MSER001@ECNCDC.BITNET (05/23/89)
Here is a good one for all people that dislike that mac... I wrote a program that would send packets of information from the mac to an IBM and visa versa to send files back and forth from my ibm to my mac.(i know..use kermit you idiot! But you never learn that way...) One of the options was to translate all cr from the mac to a cr/lf on the ibm (and visa versa). My first attempt was pretty bogus. I first tried a read/write one character at a time. My...talk about a bottleneck. It must have taken the poor mac about 15 seconds or so to write a 5k block of information to my (at the time a slow 71ms miniscribe), but less than a second with a block read/write. Needless to say..my throughput must have been <100bpi, and the block read/write bosted it up to around 6900bps or so. (se/30 is >9600 bps [usefull bps]). Now of course I was using a small buffer, but I didnt want to tax the poor IBM now did I? Small packets == lower usefull... My point is, since I was reading/writeing binary, why was the thing so damn slow? An Apple IIplus reads a TEXT file faster than that it is using DOS 3.3, and probably faster with DOS 3.2! Has anyone done checked out how long it takes to read a character at a time using Prodos, Dos 3.3, GS/OS etc with different file types? I think if this were checked out a bit more, it would be the death of the mac. Maybe we should all send letters to the Apple ceo and tell him our Apple IIplus is faster than the Macintosh line...how come? Rambling...srh (and they call that progress)