We do, we always have. Our Engineering people knock themselves silly
making the code as small as they can, and they're good at it - but
eventually, as the demand for various features makes it a diminishing
return to hold off on implementing them, the code grows. OSPF is a
very good example of both these points - we got to a point where
demand pretty much required us to include it, and when we did we
optimized the hell out of it. Our OSPF implementation is -tiny-
compared to most others out there.
And our code IS modular - you can turn off things like OSPF, reboot
the box and run without those bits of the ComOS loaded into RAM.
> You are about to turn every 1 MB PM-2E you ever sold into a dinosaur
Every product eventually reaches this point. It's inevitable. We
work hard to minimize the rate at which it happens, but eventually,
progress will overcome the 1-meg units - hell, look at the long term,
eventually progress will overcome the PM2s, the PM3s, the human race,
the planet Earth.
Nothing is forever.
> While it may not be much work for a small customer to choose option 1,
> for a big customer both option 1 and 2 are enough of a pain to force
> serious consideration of option 3, and anytime you force a customer to
> contemplate option 3 you risk losing the business.
We know, and backward compatibility vs. new features is a balancing
act. We take a certain amount of pride in our adeptness at that
balancing act - personally, I think we do it better than any other
outfit in the industry. But there comes a time, as I said above, when
you hit diminishing returns by holding back.
--G.
-- 'Twas brillig, and the slithy toves | Benjamin D. Hutchins, texaport minion Did gyre and gimble in the wabe: | Livingston Enterprises, Inc. All mimsy were the borogoves, | BellNet 800-458-9966 FAX 510-737-2110 And the mome raths outgrabe. -><- | http://www.livingston.com/ S.I.G.