background: I create user entry pages where users come and enter their details. These details go to our database and sometimes I also send them to our clients.

Sometimes there may be problems and I am saving users' data on the server outside root of course.

Now I am writing a script that will check this folder automatically ( cron job) and then try to process the data. If the data needs to go to our partner, i would have to factor this into our script as well.

so far this is how i am thinking of doing:

folder structure:

            either file.csv or files with one xml each
script will then read all directories underneath data folder, open them up read config.txt determine if it is csv or xml and then send the data to our database/

Now i am thinking that i should just have CSV because if i have xml then i would need to "extract" the data if i need to send it to my partners. For files in csv, i am thinking to provide an xml template and then use sprintf to put csv into xml.

Q; how do i factor in the clients system? i would probably need to write a class per client anyway but how do I use this in my script

Q: does anyone see any potential problems?

Q: should I use this script as a linear procedural way or try to make it all OOP. (Note: procedural way will be using some classes ) but the script is nothing more than looping through directories to get the data etc.