ISQL app migration option, Progress - "OpenEdge" Personal RDBMS with 4GL(ABL) development system?

Question

Back in 1985, when I was a Product Planner for AT&T-IS Labs (Unix Product Management), one of my primary duties was to evaluate emerging 4GL/RDBMS products, submit product plans and recommend whether AT&T should pursue a co-marketing or co-labeling agreement with the vendor or just test/certify it as compatible on our 3B-series systems. Some notable products I submitted plans for were: Informix-4GL 1.1, Microfocus VS/COBOL Workbench, etc.

One particular sexy product which captivated my attention was Progress. Although, at that time, it didn't have SQL support, its features/functionality, integrated RDBMS engine and integrated debugging editor, which allowed me to quickly execute interpreted code and compile it once it was debugged, appealed so much to me that I recommended AT&T pursue a co-marketing agreement with the condition that SQL support be added to it.

Well, 25 years later and SQL support added, I'm looking again at Progress as an option to re-write my INFORMIX-SQL based, single-user, pawnshop management app for Windows and UNIX platforms.

I don't require networking or multi-user capabilities at this time, but that could easily change and would like that option as an add-on/upgrade option, which Progress offers. The personal (single-user) RDBMS engine costs only $246/Windows computer and the 4GL (character and GUI-based) development system costs about $2,600 (one-time charge).

Anyone here in SO have development experience, exposure with or have comments about Progress?

I'm in the process of obtaining an eval copy and would appreciate any input.

Answer 1

It's a wonderful environment. I love it.

Answer 2

I am currently programming Progress (now actually known as OpenEdge ABL) for a living. I have been doing so full-time for a full year; I will be at this job for another year and will be very happy to leave.

Progress is simply a horrible language for general purpose programming. Its strength, as you mention, is the integrated RDBMS system and the ability to do querying directly within the language. This was, I'm sure, pretty amazing in 1985. However, for anything beyond easily generating some printed reports (a popular thing to do in 1985 I'm sure), it is really not that powerful. For instance, even with the 'strength' of the integrated DBMS, Progress doesn't even have a concept of primary key / foreign key relationships between tables. The closest you can come is to name your fields the exact same in separate tables, then you can do a query like

FOR EACH order-line OF order ...

but that is the only way. This can lead to maintenance nightmares when you have hundreds of tables and no ability to generate an entity-relationship diagram between them because the geniuses who designed the system 20 years ago didn't understand the severe limitations of the core DBMS. This is an enormous weakness for a language that's supposed strength is in the DB area.

As far as general purpose (non-trivial) programming, you basically have to rely on Progress to write methods that are integrated into the core ABL system for you, because the language is simply not flexible enough to extend on your own (believe me, I've pushed it to it's limits). They take the approach of "let Daddy do that for you", which was probably awesome in 1985, but today languages are all about extensibility (for obvious reasons).

If I were you, I would explore something like Java or Ruby using MySQL as the DBMS (all 100% free). I have recently been doing some hobby programming using object-relational mapping in both languages (JPA2 in Java, and DataMapper in Ruby). To me, being able to persist objects almost transparently is much, much more natural than trying to balance hundreds of tables without any PK/FK support and infinitely easier to maintain farther down the road.