On Sun, Oct 14, 2001 at 03:51:39PM +0900, Patrick Shirkey wrote: > Is anyone going to shoot me if I advertise the best best case latency? > Should I put a banner on the LAU-guide that says you can get 0.75 msec > Full duplex latency or should I tell the truth that we can only get 2.6 > msec latency? Actually I'll do both. But before I do is 0.75 msec the > best best case?
Please, please don't do this. Mis-information is bad, and purposely spreading stuff which you know is false is even worse. Think about how you would feel if you were new to Linux, and a friend told you "Yeah it can get 0.75msec latencies!" "Really?" "Yeah!" ... and a month later, after you go through all the processes required to set up a linux box, and find out that you never ever get below 2.6ms, (s)he says to you "Oh, yeah, 0.75ms is the best case, 2.6ms is more like what most people get". "So how do I get 0.75ms?" "Oh, well you have to buy this hardware to do it, and have a processor over 1.5GHz, and you have to run it in single-user mode, and you also can't run X because that floods the PCI bus for too long, and the program that you want has to be optimised for it too." You won't be left with a very nice impression, and you'll start resenting the Linux community as the whole, the bunch of liars that they are. Don't fight false marketing with false marketing. > I have many muso friends who are very interested in getting a Linux > system running. The hard part for them is finding someone who can help > them through the first few baby steps. So help them with the baby steps. Don't tout Linux as the be-all and end-all solution when it's not (yet). -- #ozone/algorithm <[EMAIL PROTECTED]> - trust.in.love.to.save
