If bandwidth is such an issue, why don't some of you all work out a way
to trade patches over high-latency (somewhere between sneakernet and
UUCP?) links?
Remember, software still got developed even in the days when it could
take several days to propagate a message across the US.
With decent modularization and regression testing, there is no reason
why patches that are developed for one snapshot of a code base would
not be applicable to several more recent versions. (techniques which
increase the abstraction and hence applicability of a patch should be
of general interest to anyone in software -- PHBs included)
A coot like me may wonder why simple hectobyte patches appear to need
such good connectivity with the current tools, but if we're all sitting
on ISDN (which, apparently, has gone from It Still Doesn't Network to
It Serves Data Now over the past, what, 20 years?) or UMTS or some
such, it's not an itch that's going to get scratched.
-Dave
- Re: [silk] FOSS.IN Dave Long
-
