Tony Wachs TW stories wanted 1286

Actually it was easily in reach from the CTY. ;-)

Mine was MTXSER. I finally persuaded the powers that be that I should do a completely new tape service routine, or rather set of routines ala FILSER. I had the basic design work done when I got that "offer I couldn't refuse" from First Data.

One of the design points I was stuck on was how to initiate an I-O operation from both UUO level and interrupt level without interference. The kludgie solution would have been to queue up the I-O request and then force an artificial tape level interrupt, where the queue could be scanned and any pending operation initiated. I mentally rejected this an inelegant, and was working on an interlock design when I left DEC, the design incomplete.

When the new TAPSER was released (as completed by other people) I quickly read the code to see how they had solved the interlock problem. I was disappointed to discover that they had implemented the "artificial interrupt" solution. :-(

Actually I'm probably most embarrbutted by my file system design for TSS-8. 201 (129 decimal) word blocks! What WAS I thinking????