* * *
Running CP/M on a $2 module the size of a postage stamp. See, this is what you get from the progression of technology. An OS that ran on computers the size of a dorm fridge now can be run on the microcontroller from a WiFi module.
The people who designed CP/M probably never imagined anything of the sort. But consider: CP/M was written at a time when computer resources like processor cycles and memory were expensive; it's only natural that it could be run on a modern microcontroller.
But the technological progression that led us to this point has probably hit the point of diminishing returns. I doubt that 40 years from now people will be running Windows 10 on an emulator, on a system the size of a credit card. It's likely that specifications for computing devices will continue to improve, but not as radically as they did over the past four decades.
In all probability the next big computer breakthrough is going to be in software; someone's going to invent a new interface paradigm or a better way of compiling code or something that we arguably cannot predict but which will make computers much easier to use and much more secure than they are now, and perform better in the bargain.
* * *
Solid state drives are coming down in price. I was talking with Og about the computer I'm working on for him. Turns out the machine (salvaged) has a pretty nifty processor in it, scoring 7.3 out of 7.9 on the Windows Experience thingy. The biggest bottleneck is video (4.1) but the video card in the computer he bought from me rates at 6.9, so that'll take care of that, leaving the hard drive at 5.9.
SSHDs are fast. I mean, they're fast. As you may recall I put one into my laptop, Seiren, and it boots in seconds to the Windows 10 desktop, even though it's going on nine years old.
Og is leaning towards the model I wanted to use for Floristica: 120 GB boot drive with essential software (whatever you need to be fast) and then use the spinning metal drive for cache and data. In light of that article, I'm starting to think he'd be better off getting a 500 GB SDHD for the thing. After all, what he spends on the drive will be his total investment in the computer.
Intel has started to release large drives using their 3D XPoint technology, which is (I think) a memristor technology. It's expensive, though, costing $1500 for a 375 GB drive.
That price will also come down, sooner or later. I can remember when a 5 MB HDD cost $3,000. That'd be almost $7,000 in today's money.
It's the whole "progression of technology" thing again. When HP introduced the Laserjet in 1984 it cost over $3,000; a little while after its introduction its price went down to just under that mark. So you'd think--all else being equal--a modern laser printer with the same capabilities would be $7,000...only I have, on my desk, a laser printer that's much faster (20 ppm vs 6 ppm) and higher resolution, and it cost under $100 in 2005 dollars. That's $55 in 1985 dollars.
This $100 printer performs better than the LaserJet IIIsi, the monstrous network printer that impressed the hell out of me in my Sears Business Center days. It could print 17 PPM and had two 500-sheet input trays; the only thing that the IIIsi could do that my $100 printer cannot is duplex printing. I wouldn't mind being able to print on both sides of a sheet of paper without manually feeding pages into the printer, but lacking that feature is not a deal breaker. And consumer printers with duplex printing are coming down in price. The IIIsi cost $5,500 around this time in 1991. That's $9,800 in 2017 dollars.