Slashdot Mirror

← Back to Stories (view on slashdot.org)

goosh, the Unofficial Google Shell

Posted by kdawson on from the land-a-gooshen dept.

ohxten writes "Stefan Grothkopp has come up with a pretty neat tool called goosh. It's essentially a browser-oriented, shell-like interface that allows you to quickly search Google (and images and news) and Wikipedia and get information in a text-only format. This is quite possibly the coolest thing I've seen in a good while."

8 of 310 comments (clear)

  1. Re:Lynx by QuantumG · · Score: 4, Informative

    links is superior.

    --
    How we know is more important than what we know.
  2. Re:What is this junk? by peragrin · · Score: 4, Informative

    Now type in one of the numbers. The link opens up in a new tab/window

    Or type

      open http://slashdot.org/

    --
    i thought once I was found, but it was only a dream.
  3. Re:Slow and doesn't work on my mobile browser by merreborn · · Score: 4, Informative

    it's in BETA and it's a Google side project
    In the block of text at the top of the page, it says "NOT an official google product!". Additionally, the whois info for the domain shows it's personally registered by Mr. Grothkopp himself, in Germany.

    As such, if by "google side project", you mean "an experimental project created by a google employee", I believe you may be wrong. It's some random hacker's side project, and it queries a google API, but that's the only resemblance to a "google side project" it bears.
  4. Re:Totally geeky by martin-boundary · · Score: 5, Informative
    There's already a better choice for command line integration: try surfraw. This lets you stay within a real command shell such as bash, and just type

    $ google what I want to know

    You'll get the results directly in a browser of your choice. If you're like me, you have the browser set up as w3m, so that the google results simply appear in the same terminal where you can click on them. Since w3m is a pager like more and less, you can postprocess the google output, eg

    $ google hello | grep Cached
    www.hello.com/ - 2k - Cached - Similar pages
    www.hellomagazine.com/ - 32k - Cached - Similar pages
    www.hellomagazine.com/royalty/ - 27k - Cached - Similar pages
    en.wikipedia.org/wiki/Hello - 39k - Cached - Similar pages
    en.wikipedia.org/wiki/Hello_world_program - 32k - Cached - Similar pages
    www.elite.net/~runner/jennifers/hello.htm - 157k - Cached - Similar pages
    www.ipl.org/div/hello/ - 20k - Cached - Similar pages
    www.mylalaland.com/hello/ - 6k - Cached - Similar pages
    publicaddress.typepad.com/ - 58k - Cached - Similar pages
    www.sanrio.com/ - 10k - Cached - Similar pages

    Best of all , surfraw is not just limited to google, so you can have a complete shell browsing experience for a lot of different sites.

  5. Re:Totally geeky by nuzak · · Score: 4, Informative

    Capturing all keyboard inputs would require javascript, and if you have that enabled for google, you'd have noticed it already sets the focus to the input box when loaded.

    --
    Done with slashdot, done with nerds, getting a life.
  6. Re:Totally geeky by smittyoneeach · · Score: 4, Informative

    Also, w3m runs nicely as an inferior process under emacs.
    http://www.emacswiki.org/cgi-bin/wiki/emacs-w3m#WThreeM

    --
    Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
  7. Re:Ironic... by oracle128 · · Score: 5, Informative

    Technically, the *player* is promised cake and grief counselling at the conclusion of the test, not necessarily the character (Chell). It is not specified that they will actually be given the cake, only that it will be available (which it was, the character presumably being dead was just an unfortunate circumstance preventing the character from eating said cake). Nor does it specify that "cake" isn't a metaphor for something else, such as "dumped into an incinerator" or that it may be available in the afterlife. We were shown the cake, we know it exists. If Chell didn't want to go back and get some, it's her own fault. Perhaps this user has only used cheats to get to the last level, instead of playing through the whole game.

  8. Re:Not particularly useful by martin-boundary · · Score: 4, Informative
    I posted this in another comment already, but here's a version with more details: if you use surfraw and w3m together, you can essentially have clickable google results inside an xterm, and a first class piping mechanism. You might have to change the color scheme in w3m if it clashes with your *term settings, or just try this out in a plain vanilla black and white terminal.

    apt-get install surfraw w3m

    export PATH=$PATH:/usr/lib/surfraw

    export SURFRAW_graphical=no

    export SURFRAW_browser=/usr/bin/w3m

    export SURFRAW_text_browser=/usr/bin/w3m

    export SURFRAW_graphical_browser=/usr/bin/iceweasel

    export SURFRAW_graphical_remote=yes

    google hello # (clickable results "in" the terminal)

    google slashdot | grep Cached | head

    slashdot.org/ - 76k - Cached - Similar pages
    slashdot.org/bookmark.pl?url - 13k - Cached - Similar pages
    en.wikipedia.org/wiki/Slashdot - 83k - Cached - Similar pages
    en.wikipedia.org/wiki/Slashdot_effect - 34k - Cached - Similar pages
    Cached - Similar pages
    yro.slashdot.org/ - 44k - Cached - Similar pages
    yro.slashdot.org/yro/07/05/02/0235228.shtml - 451k - Cached - Similar pages
    hardware.slashdot.org/ - 40k - Cached - Similar pages
    Cached - Similar pages
    politics.slashdot.org/ - 45k - Cached - Similar pages

    It's also possible to write some scripts so that w3m can open new terminals when clicking a link, and if you cannot live without images inside a terminal, there's the w3m-img package you can install.

    I also like to use w3mman as the system man pager, which lets me click on urls and file paths referenced inside a man page.