Wednesday, January 29, 2020

Computer History

A brief history ...

Around 1982 to 1983 I purchased my first computer, a Timex Sinclair 1000. This small computer had 2K of RAM, which I would later upgrade at some cost to 16K using a plugin module. It only displayed in Black and White, and both the graphics and 3.5 MHZ Z80 processor are comparable to a TRS-80 from the previous decade. It was a very limited machine whose hardware was inferior in every way to other 8-bit computers at the time.
However, it was the only thing remotely affordable, costing around $100 to $150, whereas other 8-bit computers started at around $500, which is equivalent to about $1300 today. However, it did have a rather clever operating system that made the best use of its limited resources, and I liked it for this reason. The machine was popular because it allowed many people to own a computer who otherwise could not afford one.

In 1984 I upgraded to the Timex Sinclair 2068. The biggest differences were a real keyboard, sound, more memory, color graphics (although still inferior to the competitors), and a better operating system. At $200 it was a good value compared to other computers. This computer, along with the European version, were hugely popular both in the U.S. and internationally. It allowed people to have a halfway decent computer for the time at relatively low cost compared to other computers.

For a short while, I had a thriving business selling software I wrote for this computer.

In 1986 I bought my first 16-bit computer, the Atari ST. Cost was a major factor, because it had much value for the money, but the market would be dominated by much more costly computers. I tried unsuccessfully to make a living writing programs for this machine.

By 1990 my computers were out of date. I eventually sold them to collectors. I no longer had computers at home until the internet was barely starting to become a thing.

Around 1995 a few brave souls were exploring the internet using only text because graphical web pages weren't available yet. I discovered that you could play chess on the internet, with the help of a program designed for that, so I began looking into getting a computer. What I got was bare-bones to the extreme, and cost around $1200. It didn't even come with Windows, and Windows 95 wasn't a thing yet. It had a 33mhz processor, which I would later upgrade to 66mhz. I spent a fortune upgrading this machine. Things that would cost next to nothing today, like a sound card and a CD Drive, weren't standard and cost a lot of money to get. My one regret is that I spent so much money upgrading this computer instead of just waiting for better machines to come out. By the time I had $3000 invested in the computer, it was obsolete.

Around 1998 I got a 400 MHZ machine, which was a huge improvement, but it quickly became obsolete.

So around the year 2001, I got a single-core 2.2 GHz computer. This was a big step up. However, by 2005 it was looking rather limited, so I upgraded to a new dual-core computer, which was the hot new thing at the time. This machine ran at 2.4 GHz.

However, what seemed impressive in 2005 seemed rather limited by 2010. I think that the machine was also having hardware problems, so once again I was looking again for another computer.

In February 2010 I took a bonus that my company gave me and bought a $2,000 Core-i7 iMac. It was significantly better than the average computer at the time, and I justified what seemed like an extreme expense by telling myself that I would try to make this computer last ten years. I still have the computer, but by mid-2019 the display was failing, and the repair shop told me that other components are starting to fail too. The computer still runs, barely, but it is too expensive to try to repair it.

So in mid-2019, I bought a slightly used 2017 iMac for $1150. This was a compromise between performance and price. Although it is 60% faster than my last computer and it has a much better graphics card, it is not a top of the line computer by today's standards. It is more like average.

--

Saturday, January 25, 2020

P vs. NP - The Biggest Unsolved Problem in Computer Science

https://youtu.be/EHp4FPyajKQ

A crude chess program in order to look 10 half moves ahead would take the hypothetical 25 moves possible and do roughly 25 to the 10nth power calculations, which would take a very long time. However, the alpha-beta algorithm eliminates mathematically unnecessary calculations making this more like 5 or 6 to the tenth power, which is a huge difference.

What surprises me is that program Stockfish reduces this to more like 2 to the N power, which is considerably less. Exactly how it does this I'm not sure, although I have some idea.

I would contend that looking deeper in chess will always involve an exponential increase, by definition. To not be exponential means that we could look infinitely far ahead and completely solve chess. This is kind of the point of the video.