1 to 26: 6.2 sec
26 to 1: 10.5 sec
A to Z: 3.8 sec
Z to A: 56.2 sec

As a quick note, my 1 to 26 times were slowed by multisylabic names, where A to Z was not, but each had about the same pause (none) between items.

26 to 1 was good, got slightly caught at 20 to 19, but it didn't repeat at 10 to 9.

Z to A took much backtracking. If I had to do it "straight" I would put "can not complete" instead of a time.

=Blue
...you might be eaten by a grue...

I don't know why I was faster counting down than up. I think it's because I normally count seconds when counting up, so I tend to pace myself.

As for the alphabet... I didn't recite it -that- fast. When I try to recite it as fast as possible... 2 seconds. I will also admit to having practiced reciting the alphabet quickly, forwards and back, in my misspent youth (I coulda spent it playing sports, having sex, and getting drunk, but no, I spent it on programming, Rubiks cubes, and reciting the alphabet forwards and backwards. I now know the error of my ways...)

Timing done with "date;read;date" on a WinNT box running cygwin.

OK, repeating the counting up/down without pacing myself...
• 1-26: 6 seconds
• 26-1: 7 seconds

The best way to describe it would be to say that I store the alphabet as a singly linked list, with a few pointers to "well used" parts so that I don't always start at the beginning. Pointers to "A" (the song), "L" (el-em-en-o-pee), "T" (I guess because of the hard sound), and "X" (xyz).

When going backwards, I did a section, went back to my closest pointer, went forward until I found my last backwards, keeping as much as I could in a small short term memory space (say an array of 5 to 7 elements).

As a side note, doing the alphabet forward was not traversing each element as I needed it, but a rather efficient block load which pulled in a "mouthful" of elements, spitting them out autonomously so that there were no I/O blocking.

Damn I feel geeky. 8)

=Blue
...you might be eaten by a grue...

<cite> When going backwards, I did a section, went back to my closest pointer, went forward until I found my last backwards, keeping as much as I could in a small short term memory space (say an array of 5 to 7 elements). </cite> Yes, this is because you know the proximity of some of the letter. The node for 's' might look like:

s => {
  next => 't',
  near => 'r',
}
[download]



$_="goto+F.print+chop;\n=yhpaj";F1:eval

I'd think it's a hash that looks like a linked list:

%alphabet = (
  a => 'b',
  b => 'c',
  c => 'd',
  # ...
);
[download]

This is because I can start ANYWHERE in the alphabet and continue to the end. But the structure is really a bit more complex than that, since I do know the POSITION of some of the letters, as well as the letter that comes before it. So it's a partially linked list. You'll see my report soon enough...

$_="goto+F.print+chop;\n=yhpaj";F1:eval

How about a doubly-linked list with "hash characteristsics"? The back links are "slower" than the forward links. I say "hash characteristics" since it is unneccessary to count to 5 in order to use the number...Any other ideas? Update: maybe even a hash with linked list characteristics?


AgentM Systems nor Nasca Enterprises nor Bone::Easy is responsible for the comments made by AgentM.</h6>

1 - 4 secs
2 - 7.5
3 - 3.5
4 - 27      # my brain felt like a nested ?= in reverse
[download]

• 3.21 s
• 4.41 s
• 2.82 s
• Couldn't do it consistently with the distractions at work

1. 5.9 s
2. 9.2 s
3. 3.4 s
4. 22.2 s

26-1: 11 s

A-Z: 4 s

Z-A: 16 s

AgentM Systems nor Nasca Enterprises nor Bone::Easy is responsible for the comments made by AgentM.</h6>

Don't say the alphabet one letter/second, because you can probably say it backwards at that speed, too. Try saying it backwards at the same speed you said it forwards.

$_="goto+F.print+chop;\n=yhpaj";F1:eval

• 6:50 s
• 11:92 s
• 4:39 s
• maybe some other time <\ul> Does this have something to do with perl algorithm analysis? Fill in the missin' link for me...
  

  AgentM Systems nor Nasca Enterprises nor Bone::Easy is responsible for the comments made by AgentM.</h6>