I'm working on modernizing Rosetta Code's infrastructure. Starting with communications. Please accept this time-limited open invite to RC's Slack.. --Michael Mol (talk) 20:59, 30 May 2020 (UTC)

Talk:Wordiff

From Rosetta Code

AI (artificial intelligence)?[edit]

Perhaps AI could be a stretch goal. This game sounds fun, but I suffer from a dearth of people willing to play text games with me on a single device. Idea for a simple AI: it has a 1/(x^2 + 2) chance of losing each turn, where x is the number of possible answers. (Tweak the numbers to your satisfaction). --Chunes (talk) 18:27, 30 July 2021 (UTC)

Hi Chunes, your question got me thinking about making programs to play the game, and what strategies or algorithms could be used.
Initially I thought a simple strategy of choosing the next wordiff with the most mossible wordiff from that word would be best - it would be a quick algorithm as count of wordiff from a word could be pre calculated and just looked up each turn leading to a fast response likely to not loose in a timed-out game against another program.
But then I thought about if an algorithm tried to actively steariits opponent by trying to choose a wordiff that has no wordiff response -that algo might force the first to loose given enough time - especially as most of the dictionary is used up.
An interesting question...
--Paddy3118 (talk) 07:30, 31 July 2021 (UTC)

Are we to assume what is meant by "checking"[edit]

After checking,   then what?

In the task's description,   are we to assume what is meant by "checking"?


  •   A player loses if the   wordiff   word:
  •   isn't in the dictionary,   or
  •   is a repeat of a past word,   or
  •   differs from the last word inappropriately.


-- Gerard Schildberger (talk) 02:44, 31 July 2021 (UTC)

HI Gérard. The intro before the Task heading, especially "The change can be either:"... Has the extra info.
--Paddy3118 (talk) 07:13, 31 July 2021 (UTC)
I read and understood that part.   What was not mentioned (or even inferred) was what to do with an invalid   (not found)   word.   Yes, it's not allowed, but does the player lose if an invalid word is attempted, or is that word simply rejected?   I was playing with myself   (yeah, that sounds  ...  never mind)   and when using the "normal" dictionary for this Rosetta Code task   (unixdict.txt),   I   (well, me and all my aliases)   kept losing with what I thought were simple words   (words not in the dictionary being used).
         climes  climbs
         homes  tomes combs comes domes foams
         loads  toads  roads  words
         slimer  slimes
         tamer  tames
         timer  times
         tins
         tones  hones  honed
         tools
                 ---  None of these common words have a plural in that dictionary:
         bail  fail  hail  jail  mail  nail  pail  rail  sail  tail  wail   (and  vail,  but that's not a common word)
There are hundreds of simple words, surprisingly a large number of plurals that are not in that dictionary.   It seems silly to lose a game because the plural of   tin   isn't in the dictionary.   One solution, of course, is to use a decent dictionary,   but how is anybody to know that a large number of (common) words (that are words) aren't in a particular dictionary?     It doesn't seem/feel right that a player should lose because a word isn't included, moreover, a common word.   I lost a game when the first word presented was   irs.     ... Couldn't use   ors,   firs,   its,   airs,   ids,   IVs,   irk,   ins,   mirs,   or   sirs.       I missed   IRA   and   ire.     Bummer.     As a sidenote,   when I was playing Scrabble® in my younger and leaner days,   I had quite the huge (mental) collection of three-letter words).   But I digress.     I was toying with the idea that if a word isn't in the dictionary,   that the player could attempt another word.   Eventually, the player would run out of words that   are   in the dictionary.   But that would change the flavor and/or tone of the   wordiff   game quite a bit, I would think.     -- Gerard Schildberger (talk) 15:30, 31 July 2021 (UTC)
Ah. Hi Gérard, when playing with my partner, we either agree what's acceptable, or Google the word when unsure. In the Python source I swapped from the ~25K words of unixdict.txt to another dictionary with over 400K words.
--Paddy3118 (talk) 17:35, 31 July 2021 (UTC)
Sounds like a sound approach.   So you were playing without a Rosetta Code computer program for the referee.     -- Gerard Schildberger (talk) 17:39, 31 July 2021 (UTC)
Aye, developed with my partner a couple of years back, with not a computer in sight. We played again last Friday and I thought of writing the program as something to do for RC. --Paddy3118 (talk) 18:10, 31 July 2021 (UTC)

another word game[edit]

I used to challenge anyone to name as many (human) body parts that only have three letters.   Vulgar (or impolite) words were acceptable, as everyone knew them anyway, and nobody could define vulgar in this contest.   Anything above a score makes you a wordster.   DNA and RNA don't count.     -- Gerard Schildberger (talk) 21:14, 31 July 2021 (UTC)

Sounds as if alcohol would go well with that game :-)
--Paddy3118 (talk) 07:42, 1 August 2021 (UTC)

Words should all be three letters or more[edit]

Sorry about adding that late. It was a always a rule of the game, but I just forgot to write it down initially, and my familiarity made the omission hard to spot. Sorry

--Paddy3118 (talk) 19:08, 1 August 2021 (UTC)