[Secureideas-base-devel] Data Layer
Brought to you by:
secureideas,
sinukas
From: Kevin J. <kjo...@se...> - 2004-11-24 15:39:12
|
Ok, I guess it is time to throw this out to everyone for opinion and then make a decision. We want to change the data access layer. We have a couple of choices. 1. Keep ADODB 2. Move to PEAR::DB 3. Roll our own a. Use arrays b. Use objects I believe that we have already shot down #1 and I agree with the idea. 2 will require massive rewrites to everything as all of our display code uses the object model. We also will probably need to wrap the PEAR::DB to add features. (Not sure about this) 3 sounds like the easiest and best on first glance.... 3a has the same problem as 2 where we need to re-write everything 3b would probably let us drop a replacement to ADODB in with minimal code changes. Surprisingly, from my language above I am leaning toward 2 or 3a.=20 Rewriting everything sounds like a LOT of work, but we have a number of developers on board now and we need to optimize the code anyway! 3b sounds like we could limit the object to only having the features we need, but we would be troubleshooting our own code. 2 would require that we check for the existence of PEAR::DB and if its not there what then? =20 I would like to get started on this soon, as in later today, so if we could get as many opinions as possible... Of course, I am going to get started anyway, since no matter which way we go, we are going to need to prep for it.... Thanks Kevin |