CoCo 3 core updates? Docs? Anything?

[rhester72]

I have a core from May of 2019 that won't even let you reset and came with no documentation at all, which doesn't appear to even be an officially-supported core. Anyone know anything more about it?

[mario64]

You have to become a Patreon of Roger Taylor to get the current CoCo 3 core (5/26/2020)
https://www.patreon.com/rogertaylor/posts

[kathleen]

There is also the amazing Dragon32/64 core available from Roger without the need to become a Patreon.
You can also join the realcoco group on facebook where Roger answers always very kindly to the questions.

[ericgus09]

I must have missed the post with the pandemic and all about the new unified group .. that explains why the group has been quiet.


I also think someone else talked about creating an opensource version of the CoCo3 .. I dont know where that is (or if it even got started) .. but it would be nice to have an "official" coco core .

[Cebion]

without the need to become a Patreon.

I don't know of any cores behind a paywall. What do you mean?

[Newsdee]

I don't know of any cores behind a paywall. What do you mean?

There's a few devs that only make their cores available to their Patreon subscribers, or (more commonly) that post "beta" versions for Patreons before eventually releasing the core to everybody.

[Cebion]

I don't know of any cores behind a paywall. What do you mean?

There's a few devs that only make their cores available to their Patreon subscribers, or (more commonly) that post "beta" versions for Patreons before eventually releasing the core to everybody.

That's a big difference, "eventually" , they are ALL released for free. Not eventually, the devs could also not release them at all during the work in progress. In my opinion creators with closed source cores are much more of a problem.

[Chris23235]

I don't know of any cores behind a paywall. What do you mean?

There's a few devs that only make their cores available to their Patreon subscribers, or (more commonly) that post "beta" versions for Patreons before eventually releasing the core to everybody.

Not sure, why you write beta in quotes. It sound like the developers that post beta versions for their subscribers only are doing something shady and that these versions aren't real beta versions.

First of all, in all instances I know of these versions where beta and received a lot of work before becoming open source.

Second, it is the Patreon subscribers who are not only donating money to the develper, but also are doing the work of beta testing for the users who can enjoy the cores, when they become open source (it is nothing different from any other closed beta).

And last but not least, if you don't like closed source codes, just don't use them.

[Newsdee]

Not sure, why you write beta in quotes. It sound like the developers that post beta versions for their subscribers only are doing something shady and that these versions aren't real beta versions.

You're reading too much into the quotes. I meant those beta builds are, effectively, behind a paywall.

There are also some closed source cores (like the Coco3). Strictly speaking nothing disallows devs to do that, but they won't be considered part of the MiSTer project since they cannot be maintained and upgraded with the rest of the central cores. I think that's fair enough.

[rhester72]

Not sure, why you write beta in quotes. It sound like the developers that post beta versions for their subscribers only are doing something shady and that these versions aren't real beta versions.

You're reading too much into the quotes. I meant those beta builds are, effectively, behind a paywall.

There are also some closed source cores (like the Coco3). Strictly speaking nothing disallows devs to do that, but they won't be considered part of the MiSTer project since they cannot be maintained and upgraded with the rest of the central cores. I think that's fair enough.

And from what I read, that's been a problem for this core, because it lagged over six months behind in even achieving _compatibility_ with 128MB SDRAM modules. It apparently still has basic joystick directional issues due to not keeping up with framework updates. In this case, it seems a matter of the developer just being stretched entirely too thin.

[dshadoff]

From what I understand, this core descended from the CoCo3FPGA project's source, and the MiSTer-compatible core developer wasn't pleased with the licensing requirements of making it open source under MiSTer.

In truth, I'm not sure whether this was a decision or a legal requirement (or both) - as the license used by MiSTer may not be compatible with the license for CoCo3FPGA.

Either way, he seems not to be including any of the FPGA-side framework code (thus the problem with new hardware like 128MB SDRAM). He seems to be attempting to make it compatible with the HPS side of MiSTer framework (ie. it uses the published I/O APIs to communicate), reimplementing the FPGA side as required.

This whole scenario is less than ideal, but I've heard good things about its overall operation.


As a separate topic, I believe that alanswx has started work on a CoCo2 core which would become part of the MiSTer repository. CoCo3 has quite a bit different hardware from Coco2 (despite relative compatibility for many things), so I think it's a good thing to also have an actual CoCo2 core.

[Higgy]

He posted today that he has been moving house and the equipment will be unboxed again soon.

[ericgus09]

What I played with up to when he switched to a patron model worked pretty decently .. I am not a huge CoCo3 guy so I can't really speak to specifics but the various things I tried worked well .. and the UART stuff was working well (that was my primary interest, internet BBS'sing / telecom stuff) and that worked pretty acceptably ..

[kathleen]

What I played with up to when he switched to a patron model worked pretty decently .. I am not a huge CoCo3 guy so I can't really speak to specifics but the various things I tried worked well .. and the UART stuff was working well (that was my primary interest, internet BBS'sing / telecom stuff) and that worked pretty acceptably ..

Agree, the core works already pretty well. The only issue I had with this core in addition to the reboot, is that the audio in is not yet wired but this is into the todo list according to Roger.

The audio in is important for me in order to use the cartridge Audio Spectrum Analyser. I own it (the physical edition) and use it on my real coco, but I'd like to have it also working on the Mister. Another advantage will be to load real K7 of course.

The Dragon 32/64 core (which is closer to the Coco2 than the Coco3 is, has the same issue and should be solved in a next release.

I understood from Roger that he prefer to bring several corrections in one shot instead of one by one.

Being a big fan of the coco, I find that Roger is being doing a great job to bring this core alive.

[ericgus09]

Yea I will say he is pretty dedicated .. its just too bad he and Alexy rubbed the wrong way off each other .. but thats not hard to imagine happening..

The CoCo series was a machine I always had "wanted" to play with back in the day but never really got the time/chance to do so .. so it was nice having the core on the mister to putz around with to see what it was all about other than the few odd times I would use a friends or friends of friends machine etc..

[shodge]

I realize this is a old thread.

I am trying to understand the history here for a COCO3 port. I am in discussions with Gary Becker on porting his COCO3 design to MiSTer. I previously did the port to the DE2-115 board. Who is the owner of the port referenced earlier in this thread who had issue?

Gary's port is not of the Dragon - it is a true COCO3. Disk Basic runs presently through a serial port to a PC 'server'. Step 1 would be to get this re-hosted, then move to a local filesystem implementation on the MISTer.

Questions:
Will I be stepping on others working this already?

Is there interest?

-Stan Hodge

[dshadoff]

I have worked with alanswx and others on the CoCo2 core for MiSTer - completely unrelated to CoCo3FPGA - and we hadn't even started on adding CoCo3 functionality to that core yet. As far as I'm concerned, these can be considered two separate machines, so you wouldn't be bothering me (and I don't think Alan either).

As for Roger Taylor's CoCo3 core for MiSTer mentioned above (which I think he has also changed the name of...), I have subsequently read online that he makes the statement that his code is not related to CoCo3FPGA, which means that it is not supposed to be related to CoCo3FPGA either. Since the code for this is not open-source, it is not verifiable, but I don't doubt it.

Are you (and, Gary) planning to make the code for the MiSTer port open and available ? The various versions that are around on the web aren't really available as source (except for one). It would be nice to have it in the MiSTer repository so that it can be kept up-to-date with framework upgrade and so on.

[shodge]

I have also been in discussions with alanswx. He is close to booting the same core. Yes [I will double confirm this with Gary but his source code is free now - its just a bit hard to find. I think between alanswx me we will get it working.

-Stan

[mclout@mclout.com]

I have also been in discussions with alanswx. He is close to booting the same core. Yes [I will double confirm this with Gary but his source code is free now - its just a bit hard to find. I think between alanswx me we will get it working.

-Stan

Hi, you say it is available and open but hard to find. Why not give a link to it if you have found it. That would be helpful. Thanks.

[alanswx]

I have also been in discussions with alanswx. He is close to booting the same core. Yes [I will double confirm this with Gary but his source code is free now - its just a bit hard to find. I think between alanswx me we will get it working.

-Stan

Hi, you say it is available and open but hard to find. Why not give a link to it if you have found it. That would be helpful. Thanks.

We have been chatting with Gary. Stan is working on the core. Right now we have it booting, and a working floppy. Stan and Gary are re-working the video so it will use less memory cycles, so we can try to hook the core up to SDRAM so we can have enough memory to play Leisure Suit Larry, et al.

https://github.com/alanswx/CoCo3_Mister

[Pepeart]

where can I find the 3 roms for the new core released today? Can I get a hint? Thank you
ps. Google hasnt been good for me

[CoCoDaddy]

From what I understand, this core descended from the CoCo3FPGA project's source, and the MiSTer-compatible core developer wasn't pleased with the licensing requirements of making it open source under MiSTer.

In truth, I'm not sure whether this was a decision or a legal requirement (or both) - as the license used by MiSTer may not be compatible with the license for CoCo3FPGA.

Either way, he seems not to be including any of the FPGA-side framework code (thus the problem with new hardware like 128MB SDRAM). He seems to be attempting to make it compatible with the HPS side of MiSTer framework (ie. it uses the published I/O APIs to communicate), reimplementing the FPGA side as required.

This whole scenario is less than ideal, but I've heard good things about its overall operation.


As a separate topic, I believe that alanswx has started work on a CoCo2 core which would become part of the MiSTer repository. CoCo3 has quite a bit different hardware from Coco2 (despite relative compatibility for many things), so I think it's a good thing to also have an actual CoCo2 core.

The RealCoCo core (CoCo 3) has zero to do with the failed coco3fpga project nor does it borrow a single line of code. Coco3fpga has been in the works for 10+ years and still hasn't got it right, so to base your clone on it alone is going to be quite the challenge. Fun, but challenging. As for the politics that seems to dominate most chit chat here, it's absolutely astonishing that instead of just using the best CoCo core available, one would rather spend time in politics. Everybody should know by now that certain contributors break everything they touch, and to have their way with a core that is being created and ported at the same time is not a very wise situation. I was going to start a TRS-80 core until I realized there was already one available. That made me happy instead of angry.

[dshadoff]

I notice that you are quoting me from 15 months ago.

Somewhere in the past 15 months, I found a post somewhere from the author of the Real Coco core which states the same thing. I’ll take his (and your) word for it, as I really can’t verify with no source available. But until I saw that declaration, there was no way of knowing as CoCo3FPGA had been ported to so many FPGA boards already, and seemed to be the de-facto implementation.

I don’t understand why you use the word “failed” in conjunction with Coco3FPGA. Please explain in more detail, so that your statements can be regarded as something other than the politics which you are decrying… (because that’s how they sound). But I believe that anything which is broken can and will be fixed, if it can be explained in detail.

Anybody is certainly free to use any core they feel like. But they can’t make code changes, updates, feature improvements, etc. if the source is not available.

So among other things, a closed-source core is reliant on an individual contributor… in this case, one who bristles against the MiSTer framework, and as such refuses any new features including new hardware until they can be reimplemented by him. 128MB SDRAM were unsupported for a long time, as I recall, as a single example. I’m sure he’s a talented developer, but such Quixotic tendencies are self-limiting.

[CoCoDaddy]

Failed meaning many basic CoCo 3 features do not work. No, I won't go through a list. Other projects that were based on that code are also broken and unreleasable in my opinion. However, the guy who fixes everything is certainly welcome to pass back their code to Gary.

The reason I quoted an old post is because it had false information in it about where the RealCoCo core derived from. Many users know it was based on the Matchbox CoCo core and has been also ported to the MiST, SiDi, Mistica, and will be ported back to the Matchbox soon since all of the Quartus projects reference from the same base code or framework I call RealCoCo for the time being.

I don't frequent here but I thought I'd drop in to say a word or two after finding this post somehow while googling around.