Lee Daniel Crocker wrote:
Sorry, but I have to make a big-picture comment. The
phase 3 software
was a quick hack, and reasonably effective at getting us over the
bandwidth hump until we could afford to throw more hardware at the
problem. But my impression is that it is still 10-20 times less
efficient than it could be if done right. Good software takes time to
design and build. Might it not be wise to consider starting now to
invest in some development work?
Yes, probably so.
By going out and negotiating all these deals for hardware, I have put us
in a position where the money that we have raised can hopefully be used
for something else. Instead of thinking of this as an either/or "either
more machines *or* development work" we are now in the happy situation
of "more machines are coming... free" and so we can have *both*.
I'm a little skeptical about the 10-20 times less efficient claim. But
there's surely room for improvement; there always is. Someone who is
qualified to say can comment on how much improvement there might be to
the code.
--Jimbo