xnevs 12 years ago

I think this game is a bit different than http://gabrielecirulli.github.io/2048/ Take a look at this situation:

   ___________________
  |___2|____|____|____|
  |___2|____|____|____|
  |___2|____|____|____|
  |____|____|____|____|

in the original the result of going up would be

   ___________________
  |___4|____|____|____|
  |___2|____|____|____|
  |____|____|____|____|
  |____|____|____|____|

but in this game it's

   ___________________
  |___2|____|____|____|
  |___4|____|____|____|
  |____|____|____|____|
  |____|____|____|____|
  • scope 12 years ago

    no score too! :(

    how am i suppose to know when i break my own record [i.e. 3012 - never got passed 512 on the tiles]

    • themattrix 12 years ago

      I'll look into adding this. (I think I'll have to add it to the bash wrapper.)

    • themattrix 12 years ago

      Score is now implemented -enjoy!

      • scope 12 years ago

        thanks!

        edit: you broke it :(

        • themattrix 12 years ago

          Fixed! The score is now calculated as the sum of all merge results (like the original game), instead of the sum of all cells (like I was doing).

          As a bonus, the score calculation is now done in pure sed.

          If you notice any other issues, please let me know!

          • scope 12 years ago

            thanks

            • themattrix 12 years ago

              One final gameplay tweak: there's now a 90% chance that a newly-spawned tile will be a "2" and a 10% chance that it will be a "4".

              The gameplay should now be identical to the original 2048.

  • themattrix 12 years ago

    This bug is now fixed! Thanks for pointing it out!

robinson-wall 12 years ago

> Note for OS X users: The version of sed installed has different options than the Linux one. Notably, it does not have the extended-regex option -r. I recommend building the latest version from here: http://sed.sourceforge.net/

The option is there, just different: -E (the same as grep)

  • celebril 12 years ago

    Why doesn't Apple just ship with standard GNU utils?

    They are the de facto standard. Having the need to compile all these basic stuff for any scripting to be portable is a nightmare.

    • Spiritus 12 years ago

      I believe it's a licensing issue, could be completely wrong though.

    • 13throwaway 12 years ago

      Apple has chosen not to ship any GPLv3 code.

    • gkya 12 years ago

      If your well-beloved GNU coreutils did not breach every POSIX standard and add new flags to every program, no such problem would exist. Seriously, I do not know whether Mac OS is using a custom implementation of UNIX utilities, or distributin BSD coreutils, but I know that it is a licensed, certified UNIX.

      • celebril 12 years ago

        UNIX certification means nothing if the utils are intentionally outdated and incompatible with the de facto UNIX-like implementation due to petty licensing concerns.

        At the end of the day, whether some script works is the issue, not a UNIX branding.

        • james2vegas 12 years ago

          Except Mac OS X is a de jure UNIX implementation, and concern about GPLv3 is not petty.

  • LukeShu 12 years ago

    GNU sed 4.2.0 and up also support -E for compatibility with BSD sed (though this is undocumented).

codezero 12 years ago

I am so sick of the Three's, 1024, 2048 posts on Hacker News, but I upvoted this because of sed :P

carlesfe 12 years ago

We just need now the 2048 in assembler, and we'll be done!

Just joking, I love these and I'm unemployed. Keep'em coming!

  • Ideka 12 years ago

    Well, there's this: https://github.com/Sanqui/2048-gb

    It was posted on HN but didn't really take off, sadly: https://news.ycombinator.com/item?id=7434529

    • jader201 12 years ago

      I bet had they left "Show HN: " off, it would've gained more traction. I have several theories behind that:

      1) It looks more like a discovery.

      2) It doesn't look like a self promotion.

      3) The point of the article is at the beginning. I.e. a certain percentage don't read past "Show HN".

      Again, just theories. No idea if this is really the case. But I don't remember many (maybe any) of the other 2048 clones making it to the first page being "Show HN" submissions.

  • segmondy 12 years ago

    The one that's done for Atari is in assembly. I made one in Prolog last weekend as an exercise since I'm learning Prolog. I figured it was a matter of time till the variation became about language implementation.

midas007 12 years ago

Sed awesome.

Also, disappointed no one has blatantly embedded https://github.com/progranism/Bitcoin-JavaScript-Miner to put all this time wasting to good u$e.

  • Ellipsis753 12 years ago

    Sadly that javascript miner is stupidly slow. You could have a top Hacker News post and still only make pennies.

    • nwh 12 years ago

      Not even that really. You might get a single difficulty 1 share or two, worth literally less than the maximum precision BTC allows.

    • midas007 12 years ago

      You didn't get the joke. That's okay.

gwu78 12 years ago

I also like his poker hand sed script.

Best part is he uses BSD sed. No newlines as patterns.

We use tr(1) for that.

adito 12 years ago

Have someone made Emacs version? Isn't it more easy to write it in elisp compared to sed.

  • RBerenguel 12 years ago

    Well, writing in sed has essentially the "heh, look what I did in sed!" value. Elisp would be a far more simple place to build it (we already have tetris there, there's even a very old simple version of Elite in emacs lisp!)

    • calebm 12 years ago

      Yep. "Real programmers write games in sed".

  • discardorama 12 years ago

    I'm sure Emacs has a "unix mode" hidden in there somewhere. Inside it, I bet there's a full complement of GNU utils. Sed should be in there somewhere.

calebm 12 years ago

Hey Matt... I mean "Sed and Awk the great" :)

  • themattrix 12 years ago

    You've revealed my secret identity!