Re: Let's do some coding :-)

[Jonathan S. Shapiro]

I have hesitated to suggest this after Linus's most recent faux pas, but
I think that there may be something more useful to do than coding.

First: the Hurd community must come to a decision about what the design
principles of the system will be. Is "real time" important? Is avoiding
DoR important? Is support for ACLs important given that they don't
actually work and interact badly with capabilities? At what level?
Should persistence be part of the system design?

Each of these decisions will significantly restrict the feasible design
space in certain key areas. They will determine which designs are
feasible and which can be discarded because the principles are violated.

After this, the next step is to design protocols using IDL. The IDL
should include both the statement of the interfaces and their
documentation (this is the hidden power of CapIDL -- it does something
very similar to doxygen). In particular, the low-level foundations of
the system need to have specified 

These interfaces will serve to define the contracts between various
developers. Once these contracts exist, productive coding can begin.

shap