kurt@tc.fluke.COM (Kurt Guntheroth) (03/15/90)
>> Does AmyOs multitask as well as Multifinder? Are most Amy programs >> pretty reliable when multi-tasking? > Apple's MultiFinder is a joke, basically > because the hardware itself is still not capable of true multitasking. That multitasking is impractical for personal computers is another one of the great urban myths. I hear it so frequently from computer people who ought to know better that I have to wonder what they were doing in school. 1. Hardware X is too primitive to support multitasking You can multitask on the Z80. I believe there was even a unix (the old system 3 flavor) that ran on a fancy Z80 board. There's OS/9 for the 6809, a multitasking, multiuser operating system with replaceable device drivers and other good stuff. If you don't believe it, go down to your Radio Shack store and see it running on a $200 color computer. Then say any 68000 system is "too primitive". 2. Multitasking is too expensive; it sucks too many cycles. The cost of multitasking is typically less than 1% (except on IBM mainframes, but that's a different flame), and can be made arbitrarily smaller by lengthening the time slices. Furthermore, a multitasking system usually runs FASTER than a single thread one, because it is more capable of exploiting the inherent parallelism in systems containing a printer with its own processor, disk drive with its own processor, DMA channels, I/O ports with hardware queues, etc. 3. The new Mac finder is just like multitasking, so're the TSR utilities in DOS. You can print and type at the same time, so there's no big deal. You wouldn't say that if you ever had to WRITE a TSR. On a multitasking OS, all the interlocks and time slicing are hidden. You don't have to make a system call to give up control. You don't have to think about how much processing you can do. All programs multitask without any effort on the programmers' part. In the long run this makes programs on a multitasking OS more reliable and compatible. It also makes them cheaper to write and therefore cheaper to buy. 4. Multitasking is no big deal. I don't need it. Maybe not, but why then is EVERYBODY trying to add multitasking to their single-tasking DOS? How many desk accessories or TSR programs do you run? Ever have two that wouldn't work together? Wasn't it a hassle?
king@motcid.UUCP (Steven King) (04/08/90)
In article <16965@well.sf.ca.us> farren@well.sf.ca.us (Mike Farren) writes: >kurt@tc.fluke.COM (Kurt Guntheroth) writes: >>>1. Hardware X is too primitive to support multitasking >> You can multitask on the Z80. I believe there was even a unix (the >> old system 3 flavor) that ran on a fancy Z80 board. > >I was part of a team that designed a multitasking, multiuser system in >1977 - it used 8085's, the slightly snazzier version of the 8080. Worked >like a charm. Multitasking - it's not just a good idea, it's an OLD good >idea. My current job at Motorola is to maintain and update the operating system and processor fault management on a cellular telephone switch. This system (designed back around '77 as well) consists of six multitasking Z80s tied together in a multiprocessing environment. Each Z80 has support hardware to give it an effective address space of 16 Megabytes. Don't ever let 'em tell you a processor is "too primitive" for *ANYTHING*! :-) -- ---------------------------------------------------+--------------------------- It's only impossible until it's done. | Steve King (708) 991-8056 | ...uunet!motcid!king | ...ddsw1!palnet!stevek