nedoPC.org

After some further minor fixes, I released 0.0.3 just now. =D

One more "unofficial" build for Windows is ready:

http://nedopc.org/ternary/tunguska/tung ... _1-win.zip (1.2M)

It includes new examples "snow.asm" and "ants.asm", alternative assembler RASM (now supports new commands XAM, XAX, XAY, XYX, PAUSE) and some examples for it.

Is there a way to read RTC data of host machine ? I'd like to write ternary clock for vrtual ternary machine

I looked into it a while back, but I found no way of implementing it that didn't either slaughter performance or look really ugly.

Maybe some time in the future I'll find some satisfying way of implementing it.

I can do special build with this functionality

Well, that would be just great I think there should be an address what will hold current huor, minute and second values (and of course it does not have to be updated every second, update-on-read-request is enought).

I dislike the idea of implementing what is really OS-level functionality into the hardware level of the machine: I think on a hardware level the best way to represent time is something like UNIX time_t, the number of seconds relative to some date, since that is easier for a computer to understand than a mixed-radix time measurement like (year:month:day:)hour:minute:second.

But, since Tunguska's OS is pretty lacking at this stage, it might be possible to add this sort of functionality to the AGDP until it can be properly implemented in the OS.

Well, I didn't say that you have to store date and time in separate thrytes, you can use that UNIX way as well. Just give to programmes a way to read it

Actually it has to be separate trytes, because any RTC does this. Personally I never met computer hardware that has UNIX time on RTC level

Not specifically UNIX time, perhaps, but IBM BIOS for an example, gives you time in seconds or milliseconds passed since some epoch (atleast according to wikipedia.)

It would be very strange for a real time clock to do this sort of a thing on a mixed-radix basis, since they essentially just count the number of oscillations of a quartz crystal. It is much easier to implement a simple second or msecond counter in hardware than one that does second-minute-hour-day spillover in separate fields.

That page on wikipedia describes software methods to retrieve absolute time, but RTC usually has separated values for date and time: http://linux.die.net/man/4/rtc

I stand corrected.

I've added a new instruction AGDP_WHEN that stores the first six fields from struct tm in the AGDP registers to the CVS. To access it, call the following



Fields are:
agdp.R1 -- second
agdp.R1+1 -- minute
agdp.R2 -- hour
agdp.R2+1 -- day of month
agdp.R3 -- month
agdp.R3+1 -- year

http://nedopc.org/ternary/tunguska/tung ... _2-win.zip (1.2M)

Great! I'm going to write ternary clock

Neat, would make a nice addition to programs in the main command prompt.