Why not gnat Ada in gcc?

[Robert Dewar]

<<The fact of the matter is that the development was very open, at
least to anyone who had proper Intel paperwork.  HP did the initial
gcc and binutils work; Cygnus refined both and began optimization
work; folks from CERN, Intel, IBM, SGI, SuSE, TurboLinux, and VA
were testing new code regularly.
>>

I am very aware of the process that was followed, but the issue I
raised here was that these community members who "had done the
proper Intel paperwork" were in some cases exchanging GPL'ed
code under a non-disclosure, which is clearly impermissible
under the GPL.

<<Things were done with as much community involvement as the lawyers
would allow.
>>

But the GPL does not allow community involvement to be limited in 
this manner by lawyers :-)