forrest@ux1.lbl.gov (Jon Forrest) (08/21/88)
I just received version 2.3 of the MKS Toolkit. I immediatly noticed that certain commands were having trouble finding files that both DOS and version 2.2 of MKS had no trouble with. I called MKS and found that this is a known bug and that they would send me a new set of floppies right away. I'm both pleased and bothered by this. First, I'm pleased with the excellent treatment I was given when I called. I received similar treatment the last time I called in with a bug. What bothers me is that the problems show themselves so easily (the command "ls" results in an error message) and are so fundemental that it's hard to imagine how a new release of software got out the door without this being tested. Maybe they didn't test the toolkit under DOS 2.X. If so, then they shouldn't say it works on DOS 2.X. In general I think the Toolkit is an excellent product and I would recommend it to anyone. I'm just a little worried about MKS's software engineering. Jon Forrest FORREST@LBL.GOV ucbvax!lbl-csam!ux1!forrest
brad@looking.UUCP (Brad Templeton) (08/22/88)
In general I have a lot of sympathy with this sort of thing. These days there are so many hardware and software configurations out there, that it is effectively impossible to test everything in every situation. A company like Lotus has to spend millions and millions on it just to do a halfway decent job. In the case of the MKS toolkit, I think it's one of the worst cases imaginable. Operating system dependent stuff is messy. They would have to test it not just on every version of PC-DOS, (there are now 7 from 2.0 to 4.0) on variety of machines, but also the various MS-DOS versions, and systems like VP/ix, DOS/Merge etc. On top of that there are all the networking systems like Novell. That's for file manipulation software. If you start getting direct hardware manipulation, you have to start multiplying this through hardware configurations. PC, AT, 386, clone, PS/2. With and without 8087, 80287 and 80387. With all the different video cards. Do you realize how much it costs just to OWN the hardware to do proper testing? And then, once the testing gets this complex, you have to worry about the fact that there will be mistakes in the testing. For a small to medium sized software house, the general practice is to do extensive testing on common modern equipment, such as an AT or 386 with DOS 3.3, and then to do basic testing on the variants. It's all that you can truly afford. You have to spend some time coding! And even after all this, you still get what is known as the last minute bug. Something you put in when doing a bug fix on something found during the test phase. As much as you *should* only the very rich can afford to restart all testing from then. When you make last minute fixes, you do them with care, but.. (Now I have no knowledge of the nature of MKS's problem, this is just speculation. And while the guys at MKS are friends of mine, this is really just a description of a software developer's experience.) -- Brad Templeton, Looking Glass Software Ltd. -- Waterloo, Ontario 519/884-7473
bill@carpet.WLK.COM (Bill Kennedy) (08/22/88)
In article <798@helios.ee.lbl.gov> forrest@ux1.lbl.gov (Jon Forrest) writes: >I just received version 2.3 of the MKS Toolkit. I immediatly >noticed that certain commands were having trouble finding files >that both DOS and version 2.2 of MKS had no trouble with. I >called MKS and found that this is a known bug and that they >would send me a new set of floppies right away. [ concern for MKS' testing, DOS 2.X compatibility ] >door without this being tested. Maybe they didn't test the toolkit >under DOS 2.X. If so, then they shouldn't say it works on DOS 2.X. A year ago I had a similar experience, I think the Toolkit was 2.1 at the time. The situation was greatly aggravated by trying to run MKS and UULINK at the same time, I forget which version of 2.X DOS I had. Both Vortex (UULINK) and MKS (Toolkit) urged me to upgrade to 3.X but I was furious with each of them because they claimed 2.X compatibility. Lobbing grenade after grenade at each of the vendors produced the firm but polite reaction, "try 3.1, what do you have to lose?". OK, I had another system that was already running 3.1 so I installed MKS and UULINK on them and behold! They each worked flawlessly, separately and in concert. In retrospect my reluctance to upgrade was just plain old laziness. I never felt the need to upgrade because the 3.X features were of little interest to me. I overlooked the benefits of improvements (environment space, fixes and changes to COMMAND.COM, etc.). I still don't use the 3.X features but I wasted no time upgrading to get the benefits. Also in retrospect both Vortex and MKS were more patient with me than I deserved. Was I reasonable to expect them to back up three revisions from the current release? Should I have expected them to exhaustively test an OS version they never run any more? None of the above should be construed as a flame for Jon, but rather a (hopefully plausible) explanation of how our expectations can differ from the results we achieve when we don't stay reasonably up-to-date with the rest of the marketplace. I'll bet a quick trip to 3.X would mend not only the MKS Toolkit but also some other things that have been nagging nuisances. -- Bill Kennedy Internet: bill@ssbn.WLK.COM Usenet: { killer | att | rutgers | uunet!bigtex }!ssbn!bill
forrest@ux1.lbl.gov (Jon Forrest) (08/22/88)
In response to Bill Kennedy's response (I hope this doesn't go too far), I'd love to run DOS 3.X but I have one of those Leading Edge 'M' systems that was summarized so well in a recent posting to comp.sys.ibm.pc. The upgrade path to DOS 3.x isn't clear and I don't have a lot of time to fool around. I have absolutely no problem with any company stating that their product doesn't work with DOS 2.X . But, given the fact that MKS even goes so far as to say on their distribution floppies that the toolkit works on DOS 2.X, I expect the same level of care to go into testing in the DOS 2.X environment as in any other supported environment. Jon Forrest FORREST@LBL.GOV
smvorkoetter@watmum.waterloo.edu (Stefan M. Vorkoetter) (08/23/88)
Personally, I do not have the MKS Toolkit, but even if I did, I could not upgrade to DOS 3.X or 4.0. PC-DOS 2.10 is the ONLY version of DOS that will run on my machine. Stefan Vorkoetter University of Waterloo
rengland@mntgfx.mentor.com (Richard England) (08/26/88)
From article <798@helios.ee.lbl.gov>, by forrest@ux1.lbl.gov (Jon Forrest): > I just received version 2.3 of the MKS Toolkit. I immediatly > noticed that certain commands were having trouble finding files > that both DOS and version 2.2 of MKS had no trouble with. I > called MKS and found that this is a known bug and that they > would send me a new set of floppies right away. > > I'm both pleased and bothered by this. First, I'm pleased with > the excellent treatment I was given when I called. I received > similar treatment the last time I called in with a bug. What > bothers me is that the problems show themselves so easily > (the command "ls" results in an error message) and are so fundemental > that it's hard to imagine how a new release of software got out the > door without this being tested. Maybe they didn't test the toolkit > under DOS 2.X. If so, then they shouldn't say it works on DOS 2.X. > > In general I think the Toolkit is an excellent product and I would > recommend it to anyone. I'm just a little worried about MKS's software > engineering. > > Jon Forrest > FORREST@LBL.GOV > ucbvax!lbl-csam!ux1!forrest I am interested in finding out what your actual problem was. When and how did it manifest itself. I have been useing 2.3 for about 5 weeks and haven't seen any problems! Perhaps my use patterns don't match yours but I'd certainly appreciate knowing what you found. Thanks, Rich England -- --- rengland@mntgfx.MENTOR.COM ...!tektronix!sequent!mntgfx!rengland ---- | "...Cock o' the North? Ach! Yon's a cheesey | These are all my thoughts... | tune. Weel no be playin' thaat tune!" | I wouldn't wish them on | Alec Guiness, "Tunes of Glory" | anyone else!