lauren@vortex.UUCP (Lauren Weinstein) (02/14/85)
In a recent message, Peter suggested that PKSPEEDUP was useful in general for UUCP implementations. I asked him for a clarification on one point, the gist of which I pass along here. His earlier comment should only be considered to apply to UUCP's running on systems that have a nap() function or some other way to do sleeps of less than one second AND which have "fixed" their PKSPEEDUP code as per his submission of a few days ago. In general, no form of PKSPEEDUP should be used if it is going to result in a minimum sleep length of one second. That is, if your system does NOT give you a means to do sleeps of less than one second OR if your version of PKSPEEDUP does not include his fix that makes use of fractional second sleeps, PKSPEEDUP should NOT be used, since it can drastically slow down communications under either of these conditions. --Lauren--