultaneously without destabilizing the code base.

Branching on Git is generally easier and lighter, due to the fact that a branch in Git is essentially just a reference to a part of the local repository. Meanwhile, on SVN, branching is based on the directories in the central repository – merging branches in SVN can result in problems if there are renamed directories in one of the branches. This is called a “tree conflict”. Such tree conflict, however, can be avoided by limiting or prohibiting renames, and doing refactoring only in the main branch. SVN also uses the copy-on-write strategy so that you will only copy the files to be modified rather than the entire branch from the central repository.

Git vs. SVN: User Access and Security

By default, Git assumes that all contributors have the same permissions; the access permissions for the users can be modified by implementing a separate software program. Meanwhile, SVN allows you to specify read and write permissions per directory level and per file level.

Git allows anyone to modify any part of the history in their local repository. Problems arise when a changed history is pushed while other contributors’ works rely on particular changes. Thus, changing the history of shared branches is generally discouraged. Even though Git does have a natural backup mechanism every time a developer clones a repository to their computer, it is not wise to neglect regular backups on the shared server. Meanwhile, in SVN, the change history remains the same, except in very few rare cases. In SVN, only the developers with access to the centralized server can change the history of the repository.

GitSVN
- Distributed version control system (DVCS)- Centralized VCS
- Faster local operations and enhanced collaboration- Ideal for large binary files due to the smaller required storage and faster checkout times
- Easier and lighter to branch- Branching can be trickier, though there are workaround ways
- Contributors have the same access permissions unless a separate program is implemented- Read and write permissions per directory level and per file level
- The change history may conflict with other developers if not managed well- The change history mostly remains the same

Conclusion

Git is a distributed version control system (DVCS), thus allowing each developer to have a full version history in their computer and perform local operations faster. However, a local Git repository will need a considerable storage space, and this is an issue when working with large binary files. Git allows easier branching, but provides the same permissions to all contributors. On the other hand, SVN is a centralized VCS that works in a client-server manner. The full version history is only stored in a server. SVN requires a notably smaller storage space on the client side. Branching in SVN can be quite trickier, but it allows you to specify read or write permissions.

Leave a Reply