Book a Demo

Author Topic: Working in a distributed team  (Read 2544 times)

shainegordon

  • EA Novice
  • *
  • Posts: 1
  • Karma: +0/-0
    • View Profile
Working in a distributed team
« on: September 18, 2004, 08:09:28 am »
My collegues and I work on the opposite side of the world. We run a subversion version control system.

We tried to store our models in a MSSQL DB, but this was just too slow(if you could work offline without doing a data transfer that would be great).
Replication is also a problem, because the replica's are too large for everyone to send them to a central location to be merged, and this is not immediate.

Form this it seems that using controlled packages is the best route to go.
However I have some concerns that I hope someone can help me with.

I would assume that the actual .EAP file will not be put under version control, and each user would have a local .EAP file on there pc.

My major issue is. What happens when someone creates a new package, sets it up as a controlled package and exports it. They then add that xml file to version control, the other users update there working copies and get the new XML file. But, where do you import that package, so that the tree structure is the same among all team members working on that project.


I assume it will also be a problem if two people work on the same package at the same time and try to update and then commit their work, as a conflict will occur

thomaskilian

  • Guest
Re: Working in a distributed team
« Reply #1 on: September 22, 2004, 07:40:52 am »
shainegordon,
I guess the answer is a simple as you might have already thought. If someone adds a new package and distributes it via version control then the others have to import it at the right place (you have to tell where). There is unfortunately nothing (I know of) that prevents you from importing it at the "wrong" place. The best would be to have a centrally operating person to control new packages and instruct every user to place it at the right position. Also, since this should not happen too oftern, when creating a new package you could consolidate the current model and distribute that. A bit slower but likely more safe.