OK, what about some resolution (Re: GUILE's GC - why we struggling to solve already solved problems?)

[Michael Livshin <mlivshin at bigfoot dot com>]

here's what we have, in no particular order:

* current Guile GC sucks.

* it can be made not to suck, but it's a lot of work to get where.

* BGC is probably more efficient than the current GC if integrated
  nicely.

* BGC picks GC roots from more places.  some say it's a good thing,
  others (like myself) disagree.

* if we sink enough work into Guile's native GC, we can make it better
  than BGC (I think).

* it's not that hard to provide both.  when we move marks out of
  cells, all that remains to abstract is the user-callable SCM_MARKP
  macro, which can just expand to a function.

* note that BGC can't be just used as-is as a part of Guile -- it
  should be made ./configure'able, etc.  I suppose the GNU Java folks
  already did this, though.

comments?

-- 
"Tools that are no good require more skill."