Seperation Model and Relationships

[Kouga]

I want to seperate data and layouts so I can create new layouts without having to worry about modifying or importing data when I add new layouts. I'm using FMS7 Advanced

Here's what I think the process is like:

1) Create all your data tables in some fp7 files.

2) Create Layout files

3) Create File References to needed data files in layout files.

Q1) Do the data files need to be hosted by FMS, or just the layout files?

The users should only be interfacing to the data through the layout files, but I'm not sure if the data file references MUST be through the server in order to make them multiuser.

Q2) Do the relationships between table occurances need to be made in the data files, the layout files or both?

Last Question:

Is there a good book about how to do this (yet)?

TIA~

[Ender]

1. Although it is not necessary for the data files to be hosted on the server, it is likely that they would be in most cases, as the real advantage of using Server is sharing the data. If you're worried about security, you can simply restrict access to layouts in those data files.

On a related note, I've seen solutions where much of the data was centralized on the server, but the interface file and certain static tables were installed on the local workstation. This improved the interface speed for a WAN deployment.

2. Anything the interface accesses must have a relationship in the interface file. I believe the data files need some relationships for lookups and auto-enter calcs.

[Fitch]

Most of your relationships (TOs) will be in the interface file, used for queries (GTRR) and presentation (portals). Ideally there would be NO TOs in the data files, they would just be pure data; but in practice this is difficult, since in FileMaker we are used to caclulated fields that reference related data, and the other things Ender mentioned. In any case I try to minimize the non-data elements in the data files.