ACB.TYM%office-2@sri-unix.UUCP (01/17/84)
I have tried to hold off, but I can resist no longer. Jerry tries to share his experience (is that not what this forum is for) and he is dumped on. Such activity keeps me from sharing some of my experiences. I am implementing LRU disk caches in the BIOS and am lamenting over the inability of the BIOS to tell when to flush the cache. It is very unfriendly to leave the directory tracks of an old disk in the cache (the old disk can be kept whole by using a write through cache) when allocating new files on a new disk. Most schemes that are safe (no explicit action required to insure integrity of disks) end up flushing the cache every warm boot because the BIOS cannot distinguish between the first disk selection after warm boot (innocent) and the first disk selection after a change (requires flush). This is ok for long sessions in database applications but not good when in an edit, compile, test loop. I would be interested in the experiences of others but wary of being assumed to be an idiot. Judging from the volume of the reaction to the Function 37 remarks, there are a lot of people out there waiting to pounce an a cause. Long live Function 37! I certainly won't use it until someone explain's Jerry's information. By the way, how silly it is to see 22 lines of mail header just to see some joker ask whether anyone has received BYTE. Darn! I guess I will get lectured about how, If I wanted to spend the time (or how utility "X" on system "Y" will solve the problem) I could write a program to filter the header. If all that garbage were written on a Postal envelope, the mail never would be delivered. I am not interested in the Odyssey of such a foolish letter, only the author.