making testing a higher priority

[Tom Lord]

A while back, I started to write unit tests for every
part of Systas Scheme.  I didn't get very far before finding
several bugs;  I wouldn't be surprised if Guile (from which
Systas branched) isn't in much better shape.

(I stopped this particular project when unit tests revealed
GC bugs, causing me to think the entire implementation had
to be rethought.)

What about the idea of stopping work on new language features and
performance improvements, and going all-out to make a really thorough
test suite?

Once the test suite is done, other kinds of work are likely to go
much more smoothly.

-t