Morobu performs his morning prayers, and also invokes Cuelessness to get a blank tile. This he saves in a pocket, as the tile he placed the previous day (giving his plans to follow the path taken by mscha
) is still accurate.
Envigorated by the bright, slightly cool morning, he un-kneels, arises, and packs his things.
Checking the area of the hokora
for anything that can be cleaned, he brushes aside a pile of leaves and sees two more tiles, both left by NoMouse
, on the morning of the 28th
. That was the day Morobu reached the beach of BlueCrab
. These tiles are a lengthy addendum to the NoMouse
tile Morobu had already found. In them, NoMouse
describes having gotten quite lost the preceding day, journeying easily from the balthasar_s
stream to moody7277
, but then looping around and crossing his own path while trying to find the next stream (this one, yappobiscuits
Morobu now searches the whole pond area again, thinking he may find more pilgrim's tiles. The other side of the pond is too difficult to allow full circumnavigation, so he doubles back twice, and thus scans every bit of ground twice. He finds several more branches and twigs, plenty of leaves and a couple neat-looking stones, but no more tiles.
:AZULEING: :AZULERS: GONNA :AZULE:
I haven't been able to find a clear explanation of what :azule: actually is: is it a synonym for ":lol:", meaning that I could put it anywhere I'd put a ;lol: smiley? Or is it a distinct emoticon with its own meaning, so if I insert ":azule:" in a post it means "Right now my feelings/behaviour/facial expression resemble those typical of azule"?
Making it a verb, does "to :azule:" mean "to insert a smiley into a post", or does it mean "to laugh out loud"? Or more creatively, perhaps it means "to laugh the way azule laughs", as in, @azule
has a distinctive laugh and if my laugh sounds like xis, I'm :azuling:. Or more generally, "I'm doing something @azule
would do", or "I feel like @azule
would relate to what I'm (doing/feeling/thinking)".
Does it take a direct object? Is it a ditransitive verb (meaning a verb that takes two objects: in "I gave Megan a trebuchet", "gave" is ditransitive and "Megan" and "trebuchet" are the objects)? @azule
did inform me that when conjugating :AZULE: (heh-heh), the colons should be on the end: I HAD :AZULED:, I :AZULED:, I :AZULE:, XE :AZULES:, I AM :AZULING:. So, while I still don't know what it means, at least I know how to spell it!
<- there, I just ;lol:'d, but did I :azule:?
With the spelling corrections, I also had to change the irregular transformations for certain templates in my automeme
MY XIP :AZULETH: OVER
THEY SEE ME OTTIFYIN', THEY :AZULEIN':
NOW I AM BECOME ENDISHNESS, THE :AZULER: OF BLITZROCKETS
macraw83 wrote: I'm running the spoiler-opener script to avoid those issues, which also alerts you to the presence of a spURLer, which is really handy.
Me too, but oddly my script
was not doing that on the spURLer button in @ED
's signature. I found a bug and fixed it.
For example: now I'm wondering why developers insist on taking things that work and remove functionality from them.
I would add to @ChronosDragon
's answer (all of which I agree with):
Some companies care about long-term backwards compatibility, and others don't. I think Microsoft is the best in this regard: .BAT
files still work, and everything I learned in MS-DOS
still works in the Command Prompt. Notably a lot of device drivers
have broken, and I think that's because the help of a 3rd
party manufacturer is needed.
I believe Microsoft does that because they have tons of people and can afford to employ more people to create stuff like cmd.exe
. Apple, on the other hand, is probably limiting costs by getting all of their customers up to the latest version, so there are fewer problems from customers using old apps and old OS's. I do all my real work in POSIX using methods I learned on BSD UNIX, because that's good enough to do all my work. It was a big leap of faith to start work on Q04B
Well, I said script but really meant LabVIEW program...
Oh sure, LabVIEW
(G) is fine. Though I've done dataflow programming, I've never used that specific language, but clearly you can download an image from a URL. As long as you can also get a status/result code, I think it would be just as good as my approach. All you'd want to do is slow down the timer used to trigger each load event (for the xkcd.com
images I imagine you loaded them as fast as you could).
Hey, a shoutout from @pelrigg
! Where did I put that "hello" smiley? ...
no, that's not it...
uhh, nope. Let's try again...
uhh, these are getting worse, I'd better quit now
ucim wrote: ... Each cuegan gets re-counted every time a tile which "includes" it is merged. Yes?
Yes, which is why you only took 1,300,000 moves to get to 25,000,000 points. The point value in a tile is the number of points you've gotten in the process of creating that tile. Each Trebuchet on your board represents 8 points of your score, plus twice the points you get in the process of making a Castle.
I think half of the monks
are doing DRUL, and the other half are doing LURD, both of which sound slightly steambottlish.
I'm going to have to update - I have 12143.30. Isn't it supposed to update itself?
It updated itself.
It updates itself only when you force a reload, which is typically done with... the reload command! You might be able to force a reload by quitting (or force-quitting on newer OS's) the browser and then re-launching it, or by rebooting the computer, but not always — some browsers still use their cache even when re-launching and re-opening a page.
That is one seaish battery Balthartist! It would probably power my smoke detector for a thousand lifetimes, and keep it from ch*rping.
I bought a dual-power Carbon Monoxide detector. On the box, it is advertised as being "better" because it operates from two power supplies: it takes a battery and also plugs into the mains.
One would think that means, that as long as your AC power is on, the detector is powered by it, and when you have a power failure it switches over to the battery. So the battery will last longer. I was led to believe that would mean les ch*rping... wouldn't you?
I in fact had a power failure, and guess what? It started ch*rping right away! The battery was fine, but it was ch*rping to tell me that the AC power was unavailable. Of course, it also ch*rps if the AC power is fine but the battery is mostly used up. In other words, they designed it so that it ch*rps if the battery is low OR
the AC power is out, OR
both. You don't have to do the math, because I did it for you: it works out to more overall ch*rping.
Apart from the silliness of having an alarm to tell me the mains have gone out (if I can hear the alarm, then I'm at home and I bloody well know the power's gone out), you have to imagine the product design meeting where they actually decided they should add a bunch of circuitry and cost to make something that is even more inconvenient than the original. I mean, now you have a product that can only be installed near an outlet, and it ch*rps more, and apart from that is the same as the old model. Stupid bloody ch*rpers.
Naturally, I fixed it. Using my modest knowledge of electrical engineering (a bachelor's degree in electrical engineering) I took a bunch of readings, made a schematic of the existing design, and added the components needed to make the thing work the way I had first expected: now it ch*rps only if the battery is low AND
the AC power is out.
Anyway, </rant>, but RELATED
inasmuch as I think it fulfulls the decree, at least sorta-kinda-ish.
Glad you like Q04B
! Be careful though, I've heard it can be a bit addictive...
Maybe I ought to get back to those long-postponed molpy boosts...
Others explained that having to type "traffic near Tacoma" or some such, is great for that tiny fraction of people who can type "traffic near" faster than they could click a checkbox
did you notice that big roads (like I-5) have thicker Traffic color-lines than litte roads? The Boston area seems to have 4 thicknesses of lines now. That's pretty neat!
balthapreviewbotbuilder wrote: Some of the tags do look like they are indeed flag based. But not all of them. How phpBB decides where to close a code tag with unmatched code tags inside is still kind of magic to me.
Yup. I looked at it for a while, comparing HTML to BBcode, and eventually decided I wasn't going to figure it out. Remember I have something that converts HTML to my own markup language RHTF, and I also have an RHTF-to-BBcode converter, so it wasn't just a philosophical exercise. Then you asked me (several mips ago) if I was interested in converting BBcode to HTML, but I sort of wasn't...
Yes, warnings are good. I already have some of them.
Oh, very treeish indeed! Thanks! You're clearly ahead of me, you put in some checks I hadn't thought of
ETA: Also, the checkboxen "Disable BBCode", "Disable smilies" and "Do not automatically parse URLs" actually have an effect of hom my conversion code works.
I've never actually used those. Do they affect the results of my post? I guess I should try it... but not on this
post, because maybe I won't be able to put it back to normal.
THE BIG automeme SECTION I WARNED YOU ABOUT
azule wrote:Mine does it infinitely (probably bad). Do you think 7 is a high enough number of times? I'd think more like 70 or 700 might be a good outer limit.
I have a limit just to protect me against myself: if create a template to choose one of three words: template w3: [OTTERS, MOLPIES, KEYBOARDS]
and then I use this 4 times in a row ("[w3] [w3] [w3] [w3]
") that 4th
request would search forever because it has already used up the three available options. Since this thing is running on a public server, I really don't want it to infinite-loop.
And yes, it turns out 7 was not enough for the "You are in a twisty maze of little passages" example, where it is generating a random permutation of 3 items. There are 6 possible results; with 7 as the limit, and running it 600 times, it gave each of the expected results about 95 times and other things (like "You are in a twisting little twisting passages") the rest of the time. That's about what I'd expect: 2
to the power of 7 is 6%. So I increased the limit to 27, and now the bogus phrases only come about 2 times out of 6,000. 2
to the power of 27 is about 1 out of every 60,000, so my results are worse than I expected but good enough.
I see what you're saying about allowing redundancy. If it's not required (a $0 situation) then it's simply "allowed".
As you guessed (I think), it allows repetition in all cases except where it is choosing from a list of class-names or a list of template-names. "[$0]
" explicitly repeats a previous word, so obviously that has to be allowed; but also "[$0/-past]
" (with one slash, where we're selecting another tense, conjugation, declension etc.); and "[$0/hyphenate/]
" (with two slashes, where we're applying a named regular-expression to operate on a previous string); and "[$0/[AEIOU]/O/g]
" (where we're applying an explicit regex, in this case changing all vowels to "O"): all of these forms allow repetition. Also, anything starting with a literal (like "[MOLPIES@0.4, OTTERS@0.3, noun-animal-plural]
") allows repetition whenever one of the literals was chosen.
Maybe, in terms of coding it, it'd be best just to add it in as [noun, $0], where there's the non-redundant first, but the chance of it matching the same in the second position.
I have deliberately left a lot of syntaxes undefined, including "[noun, $0]
". With that particular syntax the probability weighting could be done two different ways, and if I choose to use that syntax I'd have to answer the question of how to weight it: If there are 49 [noun]
s does that mean the odds of choosing the "$0
" are only 2%? That makes sense if $0
was a [noun]
, but what if $0
was a [verb]
? Alternately, we might want "[noun, $0]
" to mean "choose $0
with 50% probability, else choose a random [noun]
". Or perhaps the syntax "[$0, noun]
" could be used for the 50-50 case, with "[$firstname.lastname@example.org, noun]
" if you want to state the odds explicitly (which is my personal preference, as it most closely matches what we already have). In any of these of course "noun
" might be a template-name. (The current implementation throws an error if you mix class-names and template-names in the same list, another case where probabilities are ambiguous at best.) Then there's the possibility of a "[email@example.com, firstname.lastname@example.org, noun-mineral]
" syntax where you have a normal list of class-names (or template-names) and you're giving explicit probabilities of drawing from each.
I'm not going to define any of these until there's a meme that clearly needs it; and I've partly addressed most real examples with the more general-purpose named templates. For example, you can do: template t-noun: [noun]
and then mix t-noun
with other named templates in a list, giving an equal-weighted-probability choice that is similar to what you'd get if it were legal to match classes and templates in the same list (where the probailities would be undefined or ambiguous).
Therefore, no coding change needed, just template adjustment. Or for coding fun, change it to [!noun] for when it shouldn't match a previous [noun], otherwise it'd allow redundancy.
I won't do that: I will always define a new syntax that does not
require existing templates to be rewritten. So the "[!noun]
" syntax would be okay with me, but only if it is interpreted to mean "repetition allowed", since the existing interpretation of "[noun]
" does not allow repetition. And yes, I changed its behaviour when I made it prevent repetition by default, but repetition was very rare, and no rewriting of memes is needed.
Regarding the exclamation point, I'd be more likely to use a symbol that already has a special meaning elsewhere, to reduce the number of special symbols that sometimes need to be %-escaped... so it'd probably be "[, noun]
" since the comma already means something inside square brackets, but a comma at the beginning currently means nothing.
Mystery is good. Until it's revealed. Yes, actually. It's listed in the comments of my template file, actually. Now I will go cry 27 tears, ...
I guess I should have asked first. I did
really intend to have named templates from the start, and when I read through your templates file I noted xkcd #1318 as being a good example for its use.
It's wowterful stuff, of course. It is essentially what I was going to do. Kudos.
Thanks! Does my near-imitation of your plans count as sincere flattery?
I think if you do HAVE SOME [REDUNDANT]NESS, [$0/ANT/ANCY/], AND [$1/CY/CIES/]!
it would print
HAVE SOME REDUNDANTNESS, REDUNDANCY, AND REDUNDANCIES!
A "chain of regex" is that thing where it takes $0
and changes it, then it takes the result of that
and changes it again, and so on.
I'll note that Turing-completeness is not very important, and there is a strong argument for the idea that we should aim for "powerful enough". There are probably lots of reasons why Haskell isn't catching on, but I think the biggest is that it tries to be too powerful, and ends up being too difficult to use.
I think @Link
's use of Python is good, because it means he doesn't have to worry about making mistakes in parsing: he's taking advantage of a parser that already works well. My approach is good because I'm defining a syntax that is designed specifically for memes, but its worse in the sense that it's hard to do certain things. That example that counts letters, multiplies and prints the answer in binary would be really easy in @Link
's AUTOMOME. Since we don't really need to do that, it's okay.
MORE automeme THINLY DISGUISED AS A HASKELL RANT
I have a friend who helped me learn Haskell. It was an enjoyable experience — I particularly like the ability to define nearly any syntax and grammar I want, while the same parser also recognizes three types of LISP-like parenthesized syntax:
Code: Select all
Prelude> ((/)(((^)(2))(3)))(4) -- explicit cars and cdrs
Prelude> ((/ 4) ((^ 3) 2)) -- monadic functions
Prelude> ((/) ((^) 2 3) 4) -- normal Lisp syntax
Prelude> 2 ^ 3 / 4 -- colloquial non-bracketed infix
My friend worked very hard creating a Haskell program to perform the Mandelbrot iteration. In C, it is a single (not nested) loop with just two exit conditions (one of which is the loop counter) and performs about 8 arithmetic operations, about 8-10 lines of code, as you would expect. The Haskell version is also simple, until I asked him to get it to measure its own speed (whereupon we discovered it only ran at 3% the speed of mine) and then asked if there's any way of getting it to run at a decent percentage of the C version's speed. He achieved about 80% efficiency with something like 50 lines of code, but had to violate all the Haskell ideals to do it: the only way to get the speed was to force imperative behaviour. I didn't have the heart to tell him that my own Mandelbrot program runs 16 of these loops in parallel, achieving 10 times greater speed.
Thanks for the origin research on that NSFW "WELL, CH*RP ME..." meme. I also found similar quotes but didn't want to spend a lot of time looking them up, as I thought maybe I'd end up on 4chan or something. I like to trace 'em back as far as existing knowledge permits. As that movie is from 1988 I'd say it's probably close to the origin.
Ooooo, particle clicker. Very nice. Clickclickclickclickclickclickclick....
Hmmm, I just spent 100 million of my funding to buy a Nobel Prize. That doesn't seem right (kinda like selling my grandma)...
— mrob27 399: So next post I'm allowed to uhhh, what exactly?