So this sort of thing never really made much sense to me, because you could emulate this entire thing on a raspberry pi for significantly less than $500. Can anyone who is enthusiastic about this explain the appeal?
His original proposal video[0] explains it well: it's not supposed to be powerful; he is aware of the Raspberry Pi but doesn't see it fitting his vision. It's supposed to be something like the C64 and other computers of the era where you could understand how all the various chips interacted and worked.
Things like DDR memory abstract all that stuff away. Computers today don't "put an address on the bus and read the data" like parallel SRAM/DRAM; they send a command packet to the memory chips and wait for the data packet to come back. Audio interfaces, especially over PCIe, aren't as simple anymore either. You can't write direct to the screen anymore like you could with VGA cards on old IBM PCs. All these abstractions make computers so complex that one person can't possibly understand it all at once.
But a C64 already fits that criteria and costs a lot less than $500. The FPGA that runs the video, the SD card controller, and probably a few other things are more powerful than the entire rest of the device and certainly not understandable the way a C64 was. The philosophy behind it feels incoherent. He talks about some future version being as little as $50, but my understanding from other videos is that that version will lack features and probably all be running on FPGA or something anyway.
Why not just emulate the entire thing on an FPGA from the start and pretend it is simple? Or skip the FPGA and use whatever computer the user already has?
Suffice it to say I just don't get it except as some kind of artistic expression, where even then I think it fails to have coherent vision.
The decision to use an fpga in there hurt some of the appeal for me. But knowing the reason the choice was made makes it forgivable to me. And after completing Ben Eater 6502 breadboard computer, which the x16 announcement years ago inspired me to try in the first place, some of the magic of the processor was lessened in my mind. I still plan on getting one because the community is really fun and I can afford it. Plus, I already have the WASD mechanical keyboard to go with it :) Even if someone were to make it so a rpi could boot right into the x16 emulator, which I think has been done for a c64 emulator, the real hardware would still appeal to me.
One of the reasons there area couple of FPGAs in there is that some of the ICs that they really wanted to use aren't produced anymore. One of them is the FM synthesis IC that he talked about in this video. Quickly looking at Mouser and Digi-Key suggests that nobody manufactures FM synth ICs any more. Same with video display ICs. FPGAs are used to emulate these 1980s-era ICs, which seems reasonable since there's no reliable source for the Real Thing.
When I first watched his proposal I thought why not just take a NES on chip, bundle more RAM and maybe a good mapper and a ROM with BASIC on there + add support for a keyboard? Those are AFAIK still readily produced with the mountains Famiclones available on AliExpress and other places and must be dirt cheap.
The last time I asked this I was given many reasons but the reason that stuck with me is (paraphrasing): "it is a complete system small enough to be understood by a single person".
If you weren't there for the 8-bit era you'll never get it. The point of all of this (and related projects) is to create the ideal 8-bit computer that never existed in the era. Sure modern technology is better, and you can write emulators or "fantasy consoles" like PICO-8 or TIC-80 that run on modern hardware, but that seems like cheating -- like creating a biplane with a jet engine.
This project has turned into a boondoggle. Many of the original goals are not going to be met, it is looking like the sound system may need to be redesigned at the last minute because no one bothered to really verify availability of a no longer made chip (one of the original goals was all modern parts), and the cost is going to be 10x that mentioned in his original video as a target for his dream system.
the main technical side is as he has exploded on youtube he has basically became the defacto keynote speaker at most of the retro computing events. now he is quite knowledgeable in the commodore/vic/pet programming scene however a lot of his other presentations/ videos are collections of community knowledge that he has gathered . that comes part of the issue in the community freely shares tons of info for it to be greatly profited by a few. there is also many cringe at some of his videos where he is given access to rather rare items and doesn't follow 'sane' practices before applying power going as far as saying in one video he is constrained on time to release the video so with a high probability blows up a prototype pc power supply basically for purpose of providing closure to a video.
If it's not the questionable qualifications(although he is an enthusiast, he has a limited professional background in electronics repair and software development, two things that are constantly featured on his channel), it may be because of his gun-toting libertarian leanings - some years ago he accidentally posted a video where he was proudly open-carrying on a shopping trip. Of course some people voiced outrage.
I think the guy is good at making videos - but it's with an asterisk. He doesn't have a lot of expertise to share, but he wants to be seen in control. So the videos themselves are always really polished, but then the thing he does is something he should have asked for help with.
What's this about? I shared this video, and pretty much the only controversy I imagined might manifest was people getting weird about FPGAs, since FPGAs are something people get weird about.
His retrocomputing videos seem helpful enough in the common Youtubish way.
not really, he used it to connect two pins together which he thought acted as a switch, but ended up causing a short. People get way too upset about it, if you've never made a major mistake like that, you've never tried to attempt a repair at that complexity at all.
As someone with no interest in oldschool computers beyond curiosity, I think this project looks like a great opportunity to see a project collapse in real time due to feature creep and bad decisions.
Buying machinery and trying to assemble THD boards at his garage? I know assembling this board would've been expensive, but there is no way that the best decision is to try doing it at home with no skills.
Also, if the sound chip was so available when he first checked, why didn't he just bought it right then? Even one thousand of them would've been cheaper then trying to redesign now. And not only that, he is now thinking of using ANOTHER not currently in production chip BECAUSE HE THINKS IT HAS ENOUGH STOCK! facepalm, just incredible bad decision making.
Visiting a part supplier website and going by the listed inventory is not really verifying availability. It is a well-known problem that part site inventories are not accurate for out of production parts. Verification would have involved contacting the sites and having them actually check that the part was available in the quantity needed.
I'm just curious, is that really any better than using the website? I assume until money has changed hands and the shipment has been arranged, a person on the phone might provide an "optimistic" view of available inventory, in the same way their website does.
It seems like his unwillingness to put the whole computer in an FPGA has significantly harmed the project, given the cost, manufacturing, and chip availability problems he's now facing.
He may have to do the sound system that way, but in my opinion using an FPGA completely defeats the point. You might as well just do emulation (which already exists; there's a Commander X16 emulator for developers and the curious). The whole fun part of retrocomputing is that it is based on real chips, not emulators or FPGAs (which amount to the same thing).
Are there any manufacturing professionals actually involved with the project? I was under the impression that this project was mostly enthusiasts with some of them having just enough engineering chops to get some components designed. I'm not following this closely so I could be wrong... But if I'm correct I wouldn't have high expectations.
Being successful in manufacturing is more than just doing some "book learning" about how others do it. Experience counts
That's over a year earlier. Nothing new had happened, and he pretended this was a new board. An honest person who learned this might say it was deceptive, but he could get away without lying in court since, hey, it had finally arrived!
Also, the Mayflower finally Arrived in America!
He used this claim to get a huge boost in funding, and he blew all of that.
As I'm sure other's have mentioned, he went about all this in some terrible fashion (ignoring if the whole ideas is a mistake). I'm not an expert in PCBs but I have printed some. He could save a tremendous amount of time using Pick and Place on the SMT components for very little extra cash. And he's using through-hole resistors and capacitors which only makes sense if you absolutely refuse to pick and place. Even many of the through-hole could be placed using the right service and many of the SMT components are probably cheaper.
(And while we're at it, you can buy custom key-caps)
I only watch T8bG now to see what he does wrong. He's someone who thinks ancient version control for code is too modern. But it is sad to see him take other's money.
Yea, It's a bit like watching a car wreck in slow motion. I'm sad that he dragged TexElec into this. TexElec has been a good and reputable place to buy retro adapters. Hopefully, it doesn't drag down TexElec.
The board that he shows at 0:41 is the 4th prototype, aka the board that is now shipping to developers and backers.
The board he shows later in the video is the older prototype, which does have a different layout. I'm not sure why he uses this older prototype through most of the video, but I suspect that it's because he already had this footage and didn't want to (or need to) shoot that section over again, just to show a board with a few pieces in a different place.
That project has been a mess from day one. The '16' in X16 was because they were originally going to base it on the 65C816, but couldn't figure out how to deal with the address/data bus multiplexing so went with the 65C02 instead.
38 comments
[ 619 ms ] story [ 5249 ms ] threadThings like DDR memory abstract all that stuff away. Computers today don't "put an address on the bus and read the data" like parallel SRAM/DRAM; they send a command packet to the memory chips and wait for the data packet to come back. Audio interfaces, especially over PCIe, aren't as simple anymore either. You can't write direct to the screen anymore like you could with VGA cards on old IBM PCs. All these abstractions make computers so complex that one person can't possibly understand it all at once.
[0]: https://youtu.be/ayh0qebfD2g
Why not just emulate the entire thing on an FPGA from the start and pretend it is simple? Or skip the FPGA and use whatever computer the user already has?
Suffice it to say I just don't get it except as some kind of artistic expression, where even then I think it fails to have coherent vision.
Comparing this project to the Agon is quite the contrast in decisions made and outcomes reached.
David is a joke among a lot of retro computing people, and utterly loathed by many others
I think the guy is good at making videos - but it's with an asterisk. He doesn't have a lot of expertise to share, but he wants to be seen in control. So the videos themselves are always really polished, but then the thing he does is something he should have asked for help with.
His retrocomputing videos seem helpful enough in the common Youtubish way.
Buying machinery and trying to assemble THD boards at his garage? I know assembling this board would've been expensive, but there is no way that the best decision is to try doing it at home with no skills.
Also, if the sound chip was so available when he first checked, why didn't he just bought it right then? Even one thousand of them would've been cheaper then trying to redesign now. And not only that, he is now thinking of using ANOTHER not currently in production chip BECAUSE HE THINKS IT HAS ENOUGH STOCK! facepalm, just incredible bad decision making.
They verified availability of the chip. The explanation for why this wasn't effective is one of the interesting parts of this video.
Being successful in manufacturing is more than just doing some "book learning" about how others do it. Experience counts
In the video, you can tell from the manual patches he clearly shows the board from on https://www.youtube.com/watch?v=hu3QzMsKb9U on 8/14/2021
That's over a year earlier. Nothing new had happened, and he pretended this was a new board. An honest person who learned this might say it was deceptive, but he could get away without lying in court since, hey, it had finally arrived!
Also, the Mayflower finally Arrived in America!
He used this claim to get a huge boost in funding, and he blew all of that.
As I'm sure other's have mentioned, he went about all this in some terrible fashion (ignoring if the whole ideas is a mistake). I'm not an expert in PCBs but I have printed some. He could save a tremendous amount of time using Pick and Place on the SMT components for very little extra cash. And he's using through-hole resistors and capacitors which only makes sense if you absolutely refuse to pick and place. Even many of the through-hole could be placed using the right service and many of the SMT components are probably cheaper.
(And while we're at it, you can buy custom key-caps)
I only watch T8bG now to see what he does wrong. He's someone who thinks ancient version control for code is too modern. But it is sad to see him take other's money.
The board that he shows at 0:41 is the 4th prototype, aka the board that is now shipping to developers and backers.
The board he shows later in the video is the older prototype, which does have a different layout. I'm not sure why he uses this older prototype through most of the video, but I suspect that it's because he already had this footage and didn't want to (or need to) shoot that section over again, just to show a board with a few pieces in a different place.
8-Bit Guy: The Commander X16 has finally arrived - https://news.ycombinator.com/item?id=33181431 - Oct 2022 (7 comments)
The Commander X16 has finally arrived [video] - https://news.ycombinator.com/item?id=33170826 - Oct 2022 (9 comments)
Commander X16 - https://news.ycombinator.com/item?id=27204168 - May 2021 (114 comments)
Commander X16 8 bit computer now has its own website - https://news.ycombinator.com/item?id=23798823 - July 2020 (77 comments)
https://mega65.org/
And of course Agon Lite, Basic*Engine, CB2, RC2014, Tynemouth Minstrel, My4th, Maximite/Colour Maximite/Colour MaxiMite 2, Superfo Harlequin, Sizif-512, N-GO...