Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!samsung!olivea!mintaka!geech.gnu.ai.mit.edu!rjc From: rjc@geech.gnu.ai.mit.edu (Ray Cromwell) Newsgroups: comp.sys.amiga.hardware Subject: Re: Amiga Custom Chips - why hasn't C= made them faster? Summary: Marc fails to see the big picture, again! Message-ID: <1991Apr3.130218.25163@mintaka.lcs.mit.edu> Date: 3 Apr 91 13:02:18 GMT References: <1991Apr2.235710.13984@news.iastate.edu> Sender: news@mintaka.lcs.mit.edu Organization: The Internet Lines: 71 I'm not going comment on your 'charts' because I have no idea what the units on your numbers mean. And in the past you have demonstrated your inability to read market statistics, so for all I know, those numbers could be the number of pizzas C= execs ordered last year. You still have no idea what your talking about and you fail to see the big picture. Sun innovated SPARC. What have they done since then? Refined it. Amiga innovated the custom chips, they have been refined since then. You don't know what it takes to build a 32bit faster chip set, your clueless. Consider the ramifications: The Amiga is mainly a 16bit machine except for the A3000, so what about all those other models out there? The majority of the machines are A500s. Redesigning the custom chips means breaking most of all the software out there. It also means abandoning the low-end machines. Do we really want to do that? Now let's consider this question Marc. Why hasn't MS-DOS been significantly improved? The answer is, because of its design limitations any 'quantum leap' or major improvements will break BILLIONS of dollars of software. The Amiga, by design, uses a shared memory model to achieve its speed. This rules out memory protection. Also by design, resource tracking ala UNIX is impossible since Amiga programs are allowed to allocate a resource and pass it to another task, then exit cleanly. Sure, code can be added to allow programs to request tracking of certain resources, but this won't do a bit of good to poorly written programs that haven't been updated. And back to memory protection, about 80% (probably higher) of Amigas don't have MMU's, so how the hell are you going to provide memory protection. Are you suggesting Commodore release a different version of the OS for MMU machines? WHat about software that has been following Commodore rules for years where passing a memory pointer to another task is legal? Every time you press a key on your keyboard this is happening. If Amiga dropped it's memory sharing message system it would be dog slow. Resource tracking on 68000s would eat up huge amounts of ram and be dog slow. Memory protection is impossible without an MMU and so is Virtual Memory(UNIX flavor). Device independence would be good, but on 68000 machines it would slow them down. You see Marc, you fail to consider the big picture and how hard these problems are to solve. Throwing money at them doesn't help. The IBM Market beeen trying to solve the MS-DOS problem for years throwing more R&D (total market wise) than Commodore's total net worth. Do you think if we invested 100 billion dollars in AIDS research we could cure it in a month? These problems aren't impossible, but they are difficult. Somewhere along the line a trade off is to be made. Do we abadon the 16bit Amigas? Remember how many A1000 owners got hurt when it got abadoned? Do we sacrifice the Amiga's performance so it can look like UNIX and run slower? Shall we break thousands of software titles to solve a problem that can be solved by good programming and debugging? If you think these problems are easy, when don't you write/draw up your new custom chip designs and send them to Commodore? Why don't you post your solutions to the Memory Protection/Resource Tracking/Device Independence problems? If they're viable, you might(I hope not) have yourself a job working at Commodore. -- /~\_______________________________________________________________________/~\ |n| rjc@albert.ai.mit.edu Amiga, the computer for the creative mind. |n| |~| .-. .-. |~| |_|________________________________| |_| |________________________________|_|