[Coco] Re: Coco Digest, Vol 19, Issue 123

[farna at att.net]

May as well add my thoughts!

1) The ONLY good thing about using WinXP is that most people will have a copy. That said, since the goal is to create a stand-alone machine, not just another software emulator, you'd have to add the cost of a copy to the machine. Only serves to drive costs up, doesn't do anything for the machine. And as pointed out, the systtem requirements just to adequately run XP are high! FreeDOS or a bare bones Linux installation should run fast with less processor and memory.

2) FreeDOS or something similar sounds like the least intrusive solution. Linux is another possibility, since it can be pared down to just what's necessary to run the emulator. I'm thinking the idea is to have this thing boot up and look/act sort of like a CoCo for all practical purposes -- the native OS should be invisible to the user once the machine is booted. All programming  and such should be handled from the CoCo interface without having to go directly into the OS. That would make a more true CoCo 4 anyway. 

3) I like the Mini ITX idea, though with cases and such going at relatively low prices any micro ATX board should be fine. The ITX boards are probably the most stable since they don't sell in as high a volume and go through fewer changes than more standard boards. The problem with changes being made in boards can't be easily overcome. Changes in the software will just have to be made to accomodate the boards. As long as one manufacturer/series is stuck with that whould be east enough to keep up with. Trying to support multiple boards is insane! So pick one for primary support, then make the code open source. If someone wants to go through the trouble of supporting another board they can, just make it in the source license that the modifier MUST send a copy to the main support group/site. Hmmm.... instead of making the code freely downloadable, require people to request the code. That just gives contact info so you know who has it and what they are working on (must supply that info). Instead of making it truly open source, I'd keep some conttrol over it just to monitor projects. 

4) The main attractions of the CoCo to a hobbyist are ease of programming (in DECB) and ease of interfacing. Should be easy enough to make an adapter card for a PCI slot, I think. It should be a buffered card though, just in case! I'd go ahead an put a wire-wrap header on the card as well. Would have to take the cover off the case to get to it, but the header could be used for pluggin something into the board as well as wire wrap experiments. When has a hardware hacker been concerned about having the case cover off while tinkering? Not often, I'm sure!! ;>

--
Frank Swygert
Publisher, "American Independent 
Magazine" (AIM)
For all AMC enthusiasts
http://farna.home.att.net/AIM.html
(free download available!)
			

 -------------- Original message ----------------------

> Date: Sat, 23 Apr 2005 22:52:00 -0400
> From: David Hazelton <davehazelton at access-4-free.com>
> Subject: Re: [Coco] The evolution of the Coco..
> To: CoCoList for Color Computer Enthusiasts <coco at maltedmedia.com>
> Message-ID: <426B09D0.4000400 at access-4-free.com>
> Content-Type: text/plain; charset=us-ascii; format=flowed
> 
> Roger Merchberger wrote:
> > Rumor has it that Steve Bjork may have mentioned these words:
> > 
> > [[ putting the rulers away for a second... ]]
> > 
> >> As for the operating system to use, my vote goes to windows XP Pro.
> > 
> > 
> > You've got to be joking. Isn't that like using a 12 pound (5.5kg) 
> > sledgehammer to swat flies? Personally speaking, I would not want to 
> > *have* to outfit a machine with 512Meg RAM just to run a CoCo emulator, 
> > and I'd also prefer not to have to reactivate the rascal every time I 
> > diddle with the hardware... etc.
> > 
> > Speaking of hardware, can you say "Hardware Abstraction Layer?" Thats 
> > the thingydoo that doesn't let user-mode applications diddle with the 
> > hardware directly -- that's why most emulators don't work well in NT/2K/XP.
> > 
> > If we're gonna have to write the extra video drivers and stuff by hand 
> > anyway, why not go with something free, like FreeDOS or somesuch? A 
> > *lot* lower hardware requirements, the DOS can "get out of the way" of 
> > the emulator much more easily, and one heckuva lot more documentation 
> > available.
> 
> > Besides, WRT motherboard choices, I was thinking a 1GHz Via (Cyrix) core 
> > on a Mini-ITX board. No fans necessary (if using an external PS brick) 
> > so it's quiet; small enough to fit in a CoCo case if you so choose (tho 
> > I don't think _I_ would, as that would make fitting a CDRom Drive 
> > difficult), fast enough yet for an emulator in DOS to make a great 
> > platform, and the specs for Mini-ITX have been out long enough  that the 
> > boards themselves aren't that expensive or proprietary -- lots of 3rd 
> > party cases, PSs, etc. A full machine can be built for around $500, less 
> > if you forgo DVD and big HD.
> 
> I have seen these as little at $300.00 with CD-rom and 20gig HD without OS
> > 
> 
> > 
> >>  Best of all, every new hardware product these days has drivers for 
> >> Window XP.  Yes, I know, I would have to use a hard drive with XP.  
> >> But is that truly a problem?
> > 
> > 
> > My $0.0000000000000000002,
> > Roger "Merch" Merchberger