Star Raiders

" The 1979 Atari 800 Game.

Can this be ported to an enhanced Apple 2e 128k or GS, thoughts ? "

Uncle Bernie comments:

I know this game very well, as I "flew" far more hours of missions against these 'Zylons' than I flew real airplanes.

Back in the day (1979), 'Star Raiders' blew everybody who saw it away, it seemed to be impossible, but they did it.

As to porting, the source code is out there, somewhere in the web, I once had it, but lost it again, due to migration of my stuff from computer to computer. I had studied the source carefully and it's amazing how the pseudo 3D effects are made in real time, there are a lot of clever math tricks involved, which consume almost all of the CPU time. The Atari has hardware sprites called 'player missile graphics' and these are used to make the enemy ships and the photon torpedoes. IIRC, all other  moving objects such as the stars are bitmap graphics. And, IIRC, up to about 40 moving objects can be on the screen at any given time, moving through a pseudo 3D space, and the animation is smooth, except after explosions, where clouds of debris from destroyed enemies float around, and the 3D math for them then slows down the frame rate considerably. But all these objects in the debris clouds are bitmap  graphics like the Apple II has. The enemy ships and the photon torpedoes are small enough in size to use software sprite techniques, too. So there is no reason why the game could not be implemented on a normal Apple II.

There are only two problems I can see without having tried to write any code for it: the Apple II runs at 1 MHz CPU clock, and the Atari 8-bit runs at 1.78 MHz CPU clock, which is half the NTSC color subcarrier frequency. The graphics system DMA steals some of the cycles, though. But still, you can expect the Atari to run about 1.5 x times faster than the Apple II. I don't think this is a problem unless too many moving objects are on the screen (the explosions ...)

What a standard Apple II could never do is the sound effects, which are very good ... when the enemies destroy your ship, you can hear how the glass of the cockpit shatters and how you (or what is left of you) gets sucked out into the vacuum of space in a very violent / explosive way. At least, this is the impression these sound effects make.  Still, on an Apple II equipped with a Mockingboard, the same quality sound effects possibly could be had.

All we need is some volunteers to sit down and write the code. Such a translation from a different hardware platform to the Apple II also would be a tough test case for the capabilities of generative "AI" code writing tools: the "AI" can't just "steal" the code from some place on the web, and obfuscate it, to fraudulently conceal the original author(s), but in this case the "AI" must also understand how the Atari custom chip based hardware interacts with the software. And then the "AI" must get creative and replace the hardware capabilities being used with software sprite techniques (it certainly would 'steal' from some other Apple II game).

The question is if the resulting code gives the same gameplay, or not.

- Uncle Bernie

P.S.: for those interested, here is my take on "AI":

Y'all know that I'm very sceptical and critical about "AI" and that I think it's a giant fraud / scam based on theft of IP the "AI" scrapes off the internet, violating not only copyright law, but also breaching the LICENSE conditions usually attached  to "open source" software, such as the requirement to honor / mention the original author(s), to not remove their names, and in many cases, the prohibition to make profit from selling derivative works also is breached (the "AI" companies sell the stolen / plagiarized / obfuscated works for their "AI" tokens --- which is selling the stolen goods for profit). A feast for the lawyers, once they sniff the blood in the waters like the "sharks" they are. This is why the "AI" companies want to do their IPOs as soon as possible, before the law (and the lawsuits) catches up with them. IMHO, any business model that is based on IP theft, breach of copyright, plagiarism, obfuscation of sources, and at the same time greatly exaggerating the capabilities of their product(s) or service(s) to lure investors in, is unethical and doomed, if not outright criminal. But I might be wrong, it's just my opinion. Maybe "AI" turns out to be the next big revenue generating thing after the microprocessor revolution has run its course and ran out of steam. But it could also turn out to be one of the biggest tech scams ever perpetrated against an unsuspecting public. Time will tell ! I am very sceptical as I did a lot of work in the field myself, and fondly remember the hype (and the disappointment) with the neural network buzz of the late 1980s and early 1990s (see https://www.sigarch.org/neurochips-from-the-90s) and I still have lots of "Neurocomputing" books in my bookshelves. What I experienced myself is that with neural networks, you can get very stunning and amazing results for almost intractable problems quite fast, but only up to a certain point, to put it simple: "It works right 95% of the time". The problem, as I experienced it, is that if you need a solution that is 100% right all the time, the neural networks of back in the day could not deliver, no amount of time, effort and money thrown into further development could improve it in any useful way, you got maybe 1% or 2% improvement for 10 times of the initial effort that got you to the 95% success rate. But if 100% success rate is required for a product, and you have only 97%, you have no product, unless you sell it anyways, defrauding the customer, and then taking the money and run. The lawsuits which came out of this (such as over imperfect natural speech-to-text software) were epic, and the whole neural network buzz fizzled and disappeared, after huge amounts of money had been sunk into it. And if you read all the significant scientific papers from the 1950s and 1970s on the topic - which I did, but only after having wasted my time and money on various neural network kits - you will find out that these pioneers had run into exactly the same problem: getting almost there, but never reaching the goal. This is why it took ~20 years where "neurocomputing" was in limbo and got almost no funding anymore, until the IC technology of the 1980s and 1990s revived it, briefly, and encountered the same insurmountable obstacle, again, to again fizzle and get into the next limbo. And then, it took ~30 more years until the current "AI" appeared and seeks investors, but on a much, much larger scale, billions of US$ invested instead of millions of US$ back in the previous attempt of the 1980s and 1990s. And trillions of transistors more thrown into it ... modern nm CMOS technology makes this feasable. Still, as far as I can tell, they seem to run into the same fundamental problem - again ! Because it makes no economical sense if  "AI" can generate a "work" of any kind in a short period of time, which however turns out to be faulty, and the effort / time expended for humans to track down and fix the bugs costs more than the initial savings brought about by using the "AI" in the first place. There are already reports out there, claiming that the current "AI" brings no real benefit as the cost savings promised don't materialize. Same story as back in the day, when the "neural network gadget" you just had bought for lots of money allowed you to dictate a text, such as a book you were writing, by natural speech to a PC - the dream of any writer - but then you spent countless hours with proofreading and fixing all the misunderstandings, misspellings and faults. More time spent than if you had typed the text into the computer by yourself, in the traditional way, using  a keyboard.

It seems history repeates itself, in the field of "AI", but now with much larger investments at stake. - UB