What’s all the Yubnub, bub?


Yubnub, the “(social) command line for the web”, is an underappreciated website, ten years old but constantly updated, that is many things to many people, if not everything to everyone. It’s a front end to a vast variety of search engines and other web services, and is so useful that I’ve added Yubnub capability to Firefox, Emacs, and the Linux command line on my laptop. (For more information, see the Yubnub installation page.)

In case you’re curious why Yubnub has such a weird name, it’s the victory song sung by the Ewoks at the end of The Return of the Jedi. Apparently the word means “freedom” in Ewok. I don’t know what the deal is with the snail, though. An Ewok delicacy?

Here are some of the Yubnub commands I’ve found useful or interesting, including Esperanto dictionaries and Finnegans Wake search engines, as well as staples such as Google and Wikipedia. Sadly, some of the really interesting ones no longer work or have been deleted recently, such as search commands for the Encyclopedia of Science Fiction and the Chess Variant Pages.

To discover your own commands, visit Yubnub and type ls.

  • am: Amazon. Example: am duct tape
  • amzns: AmazonSmile. When you shop here, Amazon donates 0.5% of the price to the charitable organization of your choice. Example: amzns duct tape
  • bgg: BoardGameGeek, database of board games. Example: bgg cosmic encounter
  • brickipedia: Brickipedia, an encyclopedia wiki about LEGO. Example: brickipedia darth vader
  • fweet: Fweet, database of Finnegans Wake annotations. Example: fweet chaosmos
  • g: Google. The default search engine for Yubnub. Example: g random celebrity Note: omitting the initial g will retrieve results from Google by default.
  • gdef: Performs a Google define: command that gets definitions of words. Example: gdef supernaculum
  • gim: Google Image Search. Example: gim random celebrity
  • ifdb: for games at the Interactive Fiction Database. Example: ifdb zork iii
  • libt: Add books to LibraryThing by ISBN. Example: libt 978-0394718330
  • luzme: Luzme monitors ebook prices at multiple online stores and notifies you of price drops. Example: luzme grapes of wrath
  • obl: Draws a random card from the Oblique Strategies deck.
  • pinb: Pinboard.in bookmarks. Examples: pinb chicago searches your pinboard.in bookmarks for “chicago”. pinb -a chicago searches everyone’s pinboard.in bookmarks for “chicago”.
  • wa: Wolfram Alpha, “computational knowledge engine”. Unique and cool. Example: wa half a gallon in cc
  • wkeo: Esperanto Wiktionary. Includes definition of word in Esperanto and translation of word into other languages such as English. Example: wkeo hundo
  • wp: English Wikipedia. Example: wp micronation
  • wq: English Wikiquote. Example: wq simple as possible but no simpler

Industry standard manuscript word counts under Linux

Last night I did a word count on my book in progress within Emacs, my text editor of choice. I was astonished that it had apparently gained 2,000 words with only a few edits – from about 43,500 words to about 45,500 words.

I had been using the Linux command-line utility wc to count my words before that, and it had been returning the lower number. I also tested Gedit (results on the high end), and LibreOffice Writer (on the low end).

I wondered on Twitter which I should trust, and a writer friend advised that LibreOffice would probably be closest to Microsoft Word, the standard among professional editors and publishers, so I should stick to the former. However, I ran a word count under my wife Marty’s copy of Word, and it was both highest of all and furthest from LibreOffice. Emacs was closest! Here are the numbers, from high to low:

Microsoft Word 2010 = 45,653
GNU Emacs 24.2 = 45,466
Gedit 3.10.4 = 45,309
wc = 43,855
LibreOffice Writer = 43,726

Moral: M-x count-words in Emacs comes closest to the industry standard – a little low, in fact, which is better than a little high. Gedit is not bad. Stay away from wc and LibreOffice Writer for counting words if you are writing professionally.

Breaking news from my friend: With a much longer manuscript (around 190,000 words), he’s seeing a spread closer to 4,000 words than 2,000, but otherwise his results are quite similar.

A further postscript: A couple of days later, my word count dropped again by about 1,000 words for no discernible reason. I grabbed an older copy of the document from Dropbox and diffed it with the most recent version. I finally understood that I had turned section numbering off in recent versions, and those section numbers had been counted as words, sometimes more than one. For example, section would count as four words. Multiply that by a couple of hundred sections, plus their appearances in the table of contents, and you’ve got a thousand words that can evaporate invisibly.