Progress codebase analysis using Moose

Question

Has anyone used Moose to deal with a large legacy Progress ABL/4GL codebase, and if yes, what are the caveats, experiences, lessons learned you can share about this?

Answer 1

From building a Delphi parser for Moose I know that it takes quite a lot of effort to get your FAMIX model complete. I'd recommend to make your feedback cycle as short as possible. From the progress grammar I'd say the system looks a bit more complicated than Delphi. Focus on the real & perceived problems and take shortcuts to find solutions for them. So you might want to delay writing a parser in Smalltalk, and just start using the pro parse stuff to generate a MSE file containing the interesting bits.

If the system is really large you might have to be careful not to run out of memory currently.

You are aware of the humane assessment site? The blog provides lots of examples and things to do/don't.

It really is an environment for developers, extremely powerful and having still a lot of rough edges. The mailing list is very helpful

When dealing with legacy systems, there are a few things I do before starting Moose, to get a feeling for the system. I run a duplication detector and zip the individual table dumps (that works well on progress files). Large files that compress extremely well are interesting discussion topics :)