-3

I am wondering how do big development teams do it, when they all work on one project.

Options are:

  • 1 virtual machine in cloud that they ssh into and push updates via ftp ?
  • each developer have their own local virtual machine and updates are being controlled by git, when merged together by master branch on cloud?
  • any other ideas?

Basically I wanted to know what the structure and workflow are.

Workflow for developing a web app for example. Lets assume no testing needed.

I wanted to know what are the basics steps: 1 virtual machine that everybody logs into and commits changes, or virtual machines are on each developer local drives?

And how do they commit changes based on above?

I am not asking for a solution to a problem, only the general knowledge of how things are being done properly in web-dev environments. And how to avoid merging traps for example, etc.

1
  • Workflow for what? Can you edit the question to explain what would happen on the VM? Are you using this VM for testing the developer's changes? How do you currently do testing?
    – amon
    Feb 6, 2018 at 11:24

1 Answer 1

2

There is no need for a “cloud” environment at all. Every developer should have their own development environment. This may be local or on a server, and this may or may not be a virtual machine. The important point is that developers can work on their own, without accidentally overwriting the work of another team member.

Version control is important for collaboration because it allows the changes of multiple team members to be merged together without accidentally overwriting something. With Git, a pull-request based workflow gives other team members the opportunity to review proposed changes before they are merged into the master branch. Automated tests as part of this review are considered to be a best practice. These automated tests would typically run on some build server, which might be a VM on a shared server. The build server runs automated, so developers should not log in except for debugging or administration.

The master branch can then be deployed to a production environment. This may happen semi-automatically if you're trying to do Continuous Deployment. Because all changes have already been tested and reviewed as a pull request, you can always deploy the master branch. If there should be a problem, you can always revert the master branch to a previous, known-good state and redeploy.

Especially for Git, there are many workflows that discuss how these general concepts can be structured precisely. They differ in some goals, e.g. the popular Git Flow assumes infrequent, well defined releases. Especially for webdev, a workflow centered around continuous deployment may be more appropriate.

These workflows also scale differently. As long as there's a single team there shouldn't be a problem, but multiple teams collaborating on a single code base are always more difficult. It is then necessary to have a project manager or product manager who coordinates the work between teams. Neither version control nor virtual machines can solve communication problems. You need people for that.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.