This sounds like a good idea in some contexts. I like the fact that you can merely alter the tables in the database, then regen the app. But, the key to bigtop is that it leaves hand written code alone in such cases. While some apps are just database front ends, almost all of ours do at least a little bit more than that. So, they have custom code. With bigtop, we can regenerate to take new data model changes into generated code, without overwriting custom code. (Of course, we have to revisit hand written code, to make it reflect the updated data model. But regenerating leaves that code undisturbed.)
Phil | [reply] |
It's not that naive as you might think. Actually it generates only a directory structure with files with no code only a 'use base ...' declaration and a database description. I think the regeneration after a schema change can be made safe - that is it only needs to overwrite the database description and eventually add new classes but it does not need to overwrite the hand written code.
| [reply] |