Best practices for multi-user FM modelling?

[aitzazali]

Hi all,

What is the best practice for multi-user FM modelling. Branching and Merging OR Segmentation & Linking?
We have a team working on a single project.I will be working on the metadata, relationships,  of one business subject area and at the same time the other team members will be working on  other business areas.  At the end, we need to merge all our projects in a single FM project.

Second, is it possible to merge different packages in one. Like, if every team member creates his own package to test the reports query, can we merge those packages in one package in the main project?

[blom0344]

If Business Subject area's can be contained in their own namespaces using a single database layer you may consider building seperate models and importing the namespaces in the latest phase into a new 'uber' model.

Merging packages is not possible, publishing a package from a merged model is no problem

[aitzazali]

Thanks Blom0344. much appreciated.
So basically I will create segments from the main project according to the business areas and then link those segments to the main project?

[blom0344]

No , segmentation would keep the metadata  distributed over seperate models, which is an alternative.  Importing in the strictest sense is creating one ultimate metadata model. 

[wyconian]

Hi

I'm a big fan of segments especially if you have multiple work streams working on different subject areas of the model.  You can link the segements into a master model where you can publish packages. 

On the downside it can get a little difficult to manage if you have a lot of segments but on the plus side you can have developers independantly working on different areas of the model.