ijor

You can reuse the disk as long as you write them in double step mode only. When you access a disk in double stepping mode, every other track of the disk is skipped (usually odd ones). This way, the skipped tracks keep their original virgin state. So you can write the disk with the PC, take them to the Atari and read them, write them back on the PC, etc ... As soon as you write them on the Atari, or in the PC using a standard (non double stepping application); then those disks are not usable anymore for this purpose until you degauss them.

It is still possible, but ... The software must be ready to deal with an 80 tracks drive (must implement double stepping) and set the correct drive density. I don't know for sure if WriteAtr does or not. You must use virgin unformatted disks as they come from factory. If they were ever written by any kind of drive, or if they come factory pre-formatted then they won't work. There is no way to restore disks to a virgin state with a disk drive. The only way is to degauss the disk. Ideally this requires a special expensive equipment, a degausser. But some people had luck using a bulk eraser, or manually using a strong magnet.

If: 1) The ATR image is in double density 2) You can use Linux 3) Your PC drive is an older DD one (40 tracks) 4) You don't want to write to the flippy side. Then yes, use WriteAtr. Otherwise, it is still feasible, but it's more complicated.

As Dan said, it won't read anything at all, not even a single bit. A single density drive, like the 810, can't read a disk recorded in enhanced density. Or more precisely, it can't read anything that is not in FM encoding (single density). Enhanced and double density are encoded using MFM. The behavior you describe is more or less what will happen if you try to read with a 1050 and using DOS 2.0 a disk formatted with DOS 2.5 in enhanced density. This is a different scenario.

No. They will still be fully deterministic as long as you read $D208 at deterministic times. Actually, depending on the exact frequency you use, it could make the behavior much worse. Again, let’s suppose a demo that runs without human action. The result will be fully deterministic. And if you don’t poll Pokey at deterministic times (say, right after the user pressed a key or a button), then you are just adding an external seed. This has no relation to Pokey and you could use that external seed with any other random generator. Yes, agree, it is much easier and faster with Pokey that automatically “translates” the seed into randomness. In software based systems you would need to read a timer and then seed the algorithm, which could be quite slow.

If you mean using $D208 for a random seed, yes, it *might* be great. But for using its output directly (which is what most A8 programs do), then it is quite poor in comparison to any decent software. It is however, of course, much faster; and good enough for most gaming purposes. Most of the time this is true, but sometimes there are no external seeds available. What will you use for a demo on cartridge that starts automatically?

Ok, I think I might be able to understand what Matt meant in the second article. I think both of us read it too fast ... It is different than the first one. The first one claims that the computer is "measuring random noise", this is no-sense. But the second article claims something different, which is not exact but it is closer to reality. It talks about an "electronic white noise generator". I suppose he was talking about the Pokey polycounter. See the difference, "measuring" vs. "generating" ?

I don't know where the authors of those articles get that idea, but I know it is plain wrong. It is very surprising in the second link you posted, the author is a very well know A8 guru. I don't think they are talking about a different system, it's quite clear in both cases they are talking about the A8. The Pokey Random generator is a very simple one, fully deterministic. Trust me on this. The only relation it has with noise is that it uses one of the Poly counters also used for generating noise effects. So, if you want, you can say it is actually the opposite than what the articles claim. There is one polycounter that is used both for generating random number and for generating noise. You of course don't need to trust me, just check the hardware manual. Using external sources (electric noise, temperature, etc) for generating random numbers is certainly possible. It is done nowadays in some systems, such as some PC chipsets. Implementing this is more complicated than what it might seems. You normally can't use the external source directly. You use it insteed as a random seed, or as a random variable factor into the mathematical algorithm. Implementing this in the A8, of course, wouldn't make any sense. As others pointed out already, Pokey random numbers are good enough for the typical A8 purposes. You don't need anything special for making Pacman ghosts random enough. Advanced random number generators are required for modern strong encryption systems. Nobody cared about strong encryption in the A8 days.

Correct. The sequence is repeatable.

I just noted that you were initally asking about the seed, and not about the sequence generation. It is possible that you can get a truly random seed with Pokley, I'm not really sure. This is because Pokey has NO reset signal. So the initial power-up state of the polycounter is non-deterministic. It has a software reset mechanism, but I'm not sure this resets the polycounters. Note that this is completely different than what the article is saying. The sequence will still be repeatable in anycase. Anyway, the initial seed is not a big deal, there are many ways to get a good seed. The real hard work is in generating a good sequence.

Sorry to dissapoint you, but the article is wrong. The A8 seudo random generator is quite good for its age and class of machine. But it is not measuring any noise, and it is of course not truly random or un-repeatable (it is repeatable). It is derived from a Pokey polycounter which is also used for noise generation, quite different than what the article claims.

Cyber Studio is just great. Don't remember the details, but some of the components were ported to the PC and become part of AutoCAD software.

Of course it is possible. But it is not needed, at least not for a stock Atari drive. A stock C64 drive is customizable, a stock Atari one is not. You'd only need a full drive CPU emulation for emulating enhanced drives like Happy, Speed, etc. And only for special software that customizes the drive uploading their own code to the drive (and even then, it is not always needed). Doing this would be quite a task. One of the reasons is that different enhancements have different CPU variants (6507,6502, 65C02). And some enhancements have no 6502 compatible CPU at all.

If you want, you can say that the 1050 performs all the commands by itself. The format command is special because it is the only standard command that takes a long time. But you can send a write a sector command, turn off the computer, and the drive will finish the writing. Yeah, of course, you must send the sector data previously, but you don't have to wait for the actual writing of the sector on the disk. There are other 1050 commands that take a long time, even more than the format. They are the burn-in diagnostic ones.

I see, I stand corrected. However in which way this makes things significantly faster? Well, of course that there is a minor benefit in freeing the computer from the file system overhead. But this is very minor. I'd suppose, and I understand from what you and other said, that all the TURBO improvements on the C-64 are precisely at a very low level. So it is not the drive "smartness" or the "intelligence" what makes it faster. At least not in this sense of intelligence.

The 1050 (and almost all peripherals except the tape) also has its own CPU (a 6507) RAM, ROM, etc. And depending from the point of view it can be considered even more inteligent. The firmware is capable of understanding high level commands, not very differently than modern PC hard disks, something that obvious requires quite some intelligence. The drawback is you can't do anything else besides the commands supported by the firmware. Upgrade a 1050 to a Happy, Speedy or similar, and you get the possibility of both high level and low level interface. You can go to the hardware directly without using SIO, but as said you are limited to what the peripheral supports. With customizable drives you can do a lot of things.

IIRC they were branded General Electric, but I might be wrong. I think they had a dettachable interface one for each computer. I don't know for sure, but I suppose they were not available in the US. They were likely produced for countries based on tape market (and not disk drive market as the US).

There were third party tape recorders with dual interface (Atari/C-64). I have no idea how easy or difficult is to find them now.

I commented this already somewhere else ... It wasn't very common, but in developing countries you could find arcade coin-up sites where all the machines were home computers. One of those sites was exclusively built on 800XL machines, most of them running off cartridges. They built their own mechanism to press START when the user inserted a coin. I was contracted by them to adapt games and to make them compatible for this purpose. The game had to be in cartridge form and it couldn't start with the joystick button (as many, probably most games do).

It is feasible to connect a ZIP drive with a centronics/parallel interface to the ST. But not easy at all. As some already said, the ST port lacks some signals and functionality that is needed by such devices as a ZIP drive. However (and perhaps, just for the record) I remember reading that somebody developed and sold a full blown centronics interface for the ST cartridge port. I don't know if it still available (guess it's not), but even if you can get the adapter you still need a special driver that probably was never written.

Perhaps slightly off-topic, but somebody knows who are the "heavy" A8 buyers at Ebay? rareatari,doc-in-a-box,mshortland,play2day, etc. The only ones that seem to be publicly know are peterolafson and and elit.

There are/were superb hardware copy products, but not A8 based. Best, by far, is the Discovery Cartridge (Happy product for ST computers). Second best is the Catweasel for the PC. I understand that latest version of German A8 drives were very powerful as well, but don't know much details about them.

Be careful when you go for very high bps rates. The computer software might honour your Pokey divisor, but it might not be able to cope with the faster interrupt rate that this will provoke. I won't be suprised if software that runs perfectly well with the "standard" $0A divisor, it crashes when you use a $08 one. And if you go high enough, at some point the small difference in the base clock between PAL and NTSC systems starts to be relevant. So a very high speed might work in PAL but not in NTSC or vice versa.

I will be interested in ATR images of 5.0. If you want I can send you older versions. Not sure why you would want them, but if you do. Might be 5.0 fixed most of the defects in the versions I knew. The last version I have doesn't have accurate skew algn. copy (needed for disks like MULE), doesn't have fast buffered sector writing, etc. Do you still have original copy protected disks? If so, and if you are interested in beta testing some software let me know.

Of course you must send two bytes, the second is the frame's checksum.