## Wednesday, December 31, 2014

### Happy New Years Eve 2015

(Click on the comic if you can't see the full image.)

At least he came to give moral support. Two to the thirteenth won't be around for a while!

Two notes: First, 2^11 - 2^5 - 1 = 2015. That is to say, if all the powers of two, up to 10, were included, the sum would be 2047, which is one less that 2 to the 11th, which is 2048. Second, let's not forget the Y2048 bug! Or maybe the Y2(11) bug, just to be creative. Yes, it's true -- some really old computers, which will be really, really old 33 years from now, will have a problem handling the year as 2 to the 11th power. It should prove as Earth-shattering as Y2K did.

---------------

UPDATE: The Making of a Webcomic

This comic started with an 11-digit binary number, 10 ones and 1 zero. I thought it would be funnier if I used the powers of 2 instead, so I needed 11 twos. And then I decided to include the next power as an extra gag. So I needed to draw 12 twos.

Rather than use one of the two twos I usually use and instead of typing the twos, I decided to try on of the paint programs on my tablet and doodled them. I was worried that they might be too snakelike -- and then I was worried that I'd doodled a row of ducks. (I have to keep this in mind if I ever need ducks again.) Then I made the smaller numeric exponents from 12 on down to 0. Colored it in and emailed it to me PC.

Putting it together I realized that I had too many exponents and not enough twos! Oops! I made a mistake. BUT I PICKED THE WRONG MISTAKE! The problem wasn't that I didn't have enough 2s (and quickly created an extra). The problem was that if the lowest exponent was zero, then the highest exponent I needed was 11.

And somehow though all the checking and proofreading -- including all that stuff above (which I have since corrected) -- none of this popped into my head. Of course, moving to 2^11 power would be a bigger problem than moving to 2^12. Some things are stored as 10 bits (I don't know why, but they were) but nothing would be stored as 11 bits (well, maybe -- programmers are strange).

Anyway, the correction has been made. HAPPY NEW YEAR!!

Bill in Boston said...

I'm not aware of any systems storing YYYY in a 12-bit int. Maybe a PDP-8? Not to say there isn't a bitfield somewhere but that's not the one we're worried about.

There is a Y2038 bug, regarding 31-bit overflow of time_t seconds from epoch 1970-01-01-00:00:00 UTC. We've been trying to transition C codebases to 64-bit time_t (or 64-bit throughout) for some time. (Oddly, the NTP time system rolls in 2036?) Mark your calendar for 2038-01-19-03:14:07.

(x, why?) said...

Not aware of any in particular. It was one of those things tossed around back on the old BBSes and early Internet (for me) discussions in the 90s.

The 2038 bug came up, too, but that doesn't fit today's theme. Or maybe it does.

xKiv said...

I thought 2^12 is 4096, not 2048?

(x, why?) said...

You got me. I had it right in my sketch and my notes and the I screwed it up when I put it together. I even added an extra 2 that I dudnt have in the original drawing.

It'll be fixed in the morning.

Orvine said...
This comment has been removed by a blog administrator.
(x, why?) said...

Orvine said...

Happy New Years Eve 2015!! Yeah it was just an outstanding evening. I attended awesome NY events on this amazing eve. I was really surprised to see the huge arrangements. You know lots of other tourists also came to attend New Years Eve NY events !