Extended expansion connector...

[Mr_Navigator]

You might include in your poll a question that has something like 'if built would you buy it', to get some idea of numbers, its no guarantee but a bit better than holding up a wet finger.

[Dave]

I don't have poll privileges... Paging robheaton or vanpeebles!

[Nasta]

Let me do a step back here.

First, inclusion of a QL compatible bus relies on there actually being such a bus in the system in the first place. The point which might not be obvious is, that there MUST be at least a rudimentary version of a QL-like 8-bit bus in the system, to interface with 8-bit peripherals, and by this I mean _ON_BOARD_ peripherals. Apart from video and a hard disc interface, other peripherals are virtually all 8-bit. Electrical reasons require that the 8-bit bus used to transfer data to and from them, be isolated by buffers. This also includes an OS ROM, or at least a boot ROM of some kind. Given that there is no real possibility for much of an OS re-write (and, some re-writing and writing of code IS required, never the less, if nothing else to init new CPU features to emulate the old CPUs!) it pays to have at least parts of the memory map emulate the QL's fairly faithfully. So, there you have the two required components to have a QL style bus: electrical signal mapping, and logical address mapping. Adding a connector is a small step which in terms of cost adds the connector, signal routing, and minor logic. Only then is the next step (wide bus mode) to be considered. Simple thing like redefining lines to add address space are trivial, the big part of their implementation being deciding where to put them on the connector to provide (some) compatibility, which at this point comes basically for free. Also, at this point, what the actual connector is and how it looks is secondary. If someone really wanted, they could make an adapter to a standard J1.

Second, there is quite a lot to be said about what to put on the motherboard, and how flexible it should be.
A good case could be made for RAM to NOT be user upgradeable. Why? Currently it only makse sense to implement RAM sizes FAR in excess of what was thus far being used on QLs. 16 M bytes is the bar eminimum, in some scenarios, 64, 128 or even 256M. Curently only SMSQ/E supports over 16M, other OS versions need patching. A similar argument could be made for video - since the QL windowing system (based on the PE) relies on making physical copies of window contents in RAM, a certain ratio of screen size and RAM size is in order - 1:16 being a good number. Also, a good argument could be made for IDE to be included - mainly because of it's 16-bit bus width and speed requirements.
On the other hand, one could also make a good argument for phasing out certain peripherals, such as floppy discs, perhaps even parallel and serial ports, and certainly some types of keyboard and mouse ports. All of these are actually 8-bit based. Some of these would be expanded or replaced by (at least) rudimentary USB, ethernet, SDHC. Hence it's not far fetched to imagine an IO board, based on existing parts from a QL or expansion board, hooked up to an 8-bit based connector, that would eventually be removed or replaced.

Third - a simple 8-bit bus interface is FAR easier to tinker with in order to add stuff that we still do not have on the QL. However - and this might make this reasoning more important - the market for such peripherals (good example is indeed a SDHC and Ethernet controller) would be severely reduced from an already severely reduced market, if they were only made specially for the new hardware. Even if (unlikely) two versions were made, for new and old hardware. Including a J1 makes it possible toi design something that works on both, perhaps with extra features and performance on the new motherboard, if the designer is aware of said extra features to use. And I'm not talking about some complicated exotic stuff, but rather something which requires an easy way to string on a chip or two to the motherboard.

So there - some food for thought, and information to make an educated vote.

[Mr_Navigator]

Well argued and stated Nasta, giving a much broader idea to the advantages and disadvantages. I don't know enough about the intricacies of systems at that level. I was just thinking that it may be possible to have a hardware peripheral that allows the older original QL interfaces could then plug in to the new QL system for the small number of people that would want them. As I said just my 2 cents worth

[Dave]

I think a small PCIe->QL J1 adaptor is a trivial matter to put together. Just copy the lines to the appropriate pins, and chuck a 7809 on the 12v line so the card sees 9 volts.... I can make those at home.