Do you think it's worth it to use version control if you are an independent developer, and if so, why? Do you keep the repository on your own computer, or elsewhere, where it can serve as a backup?
If you use decentralized source control (Mercurial or Git or Bazaar or whatever), you get advantages over SVN/CVS that makes it easy, useful and powerful to use in case you're an indy:
- You commit locally: your project dir is your repo with FULL history. So you don't have to have a server, you commit directly in your repo, and you can have several repos in the same computer. Using a laptop that you open sometimes to continue working on your things? Great! You don't have to setup a server and if you need one later, it's easy and you just "push" and "pull" changes between repositories.
- It's made to ease experimentation: often you need to have an idea about a feature without polluting code. With SVN and CVS you can already use a branching system and ditch the branch if the feature is not as good as you wanted it to be. But if you want to merge the feature with the trunk version, you'll have a lot of hard to fix surprises. Git, Mercurial and Bazaar (at least) makes merges and branches really easy. You can even just duplicate a repo, work on it some time, still commit and kill it or push your changes in the main repo if you want.
- Flexibility of organisation: as pointed before, as you have repos that you organize as you need, it's easy to start alone and allow other people to work with you by changing your organization. No organization is imposed so you just have to set it up and voilà. I often just push/pull changes between my own computers (laptop/desktop/server) and I'm still alone on my devs. I use Mercurial and that help me duplicate my work but also work on features I thought about outside on my laptop, then continue to work on other features on my desktop, then push my laptop changes on my desktop or server and merge the whole desktop+laptop and put it (as backup and future teamwork repo) on my server.
- It helps setting up backups: if you setup a central repo (on GitHub if it's public, or in a private repo on BitBucket) you can easily write a script which will be executed each time a computer is booted, and then pass said script on to your friends so that it makes automatic backups of your work regularly. It's what I'm doing so now I'm sure it won't be easy to lose my work.
In fact, currently, you have no excuse to not use a control source tool for any project. Because they are more powerful and flexible than before and scale with your needs.
source code control is utterly useless for independent developers, because as we all know:
- independent developers never make mistakes
- independent developers never go off on revisions that don't work out
- independent developers never have more than one version so they have no use for branches
- independent developers never care about what they changed yesterday or last week
- independent developers never, ever need backups
Call me a "dependent developer": Mercurial repositories clone easily between my desktop, laptops, USB backup drives, and bitbucket.org. I've grown dependent, and I like it that way!
I'm a solo developer and I use BitBucket and Mercurial for my personal projects. Having the ability to revert and fork your code is just too good to pass up.
I find value in it, personally. My projects are all checked into git repositories (all of which I keep on multiple machines in case of a hardware failure). The most useful features are branching (so that I can run an experiment that messes with half of my codebase, and not worry about blowing anything up permanently) and reverting (which is basically just undo on steroids; in case I find that I made some mistake that is out of regular undo range).
The article is Mac & Git centric but it covers the basics.
You may also be interested in the following StackExchange Questions (and their answers).
Worth?? Must! If you don't use Source Control then you don't control you sources, and that's bad. You can't diff you can't revert, you can't track changes - you'll spend hours trying to find out the dummy bug you just entered. It is better to have it on some backed up server, but you can also you your computer and use any backup method you find appropriate.
Absolutely use source control. Then set up a build server and automate your build and testing processes. Trigger builds from your source commits of your central repo. I have been working by myself for three years in this way and it's wonderful.
Even single developers need to sometimes see the state of their code from some past revision. And it's always a good idea to backup everything important, and that applies to all people.
protected by gnat Jul 20 '15 at 8:33
Thank you for your interest in this question.
Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).
Would you like to answer one of these unanswered questions instead?