Morobu glances up at the gradually darkening sky, then back at the mottled-brown duck.
"Find his way back
? Can I ask?"
"Flight patterns are different. We all know that much. This time of yip the Mountain should be easy, just skirt around the west side, circle part way around and continue ... wherever we're heading. Time it right and you get a boost on the up-slope. But..." There is a pause.
Morobu offers more crumbs as encouragement. "What changed?"
"It's not the air, and not the Mountain. But somehow we end up in the wrong place. When I got to this pond I felt like it was too far north. Link
was leaving when I arrived. Link
felt like it was too far south
Morobu finishes the question, "... so Link
is trying to get back on course?"
"Yes," quacks the duck in reply.
Link wrote: Huh, I don't even recall seeing anything I'd call an error in any of your memes, though I did sometimes apply minor tweaks.
Oh, well, they were small I guess. But you seem to be a bit better at thinking of all the grammatical possibilities.
- In the "Billy Madison" template ("If peeing in your pants is cool, consider me Miles Davis.") I turned "Miles Davis" into [character adjecive verb-perfect], which allows intransitive verbs such as "CONSIDER ME CHEWED BEESNAKES". In yours you have [verb-tr-perfect], allowing transitive verbs only.
- In the "Boys’ Life" (1929) meme ("You fellows go out there and play like you’ve never played before!") I was using [verb-inf] for the first and [$0/-past] (the past tense of the previous choice) for the other. Your version uses the perfect tense. The perfect tense is correct (because "you've ..." is "you have ..."); my version was producing "I'M GOING TO GO CAVE LIKE I'VE NEVER WENT CAVE BEFORE."
- In the Arthur Baer (1938) meme ("You can take a boy out of the country but you can't take the country out of a boy."), I was using "OUT OF" all the time, even when that preposition doesn't match the [place]. For example, mine would generate "YOU CAN TAKE CUEGAN OUT OF THE BEACH...". It shouldn't be "out of", it should be "from" or "away from". You dealt with this by excluding the [place]s that use ON as the preposition denoting location-inclusion. After seeing your code to handle this, I realized I needed to fix mine too and addressed it by adding a directional-origin preposition to the [place] multiclass (the IN/ON test doesn't work for me because I prefer "TAKE ... BACK FROM THE FUTURE" to "TAKE ... OUT OF THE FUTURE").
Other recent changes in my AUTOMEME: Regex replacement operations can be given a name, so I now define /has-no-spaces/
to be equivalent to /^([^ ]+)$/$1/
; and a template can invoke another template as a subroutine (though I'm not using that capability yet in the main templates file).
HAT-HAIR TURNED TO LOOK AT MS. FRIZZLE. I ASKED MS. FRIZZLE, AND
XE ANSWERED, "WINGISH MIPS: THEY'RE WHAT YOU BLITZNAPPING BLITZNAP".
"REALLY, MS. FRIZZLE?" I REPLIED, "BLITZNAP? WHY NOT WAIT FOR THEM?
HAT-HAIR LOOKED BACK AT ME AND REPLIED, "BECAUSE WAITING FOR MIPS IS
I wonder how much further this work will go before we're on par with the early AI projects, such as ELIZA
has already tagged every
word as positive or negative (part of the PARRY design). I'd love to have a chOTTerbot
ucim wrote: I'm obviously mustarding the math; I figure that every tile requires two of the previous tiles, except cueballs, which just require a move. So, to make a castle (4 pts) requires two megans, which require four cueballs in all. So in a perfect game the number of points in a tile should equal the number of moves required to make it, no? You'll have more total points (because you keep points from merged tiles), but ignoring random Megans, each move generates a cueball (one potential point), which becomes a point when merged. So the number of moves should be pretty close to the total point value of the visible tiles. No?
When you make two Megans, you score two points apiece. When you eventually merge the two Megans into a castle you get another 4 points. 4+4 = 8 points, but you made 4 moves.
4 Megan-forming points
+ 4 castle-forming points
= 8 = 2 × 4 = moves × log2(moves)
Similarly, to make a trebuchet you first make four Megans (2 points each = 8 points) then you make two castles from the Megans (4 points twice), then make the trebuchet (8 points). You got 8+8+8=24 points out of 8 Cueball tiles:
8 Megan-forming points
+ 8 castle-forming points
+ 8 trebuchet-forming points
= 24 = 8 × 3 = moves × log2(moves)
I'm only counting the moves required to get the "raw materials" from which the larger tiles are made; it always takes a few extra moves to consolidate the pieces — but most of those are happening concurrently with the gruntwork of making Megans from Cueballs, and any extra moves at the end give a few extra Cueballs that count towards some future cascade, so it all works out. In general, your score is close to the number of moves you've taken, multiplied by log2
of the number of moves (an approximation that can be verified by simply starting a new game, counting how many moves you take, and pausing every now and then to write down that number and your score). Or, just watch this Numberphile video
. Start at about 14:35 in the video, and remember that in Q04B
you get more because the lowest tile (Cueball) is a "1", whereas in 2048
the lowest is a "2".
ucim wrote: You could even overlay the scores with a new message if necessary, letting it go away once it's irrelevant.
Oh yes, of course! And if you want to see the score you could touch/click on the message to make the message go away, then touch again to see the message if any. This is why I need folks to think for me!
ucim wrote: What I usually do is, just prior to and just after making a big tile, I manually copy the state string to a text file. I'll also do this prior to running Lucky, after compacting a Lucky streak, prior to a risky move, and if I'm going AFK for a while (in case there's organized sand mustard). I guess that has unknowingly cost me many boosts.
You definitely deserve to give yourself some Cueless
credits. I know you have been using the boosts (leading to our discussion of the useless messages), but I don't know how much. But you can make a pretty good guess:
- Try to estimate how many credits you've missed out on. Here's a rough guide: Each Rosetta gives you about 6.5, each laPetite+LBB gives you 13, each "Lucky on Shed" gives you 26, the Beanette/Amtoo Wedding gives about 52.5, and "Hello" (Beach Reunion) gives 105. If you made one or two moves before jumping out of the game (to Options/Info/Credits) then you got one or two credits, but the rest were "lost"
- Remember that to get the Hello you had to make two Weddings first, and so on, so it's like 105 + 2×52.5 + 4×26 + ... (which explains the 700 or so credits that you calculated you should have gotten during an interval when the score increased by 7 million)
- Take a note of how many Cueless and Snake! credits you have right now
- Export and look at your export string. The numbers of Cueless and Snake! credits are easily spotted in the export string, near the end
- Change the numbers and import
- Continue your prayers, graced by resurrected credits!
The simplicity of altering the game via export/import is intentional. Yes, that could be "abused", but so what? This hotdog is meant to be fun, and if playing a game entirely on boost credits is fun, then have at it! There is no shame, or implication of "cheating" in a non-competitive solitaire game.
In the interview, Kieryn wrote: ... Randall Munroe, he does xkcd, and there was a Time comic that he did, that lasted for three months. ...
lasted for 4 mips... but I guess that last month was so rushed it felt like 3 months overall
macraw83 wrote: Also, does anyone know offhand what the default fora font is called? I need it for a small project I'm working on.
Never mind, found it. "Trebuchet MS", in case anyone was curious
pointed out that we don't all see the same font. I'll add that Apple users will usually see Lucida Grande, and I think Windows users see Trebuchet MS; I don't know about Linux and Android.
Some users (like me) have user stylesheets, I recently changed mine to say: font-family: "Cambria", "Hoefler Text", "Averia Serif", "Rockwell", "Georgia", "Lucida Bright", "Times", serif;
and I often magnify it a fair amount:
How I make the OTT more readable
from a distance on a 22-inch monitor
Same here, I thought it was easy enough to get if you know the definition of factoring
x2 + 5x + 6 = (x + 3) • (x + 2)
though I suppose I discovered the term in the context of an IDE while programming, so it might have been a bit easier to infer.
When I first saw the word I assumed it meant "re-making", because I had studied Latin and recognized the root. The way I refactor code is to add new redundant code, test that it works the same with the new as with the old (sometimes I program it to test itself), then switch over, update the block comments and leave the old unused code in there for a year or two. I don't think this is what most people mean by refactoring, though I guess it's a bit like "re-making".