Emoji Dick
A Publishing project in New York, NY by Fred Benenson ·
-
Don't want to forget? Click the star to add this project to your profile.
Update #11: Progress from Today's Work
Posted on November 23, 2009
Hi Everyone,
So today was my first day to work exclusively on Emoji Dick. I made good progress:
* I cleaned up the database of sentences. Originally I estimated ~6500 sentences, then I got to 10,393. Now its back down to 9600. For those of you who are curious, here is the current workflow: mobydick.txt -> perl / unix scripts to tokenize via sentences -> CSV file -> MySQL db -> PHP processing into LaTeX markup -> pdfLaTeX -> PDF file.
I'm currently working with a "fake" version which has all the English sentences from the book correlated with randomly generated emoji sentences. None of them make any sense, but this gives me a feeling for how many pages there will be and what it will look like with the "real" emoji sentences.
* I'm doing a lot of tweaking to LaTeX to produce a PDF. If you don't know what LaTeX is, its a instantiation of TeX which is a markup language for publishing. Its 30-40 years old at this point, but really the only free & open system capable of handling such complex document generation. Its needlessly arcane, painful, and verbose, but I used it to write my thesis, so I'm somewhat familliar with it.
Anyway, the most compact and best version of the book so far as Emoji on top and English on the bottom for each sentence. This is coming out to around 1300 pages. This means two volumes unless I can find a printer (not Lulu) that will print that many pages in one book. Let me know if you have suggestions.
* I ordered hard-cover and soft-cover draft books from Lulu. The hard-cover is color and 40 pages and uses the original 217 sentence PDF. The soft-cover black & white version uses 700 random-emoji sentences paired with the text. These are more like sketches than proofs, but I'll be sure to post photos when they arrive.
* Started thinking out how I'm going to assign the Amazon Mechanical Turk tasks.
In the next couple of days I'll be working on how best to assign the sentences in Amazon. I'll most likely just assign a couple hundred tasks to the crowd and see if the results work and if the flow needs any tweaking. Watch this space for links to the HITs if you want to work on the translations.
I have no idea if you wanted this kind of detailed update, but now you have it.
Have a happy thanks giving and speak with you soon!
Thanks,
Fred
83
Backers
$3,676
pledged of $3,500 goal
0
seconds to go
Funding Successful
This project successfully raised its funding goal on October 19, 2009.
Pledge $5 or more
10 Backers
and receive a PDF of the translated book, and the raw data results of the Mechanical Turk tasks.
Pledge $10 or more
22 Backers
and get your name and an emoji sentence of your choice in a 'Thank You' appendix of the book, a PDF of the translated book, and the raw data results of the Mechanical Turk tasks.
Pledge $20 or more
16 Backers
Limited edition archival frontispiece signed by author, the PDF of the translated book, and the raw data results of the Mechanical Turk tasks.
Pledge $40 or more
23 Backers
and receive black and white soft cover copy of the book, a limited edition archival frontispiece signed by author, the PDF of the translated book, and the raw data results of the Mechanical Turk tasks.
Pledge $200 or more
9 Backers
•
Limited Reward
(1 of 10 remaining)
and receive a limited edition hard cover color version of the book signed, dated and numbered by me, a limited edition archival frontispiece signed by author, the PDF of the translated book, and the raw data results of the Mechanical Turk tasks.
Project By
Connected as Fred Benenson (1417 friends)
I work for Kickstarter as a data engineer.
I have interests in Free Culture, digital art, and computer science. My first project on Kickstarter was Emoji Dick and second was Hackers Party.
ROCK!
Sounds awesome! Happy Thanksgiving to you too!
Question: are you going to try and keep character name translations consistent or will it be up to the crowds to decide? You may want to consider doing mturk on the character names and providing a "cast" table for the crowd to use during the translation process - this way characters don't end up with the same emoji names etc.
nice one fred - ill give you a shout if any other binders come to mind