Linking and segmenting seem to serve the same purpose to me....If i want to divide work between my developers, I could create multiple new models and later link them into the main model...Or as user guide, you can create segments from the main model...but, in the end they do the same stuff in almost the same way...So which do we use where and whats the distinction
Hi,
Have a look at the proven practice document below. It gives a fuller description than the paltry ones found in the FM guide and also scenarios for when you would use it. Scenario 1 deals with the segmenting concept while scenario 2 the linking concept.
From my understanding Segmeting would be used where you have a layer of common metadata and you wish to create seperate projects that will utilise this common meta data and allow developers to build upon that. For example your main project may only contain information common across an organisation while the segmented projects may build upon that e.g. HR, Sales, Finance and their specific structures. As mentioned in the document this is hub and spoke.
Linking on the other hand is where you wish to bring together content from already existing projects.
http://public.dhe.ibm.com/software/dw/dm/cognos/modeling/design/enabling_a_multi-developer_environment.pdf
Hope this helps
Cheers
Rob
My preference is for segementing. This allows you to have developers working on separate 'subject area' based models which you can then bring together into a master model. Because each segment is a standalone project (or model) it is easier to deploy segments separately into different environments rather than having to deploy the whole master model where some areas may not be ready to go live.
Good luck