Jeffrey Hobbs wrote in the comp.lang.tcl newsgroup on 2002-02-15:
So how can people help out? Well, I have many partly formed ideas that I would be glad to pass on to interested parties. It would be nice if those in academia could interest a student or three to work on interesting projects in the core. I guarantee you I can find projects of interesting enough research value for their time.
Do you like compilers, GUI, IO, algorithms, automata theory, OS internals, or ??? I guarantee you, you name me a preference, and I'll return at least 3 projects or merit. Let me just give some morsels:
Compiler: (a specialty of miguel sofer at the moment)
* Simple post-pass optimization (making things like ''if 0'' completely disappear) * Byte-compiling of more commands (switch, l*, ...) - Jazz it up by better reusing code between eval'ed and bc'ed variants (this isn't so great now) * Reorganization of TclExecuteByteCode for enhanced performance (this really will require thinking)
GUI: oh, this list would go on forever
* themeing * enhanced native platform capability * work on TkGS * more widgets * a truly efficient megawidget system * improve embedding / embedability * ...
IO:
* design / write cool VFS extensions * special path optimizations for simple IO cases * core UDP support (that means fileevent and all that works too)
algorithms: this is basically looking at whatever algs Tcl uses, and seeing what might get improved
* hash table performance * UTF <> Unicode handling * ...
automata theory:
* this would be the regexp engine (maniacal laugh)
OS internals:
* loading DLLs from memory * core DLL function caller (x-platform) * threading tweaks * improved allocators
I could get into fine detail on any of the points. Look at all those fun projects! What I would give to be a student again. Find a couple bored friends. Have a semester long hack fest.