GW> IM> Because of my job I would be very, very interested in hearing
GW> IM> any factual information concerning an RTC used in a PC which
GW> IM> _does_ update the century byte.
GW>
GW> As most (I believe all) PCs made these days have the RTC integrated
GW> into the chipset with the BIOS also stored in the chipset it becomes
GW> difficult to separate chipset solutions from BIOS solutions :-(.

Uh, well... true, but from an electrical point of view it would still 
work the same way. From the point of view of the machine it makes little 
diffence whether there are individual BIOS ((E)PROM) and RTC components, 
or whether these components are just smaller parts of a VLSI chipset.

From the Systems programmer's point of view, however, there is a big 
difference between the RTC component (integrated or not) updating its 
own century byte and the BIOS (integrated or not) updating the century 
byte. If the BIOS is going to do it then you either have to use that 
short-sighted, crapola, misbegotten windowing technique, or use a 
technique which would determine whether it's appropriate to update the 
century byte or not. (Actually, not too difficult to implement. Probably 
less code than the windowing technique. Would require at least one bit 
from the CMOS, however.)


GW> Although I can see no reason for not making it a hardware solution.

Nor can I. That would be the preferable method, IMO. These days I can't 
even see why it could not be backward compatible with the current 
design.


GW> I can dig out from data sheets RTCs that do support a century byte
GW> if you want. It is doubtful if many will be compatible with the PC
GW> RTC.

Well, that's a consideration -- backward compatibility. One attribute of 
the PC is backward compatibility. I'm in favour of a hardware solution, 
but if given the choice between a hardware solution and backward 
compatibility I think I'd choose backward compatibility. Afterall, you 
only have to make a small adjustment every century. For gosh's sake -- 
it's hardly even worth worrying about.


GW> The RTC is only good to 28th February 2100, thereafter it will be
GW> wrong every century year _except_ the years divisible by 400.

Uh... Ah! Yes, you're right! I hadn't thought about that. So I guess 
it's wrong _twice_ every hundred years -- one adjustment within ten 
months of the other. [:)


GW> IM> Yes, I agree. And I've heard that criticism of the AT's RTC
GW> IM> before. Could you give me the part number for an RTC which was
GW> IM> readily available in 1983 that tracked real years -- "1983"
GW> IM> instead of just "83"?
GW>
GW> Sorry, I can't help there :-(. I don't have stock lists from then
GW> around any more, I needed the space :-).

Well... a little rhetorical that question was. [:) IIRC the first 
digital IC was put into production around 1974, for the U.S. space 
program I think. We don't re-invent digital circuits any more than we 
re-invent code to perform binary multiplication. When someone coughed up 
a decent digital RTC, that circuit was licenced and re-used in a myriad 
applications. Remember that things were changing very quickly in those 
days. IC's meant ease of implementation, and the closer your product 
could emulate another the more it would be used. The RTC in question 
tracked a single-byte year, but nobody believed that we'd still be using 
it in 1999.

The PC has an excuse -- backward compatibility. VCR's, nuclear reactors, 
and Hubble reactors don't have that excuse. [:)

Anyway, take care and TTYL.

---
 þþ VGA:  Very Good Adapter.                        

--- AdeptXBBS v1.11y (FREEWare/2)
382/92