On 15-09-11 05:29, [email protected] wrote:
Wouldn't all this "optimization" be futile if the second Ben has a pointing device?
IMO, No. On a device the size of the Ben, I don't think a pointer device would be the most comfortable way to use it. So for programs which allow it, it is useful to create alternatives with a keyboard-only interface.
OTOH, as you write in the other mail, some programs simply require a pointer device. I'm not saying we shouldn't run those. Just that for programs where this isn't the case (and it may not be trivial for all programs if it is), IMO it's worth it to design a new interface, which in many cases may require a complete rewrite of the program. And yes, I realize that's a lot of work.
Plus with USB host you could always attach a mouse...
Sure, but I like the Ben to fit in my pocket, and a (comfortable) mouse doesn't even fit there on its own, let alone together with the Ben.
As someone who hopes to see the Ben as a mini fanless laptop, custom tailoring sounds more to me like a boxing in.
I'm with you hoping this. But I also think that it should be usable without peripherals. Sure, if you attach a mouse and a keyboard and a screen, it would be great to use it as a normal computer. But it shouldn't feel crippled if you have only the NanoNote itself with you.
Especially if you're talking about re-writing non-trivial hard core code. It seems like reinventing the wheel so you can have one 10 inches smaller.
Yes. But that's not only useful for the NanoNote. I like keyboard-controlled applications also on machines which do have a mouse. And applications which can work on a small screen are useful elsewhere as well. So IMO it's worth the work. But if you don't agree, no problem. I'm sure you don't have a problem with me (or others) doing it. :-)
Not to mention that the SOC isn't even 100% libre, the best we can to is tailor the code to something manufacturers cannot make on their own...
Yes. That's why I try hard to make as little code as possible specific to the SoC. The OS I'm writing has very few parts in assembly which need to be ported if a new CPU is used. The drivers are specific to the SoC. And that's all. All other programs talk through a well-defined interface to the drivers. New drivers which control other hardware will use the same interface.
since I've learned that MIPS has its problems I think ARM is a little better because it can replace x86 more viably. The real solution, however, would be something like MilkyMist Soc.
Indeed.
That should probably be where all the "custom tailoring" should be going on. The Ben cpu will come and go, but MM is a real sign of copyleft progress. I'm saying this and I don't even own one.
:-) I agree, but I'm not working on the level where that really matters. Writing drivers for the Ben is useful, because it's nice to have things working on it. And it means I can test things now. :-) But I'm careful to not expose any hardware details to the programs using the library.
Thanks, Bas _______________________________________________ Qi Hardware Discussion List Mail to list (members only): [email protected] Subscribe or Unsubscribe: http://lists.en.qi-hardware.com/mailman/listinfo/discussion
