Tuesday, March 10, 2026

I wrote this in 2020

Where are we in terms of technology compared to 10 or 20 years ago? There have been some major advances, but most of it has been incremental. Smartphones and tablets are roughly a decade old. In the year 2000, not everybody had internet, but we were definitely headed in that direction. We have gone from VHS to Blu Ray, and from CRT to big screen TV's. I first got cable internet in the early 2000's, and since then I have gone from 3 MbpS to 100 MbpS.

It appears to me that technology in the coming decade will explode. Some of it will be incremental, but some of it will be revolutionary. I am convinced that a decade from now many of our lives will be very different.

Friday, March 6, 2026

My old house in Salt Lake City

I tried to do a screen capture, which didn't work.  

The house was built in 1939, which is the year The Wizard of Oz came out.

The landscaping is completely different.  They planted a bunch of bushes and trees.  There is a "Love is Love" sign in the yard.  It looks like a small jungle, when it was mostly grass before.

You can see an AC unit that I had installed.  The house didn't come with AC when I bought it, but an evaporative cooler which didn't cool very well.

The old dilapidated garage is gone.  It was in such bad shape I never used it except for storage.



I've seen pictures of the inside which was completely remodelled. 

From 12 years ago









Wednesday, March 4, 2026

Re: Apple’s Cheapest MacBook


The benchmark for the A18 Pro chip is not bad, but a little slower than the M1 chip.  For a low cost and likely low power laptop, I think that it will be terrific.  On a phone this seems insane.


The 2017 Core-i7 iMac I sold had a benchmark of around 10,000.  My 2020 Core-i7 model is around 19,000, but it has twice as many cores.  These are power hungry computers compared to the low power RISC style chips that Apple makes.

--

Video Games and Retro Gaming

FYI.

https://videogamesandretorgaming.blogspot.com/

I found a 2068 technical manual with a section called something like "How to make a Scramble like game".  Their solution says to use the Z80 block copy.  This means that it scrolls 8 pixels at a time.  If your objects are 16x16 then I think that 8 pixel scrolling is okay, but obviously not as good as smooth scrolling.

However, the block copy is slow.  At 21 clock cycles per byte, and with 6912 bytes counting the color attributes, this works out to be about 1/25th of a second.  However, it can only move these bytes during the vertical and horizontal blanking periods, which might make it take 4 times longer, although I'm not sure because I am trying to get more information.  I say 4 times longer because the slow mode on the TS1000 ran only at 25% speed because the processor spent most of its time sending bytes to the display hardware, which is how the Atari 2600 worked.  On the 2600 the program had to do this manually, whereas on TS1000 it is more automatic, but it still ties up the processor.

The 2068 only draws every other line on the NTSC screen, so maybe there is much more time available to draw to the screen.

I don't remember if used the block copy on my Diamond Mike game.  I think that I chose to redraw the entire screen on each frame, which is likely slower.  However, on the 2068, this appeared to be the perfect frame rate for this type of game.  If the game had been too slow then I would have tried the block copy.  I did the same full screen redraw on the Atari ST version of Diamond Mike and got roughly the same frame rate.  The Atari ST has a faster processor, but the screen memory is 4.5 times bigger.

If you know the size of your buffer, then there is a way to make the Z80 block copy slightly faster by not looping as much.  Instead of using the LDDR instruction, which does the entire block copy, a person could use the LDD instruction which only moves one byte and updates the pointers and the counter.  It only takes 17 clock cycles.  You could have 32 or 256 of these instructions in a row and then loop however many times as you need to.  Since the instruction decrements the BC counter, you can test the zero flag and loop if not zero.  This would increase the speed by about 24%.