custead@sask.UUCP (Der cuss) (08/15/86)
This discussion has probably gone on too long already, but I just can't resist. > But seriously, the problem is not the Mac os. It is the hardware. > Apple was right to not try to put multi-tasking on the Mac. Why did Apple create Desk Accessories? Why does Apple distribute the switcher? These are kludges to get around the lack of multi-tasking and let you use the mac for more than one thing at a time. In a multi-tasking system these would not exist because they would have no reason to exist. (DAs and the Switcher rely on the voluntary suspension of the running program. Multi-tasking is the INVOLUNTARY suspension of the running program. That is the whole difference, although it is no small thing; it is fundamental to the whole design of the os.) > Without an MMU, multi-tasking is dangerous. This is absolutely correct. It is also irrelevant. I hope you never use a Ramdisk, or a Desk Acessory, or the Switcher! Without an MMU, all of these are exactly as dangerous as multi-tasking: you have more than one piece of code sharing unprotected memory. Program A is exactly as vulnerable to clobbering by Program B running under the switcher as it would be in a multi-tasking environment. Multi-tasking would not increase (or decrease) the frequency of bombs. Apple was WRONG to not try to put multi-tasking on the Mac. L. R. Custead Univ of Saskatchewan ihnp4!custead
ngg@bridge2.UUCP (Norman Goodger) (08/27/86)
> This discussion has probably gone on too long already, but I > just can't resist. > Why did Apple create Desk Accessories? > Why does Apple distribute the switcher? > These are kludges to get around the lack of multi-tasking and let > you use the mac for more than one thing at a time. > In a multi-tasking system these would not exist because they would > have no reason to exist. > (DAs and the Switcher rely on the voluntary suspension of the running > program. Multi-tasking is the INVOLUNTARY suspension of the running > program. That is the whole difference, although it is no small thing; > it is fundamental to the whole design of the os.) > > Without an MMU, multi-tasking is dangerous. > Multi-tasking would not increase (or decrease) the frequency of bombs. > Apple was WRONG to not try to put multi-tasking on the Mac. > > L. R. Custead > Univ of Saskatchewan > ihnp4!custead Did anybody ever stop to consider that Apple *NEVER* considered to make the Mac a multi-tasking Machine. back in "84" when the Mac first came out no one ever thought of the Mac being a multi-tasking Machine, the Lisa in a way was, but it was incredibly slow, you think that Mac was slow try playing with a Lisa for a while, it will show you what slow is all about. It seems to me that Jobs whole intention from the outset was to create a simple and easy to use computer for all, and Apple has and is taking it way beyond Jobs initial vision, which is great, the next Mac whatever it may become, with a 68020 and probably the 68881 and possibly an MMU will become the machine you have all been waiting for. So save your pennies...its coming... Norm Goodger @ Bridge Comm. Sysop-MacInfo BBS @415-795-8862
tim@hoptoad.uucp (Tim Maroney) (09/03/86)
About memory management, it seems certain to me that the Open Mac will
contain an MC68851 Paged Memory Managment Unit chip. The Programmer's
Workshop documentation contains complete support for and a number of
references to this chip, as well as the MC68881 floating-point chip. The
thing to remember is that this is an unreleased chip, with release
tentatively scheduled for last quarter this year or first quarter next.
Motorola is not giving even tentative pricing information on it yet. Given
this, the only way for Apple to have enough of the chips to have provided
full support in MPW for them is to have made a fairly firm agreement with
motorola to purchase a shit-load of them; and it is unlikely they would have
released documentation on communicating with the chip unless they expected
it to be of wide interest in the Mac community.
I haven't gotten an MC68851 manual yet, but I hope Motorola did better than
the MC68451 segmented memory managment unit chip. The functionality of the
chip was great, but it took 200 ns to translate a logical address to a
physical one! Can you say "wait states"? I knew you could....
--
Tim Maroney, Electronic Village Idiot
{ihnp4,sun,well,ptsfa,lll-crg,frog}!hoptoad!tim (uucp)
hoptoad!tim@lll-crg (arpa)
These are the official opinions of the Vatican Council of Bishops.DMB@PSUVMA.BITNET (09/12/86)
Multitasking on the mac:
Although possible the 68000 is too slow for context switching, to really
make multitasking fast. Second DMA really doesn't exist on the mac
thus the use of context switching when an I/O op is needed can't be
done. Thus you would gain no speed up of overall programming in fact
the overhead would make it miserably slow. With the 68020 and some
intelligent scheme for I/o processes, multitasking could be worthwhile.
-----------------------------------------------------------------------
Then again, I'll probably eat my words when servant comes out.
dave
I think, therefore I progrAM.