Win (or lose) 3.10 (Was: Kaypro PC ? == Micro 1

On Sat, 8 Sep 2012, Liam Proven wrote: Hence the quotation marks around "solved". It was a ludicrous claim. Careful would not protect you if you had a disk write error during delayed writes (you were slightly lucky), but it WOULD NOT protect you from turning off the machine before writes after it came back to the prompt on exiting a program. "I'm DONE! SAVE. EXIT. At the prompt? Hit the power switch!" Agood reason for implementing the "shutdown procedure"! Infoworld printed at least a dozen articles of the form of "Users of new DOS 6 losing data!" Return to prompt before writing buffers was where the InforWorld "test" blew it. Billg tried to tell Infoworld that their testing methodology was flawed, and they misinterpreted that as "intimidation". InfoWorld's "test" of "disk compression problems" consisted of Word Pervert and Lotus running macros, Followed by a reboot, In a loop, until the disk had problems. The ACTUAL cause of the problems was that the reboot was occurring before the write-cached buffers got written to disk. Microsoft correctly insisted that the testing was flawed and did not show disk compression problems; but Microsoft would NOT admit that SMARTDRV was, indeed, damaging the test data, and causing the bad results. Correct. Nope. There was no MS-DOS 6.1 (or 6.10. THAT was PC-DOS 6.10 (with a different compression). The debacle that resulted in free "step-up" from 6.00 to 6.20 was BEFORE the suit. The lawsuit caused the change from 6.20 with compression to 6.21 without compression. Then 6.22 had a different "non-infringing" compression. The judge awarded Stac about $100 million, but awarded MICROS~1 about $30 million in their counter-suit. Billg said, "I'm having a bad day." NO. in 6.22 As to "theft of concept", STAC was NOT the only disk compression product on the market, nor the first. NB: There was never a Six point ONE. Nor a six point TWO. It was Six point TEN, and Six point TWENTY. MOV AH, 30h INT 21h returned 6 in AL and 0Ah or 1.4h in AH Microsoft version numbers are an integer "MAJOR" number, a "full stop", followed by a TWO digit decimal number for "MINOR" number There WAS a DOS "Four point one" (4.01), but never a 4.1 (4.10) You are right in concept, but your version numbers are wrong. 6.00 was the full-blown shopping excursion. Basically 5.00 with lots of bundled stuff (hence "NEW AND IMPROVED"), including the infringing disk compression. 6.10 was IBM's PC-DOS version. It had a different compression 6.20 was "to repair problems with disk compression", such as those reported by Infoworld. The "repair" consisted of reconfiguring SMARTDRV to have 1) write-cacheing OFF by default 2) IF write-cacheing were enabled to not re-arrange writes 3) IF write-cacheing were enabled, to NOT return to the DOS prompt until the buffers had been written. (6.20 may have had some additional trivial bug fixes) 6.21 was the one with disk compression REMOVED 6.22 was the re-release with "non-infringing" disk compression Microsoft would never admit that SMARTDRV had problems, but it was actually the ONLY (or at least PRIMARY) factor! -- Grumpy Ol' Fred cisin at xenosoft.com