Can anyone confirm what @textfiles says at 15m: The developers didn't have a development tool and they reverse engineered a pre launch PSX? Sounds like BS to me as it would have been an infringement of their license!
You see back in the day people were creative and always poked around and did neat stuff. It took much less effort than it does now but it most certainly is a possibility.
Also, the developer is Japanese: https://en.wikipedia.org/wiki/Tamsoft Jason worked at Psygnosis: https://en.wikipedia.org/wiki/Jason_Scott#Early_work So maybe Chinese whispers lol Edit: I wouldn't be surprised if it's made with Sony's first library, libps/libgs (ie Net Yaroze) with TMD 3D objects.. like RR and RE were
He appears to say they didn't have the SDK or documentation, but somehow they had hardware. I'd accept the team did reverse engineer the development kit and avoided the SDK for some parts, because at least Tekken, Tekken 2, Wipeout & Tomb Raider did that too. You could probably disassemble the game and find out. He made another claim that Sony decided to drop the memory from 8mb to 2mb, but I don't think there was ever a possibility it would come with 8mb and Sony made that clear.
Yeah, they certainly had the SDK, there is probably some truth in terms of having to do reverse engineering but not that they had to reverse engineer everything. Lots of companies were doing reverse engineering. After people reverse engineered the GTE library to improve the speed, Sony responded with an inline method which was faster but obsfucated it to make it harder to reverse engineer it.
It is plausible that they reverse engineered it, but I doubt that they could do it in the time frame. IIRC, Toshinden was a launch game in 1994. Making a game takes months at minimum, especially a 3d one on a brand new platform, so that means they had access to the hardware from mid 1994 or so. And the only way that could happen is having a dev kit, which would have came with documentation and some libraries - it's not much of a reverse engineering at that point.
very early dev is base on Video Edit Machine sony first demo is use Video Editor engine make few picture layer with 3D model , use Lanc remote to play it so is not hard port 3D model on there
It may be a case of "we reverse engineered how the sdk did it in order to reimplement it with better optimizations for our specific case", which wasn't uncommon at the time. I know developers did that for Dreamcast.
I was told (independently) by two different people that were involved in the very early days of the PlayStation that the first hardware Sony sent out consisted of a Sony MIPS based workstation with a custom PCB in it that contained a preliminary version of the PlayStation video hardware. There was no specific SDK or libraries - just some very basic hardware documentation and a collection of sample code that demonstrated various features of the hardware. This setup was updated several times, and eventually all the workstation parts were removed and the chassis was reused to hold the original version of the target box. That was also why the controllers had those 9 pin D type connectors on them - they originally connected to the workstation serial ports and the same pinout was retained.
I've always wondered how this relates to System G (System Gazo). i.e. were these prototypes from that project. It's unfortunate that Sony picking the name NEWS means it's pretty impossible to google for any information, if any such information exists. BIOS and OS dumps would be amazing to get hold of. I'd also love to have access to a target box, or any of the earlier prototype hardware. To see when they switched away from the R3000 and when the GPU changed (the earlier versions of the GPU are not compatible with the released version). It would also be nice to know whether the NEWS workstations always ran the games on it's own CPU and OS like the early "Dreamcast" dev kits that were just a graphics card. Or if they ever had a solution like the PC and DTL-H2000 or the PS2 TOOL where the workstation was used for controlling and debugging but the game ran on a card.
Unfortunately, these were conversations I had some time ago, and in a lot of cases even the people I was asking were unclear about the details. The strong impression I got was that at that stage in the development the hardware was very much a moving target and went through multiple changes. The OS seems to have been some Unix variant - possibly BSD. Apparently it was a nightmare to install even by the standards of early 90's Unix machines - you had to format the drive using floppy disc based utilties, write a disklabel on it so the boot ROM could see it then finally boot the installer using more floppies and a CD - at which point it exited with an error because it didn't like the model of SCSI drive you were using. You also had to make sure you were using the correct build for your machine since they were all different, and getting anything else to run on the hardware was basically out of the question because the machines had some sort of IOP in them and nothing but the Sony supplied OS knew how to talk to it. Even in Japan, they seem to have been uncommon machines - the one they were using appears to have been a NWS-3865, but I was never able to find one - I did locate some 68K based NEWS machines and another with an R4000 in it, but nothing from the R3K generation. The point the CPU changed over from the one in the workstation to the PSX one seems to have been the same point that it the hardware configuration changed from workstation to target box.
first buil i see in Hong Kong it a normal news unix system target to video editing system run demo , it every time load pic and graphic by 2.5” FDD in video edit machince . they label it call : DEX-???? . with out DTL-2000 , they also have PS Acrade target box for dev ...
There were multiple PS arcade dev boxes, ZN1 & ZN2 were used by a lot of companies. However there were namco system 10/11/12 & konami gv, gq & 573. System G was designed for video editing, the prototype could take a jpg and turn it into a 3d polygon.
"I'd accept the team did reverse engineer the development kit and avoided the SDK for some parts" I'd say you're correct, I was looking into early PS1 games and that makes sense! Most likely avoiding the huge and horrid Sony 3D formats and it's really slow draw calls.
There is an excellent set of articles out there (cant find anything off the top of my head though) showing how Naughty Dog reverse engineered the hardware and talked to it directly (bypassing the Sony libraries) in the first Crash game (to get the speed they needed for the game they wanted to make) even though such reverse engineering was a violation of the license. By the time Sony found out, Crash was already so awesome that they overlooked the license violations and allowed it to happen anyway.
I read Andy gavin's blog ages ago, good insight, I'll need to reread it, but IIRC, he doesn't credit were or how they reversed engineered the PS1, just how they developed their tech, ie BSP, animation, etc. IE: Bill Guschwan ( C++ Software Engineer Company Name Sony Computer Entertainment America Dates Employed May 1993 – Dec 1993 Employment Duration 8 mos Worked on Launch of Playstation and helped support of Crash Bandicoot and it's relationship with Sony https://twitter.com/mistagogue/status/705481323016818689 Listen from 8minutes He's pretty clear.. I'm starting to think there wouldn't have been a PS2 without him! This is what I've been researching on my blog post about Bubsy 3D (WIP)
You guys seen this... System G (System Gazo) video? Edit: it's not System G ...It's a game https://www.giantbomb.com/profile/m...ies-at-sony-the-making-of-the-sony-pla/71709/ System G was a geometric engine for 3-D processing developed for broadcasting networks. An example of a current application is its use by Nippon Television Network Co. to change the size of computer-generated faces in the popular TV program Denpa Shonen. Anyone got an example of System G doing Denpa Shonen?
This could be it.... I personally think System G and the PS1 only have similarities in the theory of texture mapping.
