<joshcough> sure. and, anything in a binding that's free is just
a free var. but, how will the type checker know where
to find it
<joshcough> it seems like the type checker will have to redo some
of the work that findGlobals is doing
<joshcough> maybe its not that big of a deal.
<joshcough> s11001001: i guess i'll just close 15. it seems to be
working and tested for terms and types.
<S11001001> joshcough: Actually, this is a good point.
<S11001001> joshcough: Totally, 15 is done.
<S11001001> joshcough: But now maybe we should consider dataflow.
<joshcough> yeah it seems like we are throwing data away here
<S11001001> joshcough: basically findGlobals computes NAMEBLOB,
an index of types and terms for each imported module
<S11001001> joshcough: it uses that only to compute explicits
<S11001001> joshcough: but we can use that for linking names
while parsing statements
<S11001001> joshcough: so maybe we should separate the phases
<S11001001> joshcough: findGlobals can *also* filter the NAMEBLOB
so that the statements only see the names that should
be imported
