I am an inhouse developer for a solution that is 10 years old, Originally in Filemaker v5 and was 102 files, with the release of v7 it was converted down to 13 files, and now more recently I have created a UI file for the interface but the "data" files still contain the layouts and scripts that formed the basis of the UI (incase I had to backtrack).
The users have been in the UI file for 12 months now.
I am trying to make the decision whether to now make the UI file 1 big data AND UI file which would effectively consolidate the following files:
Which in one folder = 4.55GB
The UI currently has just over 1000 TO's which would all need repointing to local tables once I imported them and the broken calcs would need uncommenting... (this could cause some RSI as about 3,700 of the 10,500 fields are calcs!).
Some futher TO's would need to be to be added that are not yet created in the UI as I did not need them for scripts or portals but will need them for calcs if I make this change.
Then the data would need importing. Not sure that it would be possible to get this much data imported over a weekend period?
I think it would be about 2 weeks solid development work (not sure if this is optimistic or not) and I would have to put a freeze on any schema changes in the live system during that time.
The other option in to leave the data files separate but go though them and clean out all the layouts and scripts so they really are more like data files. This would be less work and potentially has a lot less room for errors but the TO's in these files are a REAL MESS whereas the UI file has been kept fairly clean in terms of anchor bouy. Truth be told I am always a bit embassaed when I look at the graphs in these files!
There are people that do not like the separation model but then there are others that swear by it. I am not fussed either way but I want to make a sensible decision and then get on and clean things up either way. As our solution is not a shrink wrap we do not have a need to "swap out" the UI, its creation was originally motivated by a corrupted file that needed re-doing from scratch and it grew from there once I realised I could develop it alongside the production version without having to re-import the production data.
We are lucky enough to have a Kove server so backup time of a large file would not be an issue.
I have always thought it was best to have the data separated out into multiple files in case we have a file corruption and need to reimport the data (as this can be very slow… days slow) but there seem to be two schools of thought on this also?
I am not scared of a hard slog but definately want to make sure the path I go down is the better advised one before I start out. I hope you do not mind me using the forum to seek some opinions.