In a document application I had to interface to, it was envisioned to have 2 large components - one to do with the customers clients and the other to do with "miscellaneous" documents (regulatory matters, contracts etc - stuff not specific to a client but to their industry).

This did not seriously impact the overall design as the software allowed for 2 "sites" - except for user permissions which were drastically different between the two sites.

Both got implemented and I had to spend tons of hours testing the permissions logged in as different user levels. Then when I found an incorrect permission, send it back to the developer and around we go again until all works.

The "client documents" side got used right off the bat and is heavily used - this is where the money is made for this customer. The "miscellaneous" branch, except for one feature (contracts), all the other "miscellaneous" stuff that was supposed to go in there (like announcements regarding the merging of companies etc), after 4 years, not implemented. In the end, last month they decided to not use it as they said it was just easier to maintain the stuff on their file server. Oh well, probably 50k of development and testing time mostly down the tubes...

It's made me a bit more leery when planning - to try and ascertain whether they have the "human ability" to actually implement something - do people have the time to use this new fangled system? If the system does not save them time and is at least as easy to use, they ain't going to use it (other than by edict from the higher-ups). In this case, it was easier to just use the old - the only advantage to the new system was "security" - it was easier to lock down the documents in the new system - but the only people who cared about that was one of the owners and their IT staff.

Albert

>I wish I have written down every bloody feature from day one (back in 1986) which got implemented by special request and then got ditched because the person asking for it 1) changed his mind, 2) got promoted, 3) just vanished or, maybe the worst, 4) kept using it but nobody else in the world didn't want it.
>
>The provisions for future expansion may be just places where code may be cut into smaller pieces, because such pieces may be used elsewhere, or various before* and after* methods, where only a slot is opened but no actual code written (but a simple code to such an empty method). But that's as far as I'd go. Fields in the tables designated for future use... nyah.