The humongous Sandy Electronics thread...

[Dave]

Did you make the custom Minerva by compilation, or byte-editing the image?

We'd definitely like to get Minerva back in active development, in the most open way imaginable. We think there's a lot of room for improvement in the whole area of storage drivers, plus the ability to best use large memory systems. When the memory map is expanded and there's no need to protect the ROM slot in the memory map, Minerva has room to grow to 64K and larger, which would also allow even more speed optimizations than Lau managed to squeeze into the incredibly small 48K.

Separately, we're still looking to implement ethernet and USB, but without drivers the hardware is unlikely to appear. With drivers, the hardware can be incorporated into everything.

[tofro]

Separately, we're still looking to implement ethernet and USB, but without drivers the hardware is unlikely to appear. With drivers, the hardware can be incorporated into everything.

Without hardware, there won't be any drivers - Catch 22

Tobias

[Dave]

It is a catch-22.

Here's our commitment: If someone makes a firm offer to work with us on the driver side of things, we will develop the hardware, and they will be rewarded to the best of our ability.

The situation with ethernet and USB are slightly different.

Ethernet:
This can potentially work with any system, including a BBQL. We'd like to incorporate it into expansions that make it available to old and new QL systems. We think ethernet is a big deal.

USB:
This is tougher. USB doesn't seem to have the range of hardware available for use in lower power systems. It's not possible to offload most of the work onto a dedicated IC. For this reason, it is not suited for slower systems. Our minimal hope would be to implement the HID generic driver for input, with extended options for other driver classes. It's possible that at just the restricted HID level, USB could work on a BBQL. This involves research we haven't done yet.

[dilwyn]

It is a catch-22.
USB:
This is tougher. USB doesn't seem to have the range of hardware available for use in lower power systems. It's not possible to offload most of the work onto a dedicated IC. For this reason, it is not suited for slower systems. Our minimal hope would be to implement the HID generic driver for input, with extended options for other driver classes. It's possible that at just the restricted HID level, USB could work on a BBQL. This involves research we haven't done yet.

I'd hazard a guess that implementing just enough USB for keyboard and mouse would fulfil much of what it would be used for initially at least - so as not to have to build and include SuperHermesLite/QIMI for example.

[Dave]

That's the choice we have.... Incorporate superHermes, or work on USB.

We think superHermes is good, but at the same time it still just gives us a now defunct 1987 technology for keyboard interconnect and it does nothing to move the QL forward into this century. We'll agitate a little more for USB, and if nothing firms up we'll let it go and incorporate superHermes.

[Ralf R.]

Hi Ralf,
I'm interested to your custom Minnie, if you could share with me a bin file I'll appreciate a lot

No problem, just drop me an email address .

[Ralf R.]

It is a catch-22.

Here's our commitment: If someone makes a firm offer to work with us on the driver side of things, we will develop the hardware, and they will be rewarded to the best of our ability.

Should we try to invoke Tony T.? Crowdfunding...?

[Dave]

Any offers, suggestions, attempts will be heard.

I just want to clarify: Sandy isn't looking to own or control the results of any work done. Everything will be open-sourced and left free for the benefit of the entire QL community.

[RWAP]

It is a catch-22.

Here's our commitment: If someone makes a firm offer to work with us on the driver side of things, we will develop the hardware, and they will be rewarded to the best of our ability.

Should we try to invoke Tony T.? Crowdfunding...?

I do think that development of new drivers is the one area where the QL community is lacking the knowledge and expertise.

If Tony Tebby could be persuaded to develop the drivers, albeit for a fee, then I do think that would be the best solution. Alas, I have had no luck getting an email response out of him for anything to date.

Others in the QL community (maybe some on ql-users) may be in a better position to talk to Tony about this, or possibly others may be able to develop the drivers for a fee.

If crowd-sourcing would be a possibility, then we would need to clearly define what drivers are needed and would need to be developed.

From my point of view I think:

a) An ethernet driver for communicating with an ethernet device - although the question here is where will the TCP/IP stack reside, or could that be provided on the device itself (one solution may be to look at how the Spectranet has implemented TCP/IP and DCHP for the ZX Spectrum).

b) A driver for the SD microdrive emulator - presuming that is the stumbling block here.

c) A USB Keyboard driver

[Dave]

If we can crowd-fund £1,000 for a passing event like QL-is-30, I am sure we can crowd-fund a bit more for something enduring - like a driver for a targeted ethernet chip. There is already a functional TCP/IP stack, so the work is not insurmountable.

That said, USB is much harder.