CompuServe Thread

#Hard errors?

18 messages in this thread
#110757From: Andy FinkelFeb 24, 1988 5:12 PM
Actually, we do care about Workbench. Copy is single threaded because there's no better way for a new user to thrash (and, under V1.1) trash a file system then allowing copies and deletes to happen at the same time. Remember that there really is no firm indication of directory structure from Workbench. So if a Workbench user launches a recursive copy, then a delete of one of the files in the hash chain… Workbench is a single application. Its not always desirable to put multitasking within an application. Look at the early Textcraft, with its seperate tasks to handle input, display, and scrolling. Multitasking should be used judiciously within a program. Now that I've said that; there are places in Workbench where use of seperate tasks would do wonders. A task dedicated to refreshing the windows, for instance. Workbench is coming out of the ROMs for V1.4, to make room. We haven't made the 512K decision yet (the one that forces A1000 owners to use roms as well) so something had to give, and it turned out to be Workbench. Once out of the ROMs, Workbench will be a bit freer to stretch a little, so to speak, and can gain functions other than just bug fixes. A handy tip: when using Workbench to duplicate drawers (or whatever), move the thing being duplicated onto the backdrop window while duplicating. This will speed the operation greatly, and cut way down on memory usage. Ben: pretend I didn't say that stuff up there about 1.4, and just mentioned the handy tip (grin)
#110808From: greg gossFeb 24, 1988 8:38 PM
Are you gonna fix the bug in INFO where INFO refuses to operate on a remote icon? Project icons can be used to start programs from another directory, with the annoyance that you cannot use INFO without building a dummy data file for it first. Then you have to delete that file again so it doesn't clog the slow directory search after you're done with info.
#110839From: Andy FinkelFeb 24, 1988 11:05 PM
Yes, I fixed the INFO thing. It was intended to do that, btw. I broke its testing mechanism. It wanted a lock, in case you play with the protection bits.
#110809From: greg gossFeb 24, 1988 8:38 PM
Workbench wish list: How about a selection, either a menu toggle or a PREFERENCES setting to create a window with all the workbench stuff in it. There are very many programs that take over the entire bench screen, but get along well with WINDOWS on the workbench screen. Examples are PREFERENCES or ATERM.
#110841From: Andy FinkelFeb 24, 1988 11:10 PM
For the advanced user, of course. That would work for the experienced user. andy
#111333From: Jim VentolaFeb 27, 1988 4:51 PM
Andy– Since I sorta started this long thread, I would like to return to my original question: what are the reasons why it is possible to eject a disk while it is being written to? It would seem to me that if we do a cost/benefit analysis, far more data is lost from folks ejecting disks accidentally than from their not being able to abort disk initializations. Is it the cost of implementing this protection in the hardware or software considerations that make the current situation. I absolutely do not buy the "multitasking" argument as a reason for allowing a disk to be written to and ejected at the same nanosecond. // Jim
#111372From: John DraperFeb 27, 1988 10:06 PM
Jim, The point is not that you are able to eject the disk at the same nanosecond it is being written to that makes the multitasking argument, but the fact that it is a software considereation. I _should_ be able to pull a disk that is between actual writes, even though a file is in the process of being written to. If ever CBM offers drives with only software eject, i certainly won't buy them, nor will anyone else who is past the beginner stage. Regards, Larry.
#111550From: Jim VentolaFeb 29, 1988 12:47 AM
Larry- I don't think I am talking about software ejects. Maybe I am though. What I had in mind was something within the disk (having nothing to do with the computer) which would prevent a disk from being ejected while it is being written to. Of course, the drive would have to _know_ it was writing to a disk, so in that sense I guess software is involved. BTW–I don't see this as a beginner/expert thing at all. Anyone can forget to look at a light. // Jim
#111555From: John DraperFeb 29, 1988 1:20 AM
Ahh… but the mark of a beginner is that he does forget… the non-beginner doesn't. It is because it's so important that the learning separates them. Regards, Larry.
#111590From: Douglas MartinFeb 29, 1988 5:44 PM
I don't accept that absolutely; the mark of a beginner is that he forgets MORE. I take it you are careful never to go near your Amiga when you're tired; you make sure you have your TV and radio off (probably a good idea to unplug your phone too); and you have no other distractions to make sure you absolutely will not absent-mindedly do something you shouldn't?! My point there is that I don't believe anyone ever is 100% perfect about doing any reasonably involved task over an extended period of time. It could be argued that software ejects on a multi-tasking system are tacky (I'm not absolutely convinced, but I won't try to argue); however I fail to see any good argument against software locking drives against ejection (assuming fail-safes ala paper-clipping). There are cases that not being able to get at a disk FAST would cause problems, just as wearing seat-belts can cause injuries. It's just that it seems to me that the trade off works the other way (there's always the Vulcan Nerve Pinch and the switch on my power bar). Another point – we all want increased sales of the Amiga. On that basis we can't argue against putting them where lots of novices can get at them (schools, universities, a great many offices, etc.). Unless you want one Amiga for idiots and one Amiga for non-idiots consideration should be given to a compromise between driving experienced users up that wall with excessive hand-holding and letting naive users destroy a few dozen disks before they get procedures straight. While I'd love to tell many of our customers what I think of their mentality for some of the things they do, the fact that people screw up and expect other people to unscrew up their problems is a fact of life.
#111628From: John DraperFeb 29, 1988 9:47 PM
Douglas, The one thing I _don't_ do any more, tired or not, is pull my disks out or reset the machine early. Regards, Larry.
#111767From: Jim VentolaMar 1, 1988 10:12 PM
Larry– Since this is one area where I do not want to learn by experience, what is this "resetting the machine early" I am hearing about. BTW, I agree with what Douglas Martin said, but this thread is verging on religion, I think. Making Amigas safe is like making anything safe: a distasteful and occasionally necessary thing. // Jim
#111668From: Don Curtis/SYSOPMar 1, 1988 1:17 AM
Douglas, NO user should be turned loose on any computer until they have at least rudimentary training…about the first thing that a teacher should tell a student looking at a computer is to never turn the computer off or take a disk out of a drive or reset the computer while the disk light is on. This certainly isn't Amiga specific….try using any computer while it's writing to a disk and shut it off…bye bye disk. Don
#111633From: Andy FinkelFeb 29, 1988 10:03 PM
There are/were at least 3 reasons; depending on who you are, one of the reasons may be the one that convinced you… 1) Cost 2) Multitasking arguments (no need to repeat them again) 3) dislike of the way it worked on a Mac, trying to use the compiler available way back then (128K Mac as well) 4) religious feelings that the user should be in control at all times 5) extensive use of the paper-clip when doing 3) above I once had the task on attempting to trash a disk using the standard programs on Workbench, and pulling the disk out at odd intervals, while writing, etc. Aside from seeing the "You Must Replace…" requester, and spending a lot of time revalidating, I didn't lose many disks that day. Do you guys think its really that common ? My informal polls of new users don't indicate that many people really trash many disks that way. andy
#111679From: Don Curtis/SYSOPMar 1, 1988 1:54 AM
Andy, Using the original OS (1.0) where you got 2 distinct disk writes with the LED going out between them…then quite a few folks did lose disks by re-setting the machine too soon. I lost one that way myself, but only 1 because I have a tendency to learn from my mistakes. With the current OS, I don't believe it's a problem any more as the LED stays on all the time between the 2 writes and the fact that I haven't near the volumn of messages about trashing disks that we saw back in the 1.0 days tells me that 'cured' alot of folks too quick on the reset keys. Don
#111765From: Jim VentolaMar 1, 1988 10:08 PM
Don- From Workbench after a RENAME there is a noticeable delay and one could easily, I imagine, pull a disk just as the light decides to come on. // Jim
#111763From: Jim VentolaMar 1, 1988 10:06 PM
I think I have trashed three disks that way. The first was from dnot knowing, the next two were the sort of things where the nanosecond you do something you regret it. Two that I know of have been trashed at the TV studio at work. Let's see, that makes five… It may not be worth changing any hardware over, I guess. Why *not* fly in an airplane with an ordinary door, eh? // Jim
#111787From: Andy FinkelMar 1, 1988 11:56 PM
All by the 'pulling out too soon' method ? Do you remember the error (key already set, key already free, key out of range, or just not a DOS disk) ? andy