It is interesting to look back at one's own code from years ago. It's like having one's own Wayback Machine.
FIJI the ForthIsh Java Interpreter is open source I started in 1998, posted in 1999, and updated in 2001. Scripting Java was hot stuff and FIJI garnered some notice inside and outside Sun. Downloads have continued over the years but nothing like an active FIJI community grew. I stopped working much on FIJI by 2002.
This year I wrote PigIron which deals with mainframe administrative automation. Since this stuff can format the disks, create users, etc., you can imagine it's pretty complex to test in an automated fashion. I turned to Open ObjectRexx plus BSF4Rexx.
Then a funny thing happened. Having returned to serious coding after years of team leading, the simplicity bug bit again. Ditched OpenSolaris with Gnome and KDE. Back to OpenBSD with DWM . And discovered this bug in OpenObjectRexx on OpenBSD that I traced through for days ...
I switched to FIJI for testing PigIron. I started writing in Java in 1997 because it simplified my life, specifically portability. Why was I maintaining someone else's complicated albeit lovely open source tool when I had my own much simpler code, FIJI, for the same tasks?
How good is FIJI anyway? I'd certainly spent many more hours coding Rexx than FIJI. FIJI is certainly simpler, the way Forth is always simpler. You don't spend time declaring stuff and then implementing it and then calling it. You name something. My fingers appreciate the smaller number of key impacts. But is FIJI really all there?
I kept finding gaps in FIJI, dusting off the source, starting to code. Then I'd discover that actually everything was already there. I checked in a few changes, none of which were really necessary. Re-reading the manual that this younger me authored, I remembered how simple it was.
FIJI is weird but solid. It never yet learned to save compiled wordlists but someday I'll fix that. If anyone ever needs it.