Just a quick comment about your article in The Amiga Monitor: you said "rename" has the disadvantage of not working across devices. You didn't mention the very big advantage - that it only changes the name and you don't have to wait for the whole file to be copied, and that if you're moving a file on a floppy disk, then there may not be enough room to copy it and then delete it.
Fergal Daly
True, very true. That is a very notable advantage of the "Rename" command. So, dear readers, you may supplement my original report (see "Mike Webb's investigation of AmigaDOS" from the September 1996 issue) with this information. Thanks for pointing that out, Fergal.
Michael Webb
Editor-in-Chief
In reply to Robin L Clifford's letter published in issue 2 of AM, I think several people are missing the point (most of the Amiga world are missing the point). Providing the low-end Amiga systems with a processor that operates like grease lightning will not help! For the last five years the Amiga's problem has been the design of its custom hardware - it is just not good enough!
The Amiga works in a totally different way to most other architectures. Firstly, it has two independent buses - Chip and Fast. The trouble is that the CPU (any CPU - even a 300Mhz DEC Alpha if you could use one) has the lowest CHIP RAM acceess priority of all major system silicon, because if the display hardware could not get access to the Chip RAM bus, you would not have a display. Add to this that the display hardware is totally planar in design (forget the rubbish about using the copper), which means that for an 8bit display, LISA (I'm talking AA here) has to perform eight memory fetches for every 32 pixels (the same as an 8bit PC display). However, due to 8bit PC displays being CHUNKY, it is possible for the memory access to be pipelined. LISA on the other hand has to change her address eight times per 32 pixels. This in itself doesn't prove to be problematic (even with high resolution) until you also realize that the BLITTER, COPPER, PAULA, DISK, SERIAL_PORT, etc. hardware are all fighting to get on the same bus - if you have not got any Fast RAM you are in real trouble.
Michael Webb states that the ECS is "more of a limitation on performance than the '030", and I agree, but I also hate to shatter everyone's dreams - so is AA. You cannot expect games like DOOM (Amiga clones are a joke!) when it takes eight Chip RAM accesses to write 32 pixels (plus a ridiculous amount of processing in the registers - setting single bits isn't very efficient). OK! You can do it much faster when you hold a chunky buffer in fast RAM, and then convert the whole thing, but it is still much slower than if you had a real chunky display. What I am trying to say is that having a processor running at 1 BIPS would not make a huge difference - especially when the speed of Chip RAM bus is linked directly to the colour clock! PeeCees will always be faster, and will keep on getting faster. Why do you think anyone who uses his A4000 for serious reasons also has a huge 64bit graphics card taking up a fifth of its innards? It is not because he likes 16 Million colours on his Workbench, it is because it provides SPEED.
What the Amiga really needs is a complete revision of its chip set! However, this is not a realistic proposition, as the cost of developing such a chip set would be far too high. There are people that advocate the selling of such chips as a high end graphics cards for PC's, but that market is already dominated by large corporations with huge budgets and very successful (very cheap) products. No! The Amiga must move on. I advocate that the whole graphics and sound system should be made retargetable, with the user selecting the appropriate hardware drivers as they currently do with printers. There are many video and audio hardware manufacturers that can make silicon of a far superior quality (using AAA is a joke) and a much cheaper price than any Amiga company could do, that the Amiga must forget its past. In 1996 there is no need to "bash the metal" (no matter how much I like programming the Amiga hardware!). What the Amiga needs is a scalable PCI architecture (for the whole range), PPC of various performance accross the range, and a significant update to the OS design whilst keeping the Amiga ideal intact. All machines should have audio-video functions performed on separate daughter-boards (with their own memory - preferably VRAM for the high-end machines). High end machines should use very high end graphics (possibly Silicon Graphics) and sound hardware (many to choose from), with the home machines using similar hardware to current multimedia PC's. High end machines should also include the current architecture on a separate daughter-board.
Obviously the high end machines should be released as soon as possible, and sold in parallel with the current A1200 (sold with 4Mb Fast RAM, and 50Mhz 68030). The whole range should then develop over a period of two to four years (I'm being realistic *), with the high end architecture migrating down through a mid-range machine, and eventually to the low end. This would allow software suppliers to migrate their products to the new architecture, and provide the low end versions of the new architecture with a ready built library of software. Obviously, the high end machines MUST keep pace with PC technology over that time scale.
Martin Cordingley (An Englishman in Germany).
Well, where shall I begin...
The Amiga has always had the custom chipset and bus design, as well as planar screenmodes, and it's just a bit odd that people have only begun to question them recently...but more on that later.
Your point stands that the current line of Amigas are performance-deficient, and you begin to hint at the idea that there are a number of reasons for that, but I think some of those ideas (and don't get me wrong; a definite portion of the Amiga community has been adopting these) are counterproductive at best.
For one thing: yes, we DO need faster processors. I reiterate what I said last month: we have to keep everything within reason if we want to have a chance of surviving. And from a technical standpoint, the 680x0 series still has plenty of life in it. That is, a 50mHz 68030 for the low range, and the 66mHz 68060 for the high end. I also reemphasize that it would be a mistake to divide the mainstream line among two processor families.
Now, the chipset is turning into a very heated, controversial issue these days. Let's just establish right off that the reason the Amiga has been so successful as a multimedia computer over the years is that it has had that efficient bus built in, including the custom chipset, which has always been a powerful and modular coprocessing subsystem. If you compare it to PC's, in which the entire platform has been crippled by a fundamentally poor software AND hardware design architecture that has been built, and continues to stand in part atop the flimsy foundation of the MS-DOS days, you see the incredible brilliance of the Amiga's design. Even though it IS dated. Let's just say a haphazard design would NOT have lasted over a decade despite corporate incompetence.
The Amiga has always had planar screenmodes, and there are perfectly good reasons for having them, particularly notable being use with video devices, one of THE reasons the Amiga achieved any level of success. But it is clear that it needs chunky pixel screenmodes.
If you are hanging from a cliff by precious few threads, you don't cut one of them. Similarly, eliminating the current Amiga design architecture in favor of, essentially, 3rd party graphics and sound boards would be the quickest road to technological suicide. Again, one of the Amiga's greatest strengths has always been the custom bus and chipset design. It needs to be improved, not killed off! Not only would switching to a generic PC architecture create more incompatibilities than you could dream of, but it would also accomplish next to nothing.
As far as I am concerned, people who cite cost as a reason for dumping the custom chipset are those who have truly missed the point. The Amiga was an inexpensive computer from day one, often selling for far less than an IBM-PC, and Commodore continued that trend up until their very end. It was only later, when ESCOM took over, that prices truly got out of hand. The custom chipset is NOT a price concern. To say it is is to point the finger out of desperation.
You say that "using AAA is a joke" without offering any support for that supposition. AAA is, incidentally, the next step; it is FAR ahead of AGA, and on a par with nearly any 3rd party board available for the Amiga today. It's also almost complete! Without going into all the technical details, the AAA chipset would significantly increase the bandwidth, speed, efficiency, and overall performance of the chipset. And yes, it WOULD add chunky pixel graphics modes, which would single-handedly solve many of the performance problems associated with the custom chipset now. In conjunction with all of this, it also has the nice side benefit of being backwards-compatible with existing Amiga technology. Furthermore, dare I say it...regardless of the technical merits of the AAA chipset OR something else, the AAA chipset would be the fastest way to get new Amigas with *significantly* enhanced capabilities on the market (and then give time and resources to develop a successor to AAA, as has been the progression for every chipset since the first). And if we want to have ANY chance of surviving as a platform..."soon" is the operative word.
So yes, blaming the Amiga's microprocessor for it's current problems is somewhat shallow...but so is blaming the custom chipset.
Michael Webb
Publisher and Editor-in-Chief
Users seem to be in the habit of mailing wishlists these days. 'I don't want a 030 in the lowend Amiga...' or 'The PowerPC should be able to run windows...' etc. Hmm...I think that TALKING about it is a big waste of time since no one listens. I've been with Amiga since day-one and believe me, they never listen.
Now for some ArmChair quarterbacking of my own...I think the PowerPC plan is a dumb idea.Yeh, I know RISC is the future and all that, BUT:
Keep that in mind! When the first Amigas came out C= opted for 68000 because it was 16-bit, but by getting this we don't improve on Intel at all in performance!
I think the only viable idea to do right now is to port the thing to 586 code and run it on a Intel motherboard. A top of the notch PC costs 1500$ these days. Cheap! If it wasn't for the crummy Windows it would even be usable. By doing this Amiga would at least have a safe future hardware-wise,not to mention the big bust in mHz! But then again , I can trust VIScorp/AT to make the wrong move again...
Ariel Magnum
Tel Aviv, Israel
I tend to believe that speaking your mind is the least people can do to enact change, but that's beside the point.
I agree with you that the PPC isn't a good choice now, but from there we diverge. Now, granted, PPC's are a bit pricey, but I think that $5000 mark was a bit overexaggerated, considering that Macintoshes generally don't cost anywhere near that much.
"Commodore" (we're actually dealing with the original Amiga Inc. here) decided to use the 68000 for a number of reasons: better architecture, far more coherent register-programmability, and the fact that it was 16/32-bit didn't hurt. In addition to that, it was faster than the offering from Intel at the time. Believe it or not, 680x0's have kept up with 80x86's, for the most part, until only recently when Motorola basically ceased development of the line. For instance, a 68020 or 68030 could compare to a 80486, a 68040 to a high-end 80486 or low-end Pentium, and more recently, a 68060 stacks up nicely against a midrange pentium (and nobody's even clock-doubled it yet). By switching to the 80x86 line, which itself might be dying (at least as we know it) at the hands of RISC development, we would gain nothing at this time besides making the Amiga a more generic PC-flavored box.
Lastly, it's lamentable that our experiences with Amiga companies over the years have had the potential to foster such cynicism, but it's difficult to not take on a pessimistic view considering what we've been through. Maybe the future will be better; we can only hope, wait, and see...
Michael Webb
Publisher and Editor-in-Chief
A few points (long points ...):
Motorola's largest 68k series customers have always been industries other than computer manufacturers. Motorola 68000, 020, 030, 040's are found in printers, robots, cars ... heck, your microwave may even have a 68000 -- I know my printer has an 020! Macs and Amigas made up a significant share of 68k chips, but they were not the entire market. Motorola are not even trying to market the 060 as a CPU as such but as a controller chip for industrial uses. I cannot remember the term used, but they are not talking about it in terms of desktop computers.
As an owner of an A3000 with an 060 and a graphics card (CV640) with Cybergraphx, I have learned the hard way that AGA-like hardware standards have to go. You talk of future machines needing chunky-pixel and of gaining back the edge once held in graphics. One way is to embrace Cybergraphx -- the only RTG pretender out there in any numbers. I have a machine which can play some of the most advanced texture-mapped games out there, playing through Mac-emulation on Shapeshifter, but I cannot run The Killing Grounds AB3D II. I can drag windows with 8-24 bit pictures as a background for the windows (with a different 8-24 bit background on Workbench) as a solid window, resizing the window as a solid as well (I don't know the technical term for the real-time resizing). What does this mean? Well, my machine at work is a P-100 running Windows NT and a 64 bit graphics card. It can also move windows around in solid view and resize as a solid window, but you can forget backdrops of any sort if you want the display to be smooth and the machine functions not to grind to a halt. Why? Because the OS gets in the way. The Amiga still has its 'graphics' advantage and many others too once a faster graphics and central processor are added. An 060-based Amiga runs Mac software roughly as fast as a 9500 - 100 based PPC Mac, (excluding FPU).
Not as much has to change for the Amiga as some may think. Phase 5 has the right idea by making the first multi-processor Amiga with their 060/PPC tandem boards. Imagine -- the OS and smaller apps running from the 060 while the PPC answers the calls from ImageFX, Shapeshifter, LightWave, AlienBreed 3D IV :) Now there is a computer to drool over, no? My Amiga has cost me many dollars, maybe around $3000 -- that includes a CV64, 060/50, a 3000 (used), and 32 megs of ram and a lot of great software. Even with its dated OS, my Amiga outclasses machines that I could have purchased new at that price. Maybe not in raw MIP speed of such and such a test, but in every day real use in programming C++, making Web pages and graphics, running rendering, and writing letters. Whether the Amiga goes bust in the end or not, it really does not matter. I still have a great machine that can do fantastic work.
It is not about beating Wintel (good luck!) or becoming number one in everyone's minds, it is about surviving and being able to enjoy the computer of your choice for years to come ... and I stress choice. I don't want to have to choose from IBM's Wintel or Compaq's Wintel. I don't call that choice. I want to have an alternative that is right for me. Macs are great, but I like more control and power, not to mention multi-tasking. Maybe I should go get a Be computer? They sound great and maybe in a few years I will get one, but it is no Amiga. It has an Up-to-date OS alright -- it takes at least 8 megs to start the system up! Yes, yes, I know, it has many great OS features and lots of risc cpus which require this ram. When Be releases a home-user based 1990's 'Amiga 500' equilvalent, then they can call themselves the spiritual successor to the Amiga. I am all for making big-box Amigas, but I for one would like to remember when I saved all the money I could and went out and bought an A500 in '86/87. I think it was $400 and it was not a main-stream computer at that time either. I highly doubt the average 15 year old today can afford a $1500 base computer. Sure, the kid could buy a used 386/486 for a little more than I paid for my 500, but would he get the same level of equivalent power that I had at the time? How about the pride in being able to buy a first computer? No, I am one of the first to say to Be, "great, another choice in computing!" Be developers and Mac users also talk of beating Wintel, and 'winning.' Beating Wintel is not going to happen ... not in this century :) The focus of the Amiga should be survival -- not teetering-on-the-edge survival which has been going on for far too long, but a healthy, strengths-in-certain-markets kind of survival (including the home market). I want the Amiga to have a healthy market share and receive some respect as a great computer platform, but main-stream? I think not.
James Ceraldi
Kingston, Ontario
Canada
It's always interesting to see how microprocessors are being used as simple controller chips in various domestic or industrial applications other than personal computers. Especially since, in many cases, the older chips that we once regarded as so cutting-edge are now being used, and some manufacturers are just beginning to get beyond the old 8-bit Z80's and 6502's and moving on to the 16- or 32-bit units we've been using in our Amigas all these years.
I won't repeat what I said earlier on this page, beyond saying that yes, AGA itself might be dated, but the logical progression would be to the next level of the chipset, that being AAA. But I won't get into that any more at this time, other than to say that CyberGraphx is a very good piece of software, and that the Amiga would do best with something like it built into the OS...something that can address the custom chipset as well as a graphics/video board. That would be versatile. But, fortunately, the OS is very modular (object-oriented for those buzzword-inclined) and easily supports and integrates such extensions.
You make a lot of good points in your letter, however, so leaving well enough alone, I'll let you speak for yourself without any further comment on my part. It nicely embodies the philosophy by which we are endeavoring to succeed in some way.
Well said. I also remind you and the rest of the community to keep the faith, and remember that as long as we have a reason for doing what we're doing, it's not over. And who knows, we may actually have a renewed reason sometime soon.
Michael Webb
Publisher and Editor-in-Chief
Editor's Note: The following is a letter that a reader wrote to Softwood regarding Phase 5 products (it has intentionally been left unedited).
Hello,
My name is Andrew Gordon, I am writing from Sydney,Australia as a purchaser of a Cybervision 64 card + Cyberstorm 060 (still awaiting Cyberstorm 060) from Anahiem,California based Amiga dealer Transdata Intl.(Cybervision $600,Cstorm $1000)
A little history on myself as an Amiga user for over 10yrs: my current system is one of the A4000Ts made by the now defunct Amiga Technologies, purchased only in April '96 after a 3 year break from computing.(my 4000T is fitted out with 16megs ram,quad scsi cdrom at a cost of U.S. $4300).
I felt I needed to write directly to Phase 5 Germany, but found they have no email access on their web page.So I figured I would write to you as my boards were sourced from the U.S.
My concerns surround my visit to Phase 5's German page and reading about the new Cybervision 64/3D update.I ordered my Cybervision board in May on a trip to L.A. only to recieve it this week after it arrived at Transdata from you.
I am very dissapointed to find I have an instantly obsolete board and to see Phase 5 trying to suggest an extortionate trade-in fee for existing owners, crying that the makers of the Trio64, S3 chip have dropped support of theold MPEG expansion option.
As one of many Amiga owner really struggling to justify pricy expansion, which is fast slipping behind PC machines, I am left feeling very cold to third party hardware makers.My chief point of reference has been me having a technician whom works for my company here in Australia build me a Pentium 100 based PC as an interim home computer/eventual work computer. I had some components from a laptop I once owned, such as a ext. 4x cdrom drive, keyboard, and mouse, as well as my 4000Ts 17" multisync. With these parts I asked my technician (a skilled PC Builder) to build me a PC with 16megs (U.S. $120),1.02 gig HD(U.S. $210),a S3 Trio32/64 based 2meg video card (U.S. $140),Soundblaster pro compat. audio (U.S.$100),motherboard + pent 100 (U.S. $400). This is my second day using this PC (I have been using other PCs at work and with friends, particularly helping install internet browers and enjoying them) and already have it running in 16-bit color at 1024x768 mode in Win95. Quite frankly I am finding it difficult to knock its performance and ease use.
In this kind of battle were does a formly anti-msdos/Win3.11 Amiga lover stand?
By comparision I have had dramas akin to the bad old days of PCs with my 4000T, such as no Hi density disk drive (my biggest gripe yet), next to impossible, user-unfriendly internet software(I still do not have it running), a general lack of decent applications.
To add insult to injury, upon obtaining a rare Hi-density floppy I struck a still unexplanable compatiblity problem, so I sent the drive back to the U.S. supplier(no drives have been available in Aust for 6mths.).Because I Knew the drive would not neccessarily be returned quickly (I had burned it out through frustration), I decided to try other U.S. suppliers and after many days found that Visionsoft in Carmel, C.A. had the U.K. Power Computing adapted 1.76m drives.After the drive arrived I Still had compatibily problem (the drive could not be controlled).After speaking with my Sydney Amiga supplier, I spoke with the Amiga Importers,then Australian Power Computing Importers (neither had tried a 1.76 drive of any sort with a '96 4000T),and finally a Sydney based Amiga dealer/technician recommended by several dealers.
After phone conversions with the aforementioned repairer, I carted my machine to this incredibly difficult to find, small backyard-ish operator.
This brings me to my current position, 3 weeks later, no computer although the Hi-density problem is supposedly fixed (the stubborn diehard Amiga technician insists he wants to get both the original 880k floppy and the 1.76k one are work, inconsequential to me!).
After this epic can you honestly say you would stand by a platform which has given you so many dramas?
On top of that can you imagine paying seeming ludicrous money for expansion hardware- $2000+ ,for an outdated video card,accellerator board, and emplant emulator.(In reflection, a momentary lapse in sanity it would seem)
As I write this letter I feel more and more let down by my Amiga and hardware makers, really how could I not? Not even the Video Toaster has remained a shining example of Amiga only brilliance! other examples are too numerous as I'm sure you know.(Gone it would seem are the days of Amiga graphics and OS power and superiority) Ofcourse If was an Amiga hardware/software company, I too would want to find security in other platforms.
If Phase 5 could present a third party perspective of a shining white hope for Amiga to get ahead of the rest once again,I, along with many other Amiga users may be able relight the fires of past. So far I have only read of possibilties and vague news bits in the press.Will it be Power PC as long lauded or now maybe HP RISC?.Who could be bother waiting!
To sum it It seems all too hard for even an Amiga devotee to put up with.To other computer users I know well, having an Amiga is like living on an desert island surround by sharks!Not to mention finding it nearly impossible to defend the cost when dicussing platforms with Mac and PC users.
Sorry to have expanded on my situation to the nth degree, but I feel it was the only way to fully explain my frustration in owning an increasingly difficult to justify computer that's so new.
Sincerely,
A. GordonUpdate - 5 Sept., Still no Amiga,no response from Softwood,No Cyberstorm. A faultless PC Now with an extra 32megs ram (only US$200) or 48megs total-my quest for ultra-high memory happening this quickly and cheaply!Tommorow a Matrox Millenium 2mb video card (real 2D+3D processing power for US$280). Also who needs the flooded,difficult to wade through Aminet archive,when you can easy and simply access the Windows 95.com archive full of Microsoft approved freeware and shareware.I feel my Amiga's ultimatum is nearing, who can afford to keep such an expensive machine, that seems like its devaluating by the second!
Regards,
A. Gordon
I hate to end this column for the month on such a negative note, but there are several good reasons for printing this letter.
I personally don't believe everybody has experienced the problems Mr. Gordon has, or else we never would have owned Amigas for very long. But the comparison between the Amiga and the PC truly exemplifies our modern-day quandary:
How can we go on when there are such problems facing us? That is, if the Amiga's technology is aging, and PC's continue to progress?
I'll let the letter tell the story. But personally, I feel that this condition is not a terminal disease. That we can get beyond it. That we can excel!
But success is more than the letters that comprise it. It's dedication and hard work. And to whom do we owe the duty of this work?
Why, ourselves. Yes, we, the Amiga community...and the myriad companies involved. The integrity that has kept us as one for so long much continue.
Devaluing by the second? A dire prediction, indeed...but possibly one with the potential to come true.
We're at that stage when either pure success or total failure is imminent. But either way, it's going to be quite a show.
We'll see you along the way.
Michael Webb
Publisher and Editor-in-Chief