4 hours ago, Radfoo said:
It sounds like the decisions have been made already. I can understand the procrastinating but think just need to get on with it, I am ready to buy an x8 and I expect thousands of others will too.
As someone else suggested, I think the x8 should be 8mhz by default, with 12mhz being a turbo mode you can enable. Seems to make more sense to me if the x16 happens alongside. Most would then code for the 8mhz and should be plenty for this kind of computer, but if you need a specific use case then they can bump to the 12mhz.
I think the whole point of the post and the poll is that the decision HASN'T been made but David is getting to the point where he thinks its time to make the decision.
I agree with launching by default in 8MHz (or "8MHz-ish" if it's the one-sixth of 50MHz 8.33MHz) and having an internal setting that can speed it up to 12MHz. The IBM-AT and a number of 286 clones did that with a turbo-mode button. That was introduced in 1984, so it seems appropriately retro to me.
3 hours ago, Shauny said:
The X16 turned out to be an odd abomination, with it's mix of fpga and old ICs, certainly not a computer built from off the shelf parts originally envisaged. The X8 is not what was originally envisaged either, it's just fpga, a poor relation to the likes of the Mega 65 or the Sinclair Next and the Raspberry Pi4.
What's worse is it could be seen as a poor attempt to clone the C64 mini.
Seems madness that there is no crowdfunding at this stage, I think they would get the funds they need no problem. Which makes me wonder why not then? Maybe it's the company structure, or is there in-fighting over who owns what, who is owed money and how much.
I do concede that while it seems like nobody could fairly describe the CX16 it as a "poor attempt to clone a C64 mini", nobody could
fairly describe one FPGA, one old IC, and 25 or more off the shelf through hole ASIC parts ... one off the shelf 8bit CPU, one off the shelf 8bit microcontroller, 2 off the shelf 8bit I/O support chips, one off the shelf flashROM, five off the shelf SRAM, one off the shelf RTC chip, and over 10 of those lovely off the shelf glue logic chips as "it's mix of FPGA and old ICs", and yet that is your description.
But they can't really let the fact that some people will make unfair descriptions of things stop them, unless something is perfect, because that is the recipe for never accomplishing anything at all.
They are definitely taking a cautious approach to the crowdfunding, but it is true that if they are building boards to send to beta testers, they really
should be funding those board and their construction out of the crowdfunding of the CX16p.
14 hours ago, xcv330 said:
I am coming late to this, from the facebook group, but I wanted to say for whatever its worth, the X8 sounds exactly like what I was hoping for when this project first began. I think if the price point is low enough, this really would be useful for teaching as well. If it sells in numbers you can get a much broader base of people wanting to develop for it.
Someone mentioned slowing it down so that it would not be faster than the X16. That strikes me as a terrible idea that's been tried before.
FPGA simulators being cycle exact to what they are simulating is not uncommon, but since this is not a simulator of the X16 but rather a simulator of a non-existent little brother to the X16, I do think they should include a turbo mode to let it run at 12MHz.
12 hours ago, EMwhite said:
I just re-watched the Part 1 and Part 2 of 'my dream' and saw that $50 was one of the initial price point targets, or upwards of $100.
Then later, David was talking about scope creep and showed a board design that looked nothing like the initial 'must haves'. I think he said $250-$300 at that point.
But note that even without feature creep,
which was then pruned out, the initial "must haves" could never be done anywhere close to under $100.
The real CX16s cannot possibly hit that price point. But they can have the real ASIC parts, the ability to understand all the way to the bottom, the extended RAM, and the expand-ability.
The CX16e can't really hit that price point, but it can come close, at least for a bare board, and retain full compatibility with the CX16.
The X8 hits that price point, and can run the same Basic programs as the CX16 up to a point, but sacrifices full compatibility with the CX16.
And (1) they could have launched the crowdfunding for the X8 in mid April (never launch a crowdfund too near to April Fool's Day), (2) the CX16p is at the stage where the crowdfund could be launched if they would just bite the bullet, work out the final development budget and do it and (3) there is not technical obstacle to the CX16c design that they haven't solved with the CX16p design, they just have to finalize their product design decisions to be able to launch that crowdfund at the same time.
So I reckon, just do it. Aim to have a crowdfund launched sometime in the next month or two.
8 hours ago, Getafix said:
I think we can argue over a lot in part because there is no certainty about what's to come but I'll explain what I meant when I wrote this. Perhaps best explained as a sequence of events that I think will unfold.
The Commander X8 will become available and people will post their unboxing and experiences. That will make more people buy it (also, a bird in hand is better than 2 in the bush - the CX16 is in uncertain territory till the finances get settled so may as well get the X8 now). Then, those of us that have made software for the CX16 will convert our software to run on the CX8, and go on to write more software for it. Soon, there will be more X8 software than CX16 software. When the CX16 finally becomes available, will it make financial sense to buy machine that's slower but has more RAM and isn't compatible with all the software now out there? (this all depends on how long it takes to come out, of course). I think for many the answer would be no. For some the answer would be yes - those that care about the fact that the machine isn't all FPGA or maybe want to build the kit or can afford both and see value in the CX16. It will also absolutely sell less than if the X8 doesn't exist (if only by 1 - me) - and this is more true if the CX16 is only in kit form. Retro computers are loved for their look and feel along with their inner workings. A kit computer just doesn't have that same identity.
Obviously I made assumptions and as the facts cement it could alter the script but for now I think the CX16 launches later, probably as a kit or maybe as a somewhat expensive ready to go system and the above narrative plays out.
Speaking just for me - If the X8 launches, I'll buy it and be done. If the X8 doesn't launch, and a CX16 ready-to-go launches, I'll try to buy that if I can afford it. I don't make the distinction between all FPGA and partially FPGA. It's real, it's hardware and I can make something for it. It's not a simulator or emulator. What I make runs exactly the same for all others that have it - unlike on an emulator. I am not a collector that want's something to display only - I want to use it and share my experiences with others that use it and the X8 will do all of that nicely at a low cost allowing more people in the ecosystem and that works for me.
One thing is, the development of CX16 applications on the emulator won't stop just because the LX8 ships. That is a big difference with a system that is going to be primarily supported by commercial software ... if it is support by hobbyist and open source community contributions, and it is definitely going to come out, the application development won't halt.
And another thing that people lose track of is that the most popular open source software is going to get ported whether or not the original author is interested in porting it, since that is the kind of "learning to program" challenge that someone can set for themselves before they feel confident to tackle designing their own software from scratch.
So the image of "all of the software is for the LX8, little is for the CX16" ... I'm not sure I buy that in terms of the most popular applications and games. The ones that won't be ported as much are the most demanding of the CX16 applications and games that simply cannot run effectively on 64K of system RAM.
For the core markets for the CX16p and CX16e, there isn't going to be any appreciable cannibalization. For the more fringe markets, as you describe above there might be some, but I think it is less than you are describing above, and as implied by what you describe above, what cannibalization there might be can be avoided
by simply launching the crowdfunding campaigns in parallel.