On Tue, Jan 25, 2011 at 5:00 AM, Peter <[email protected]> wrote: > Plastic wrote: > > > The reason I said ARM is because it's readily available extremely > > cheaply, easy to design with, and there's a plethora of available > > boards already. > > You could use any Microcontroller with Linux. It is debatable which non- > 68K controller would be the best. I might vote for SuperH instead of ARM, > but don't you think all of them are boring mainstream? :-) >
"Boring mainstream means cheap mainstream means long term availability of standard designs at commodity prices. One of the problems hardware developers have faced in the past is they'd pick a suitable programmable logic chip, then it would be superseded, and they'd have to modify their designs. The idea of a QL in a CPLD is nice. The Q60 split it across four - a no-brainer at the time - but I am sure there's lots you could do differently with newer devices on the market. I favor the ARM family for a couple of reasons: it's a standard instruction set that is simple, fast, well documented and incredibly cheap. My 2W power statements are for high end devices with dual cores. Many bottom end system-on-chip ARM devices have standard 500MHz, IDE, SATA, USB, video, I2C, etc and could easily emulate a SGC QL. Inside a QL case, nobody would even know they were not a QL unless told, shown, or very VERY observant ;) This raises the question of what is a "real" QL. Different people have different answers. I would expect a hardware designer to find a hardware aspect to be primary, and a OS user to favor the OS experience. Finally, ARM <> Intel. Intel licensed the ARM design when they acquired Digital and the original StrongARM designs. They lost interest in the market segment and sold it off, and not ten minutes later, it got interesting in that segment and they started the mobile designs which led to the current Atom chips. Atom is okay, but it draws "an order of magnitude" more power than an ARM board and has approx. 3x the hardware cost. Embedded linux is already on ARM, if Clive were active today in this market, I think ARM is what he'd choose. > Following the ARM route, we can easily obtain ready made boards for > > around $100 (70 ukp) complete, or design our own (where are you, Nasta?) > > and build them for around $150 (100 ukp). > > Add margin and it sounds expensive compared to a QL on a chip :-) > When you add the hard drive, interfaces, PCB, the cost would be about the same. However, the quirky, hardly any of them exist, everything is a little bit experimental, a little bit of a hack, using NOS components with no reliable supply. It would be attractive to some in the community, but nobody outside of it. It's a closed market. ARM (or Intel ;) puts QDOS/SMSQ in an open market of standard, supported hardware where traders can sell the boards into embedded markets too, gain economies of scale, and promote the QDOS aspect to new people who might like an easy programmable device for all kinds of embedded applications - I'm convinced (though not being very realistic) that SuperBASIC is a quick developer's dream, and the SBASIC -> C converters etc would do well, also. You say it would be a "massive" amount of Linux work, but the work has already been done. Linux is there and is self-supported and developing. uQLx is there, and works quite well on Linux. It could use some development to increase options, but that's right up this community's street ;) Same would be possible for x86, and obviously nobody cares. You still boot > something which is multitudes the size and complexity of Minerva. I can > see no news and no sensation here. ARM or not ARM, running a different OS > with emulator will never be as cool as the real thing :-) There aren't ever going to be more new QLs. There may be one last gasp Motorola-based board, but I suspect not because the Q60 already fills that role very well. It's neat. And costs more than a high end laptop here in the States. At some point in the next 5-10 years all these QLs will start to become unreliable and die. We need to replace them with something, or the community will disperse. My "heart's desire" is that QDOS, in all its forms, carries on and grows and is seen by new people who are not a captive audience. Our poor kids! It's a cool and capable embedded OS that could do need things... like be in micro-sats, robots, washing machine controllers, home monitoring systems, car entertainment system... Anywhere linux on ARM naturally goes. For me, ARM is a no-brainer. For you, the argument is less compelling, and you're *absolutely right* to feel that way. If the QL was ever just one person's direction, we wouldn't all still be using it, and SMSQ/E, the Q60, the pointer environment, and emulations wouldn't exist. Dave _______________________________________________ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
