Commit cfd0effa682d0634ead67e566a5e57bc46ef2077 - NaNoGenLab

+13

-6

ending-concordance/README.md less more

0	0	ending-concordance
1	1	==================
2	2
3		Requirements
4		------------
	3	Hypothesis
	4	----------
	5
	6	We hypothesize that words in English can be roughly categorized using a
	7	characteristic as simple as the pair of letters that they end with, and
	8	that this can be exploited to form sentences which look almost plausible.
	9
	10	Apparatus
	11	---------
5	12
6	13	* Python 2.7.6 (probably works with older versions too)
7	14	* The `gutenberg.py` module from [gutenizer](https://github.com/okfn/gutenizer/)
8	15	* A bunch of Project Gutenberg texts in plain text format
9	16
10		Basic Strategy
11		--------------
	17	Method
	18	------
12	19
13	20	* Read in all the words.
14	21	* Index the words based on the final two letters in each word that is four or
15	22	more letters long.
16	23	* Write out words randomly chosen from two alternating end-two-letters groups.
17	24
18		Sample Output
19		-------------
	25	Observations
	26	------------
20	27
21	28	When run on _Principles of Scientific Management_, I got:
22	29

+11

-6

evaporating-text/README.md less more

0	0	evaporating-text
1	1	================
2	2
3		Requirements
4		------------
	3	Hypothesis
	4	----------
	5
	6	We hypothesize that a novel, under the right circumstances, can evaporate.
	7
	8	Apparatus
	9	---------
5	10
6	11	* Python 2.7.6 (probably works with older versions too)
7	12	* The `gutenberg.py` module from [gutenizer](https://github.com/okfn/gutenizer/)
8	13	* An input text (possibly from Project Gutenberg)
9	14
10		Basic Strategy
11		--------------
	15	Method
	16	------
12	17
13	18	* Collect all the sentences and count them: _s_ is the number of sentences.
14	19	* In each sentence, erase words. The probability of a word being erased
15	20	is _n_/_s_ where _n_ is the sentence number;.the first sentence is
16	21	numbered 0.
17	22
18		Sample Output
19		-------------
	23	Observations
	24	------------
20	25
21	26	When run on Voltaire's "Candide": at the beginning...
22	27

+12

-6

infix-neologisms/README.md less more

0	0	infix-neologisms
1	1	================
2	2
3		Requirements
4		------------
	3	Hypothesis
	4	----------
	5
	6	We hypothesize that new words can be formed from existing words by splitting
	7	them open and sticking a word inside.
	8
	9	Apparatus
	10	---------
5	11
6	12	* Python 2.7.6 (probably works with older versions too)
7	13	* A set of input words
8	14
9		Basic Strategy
10		--------------
	15	Method
	16	------
11	17
12	18	* Pick an input word.
13	19	* Split it into two parts. Pick another input word and insert it in

15	21	* Possibly repeat step #2.
16	22	* Output the word and repeat from step #1.
17	23
18		Sample Output
19		-------------
	24	Observations
	25	------------
20	26
21	27	Running it on `../generic-corpora/containers.txt` which I just threw together
22	28	after a few internet searches, you might get

+13

-8

naive-cut-up/README.md less more

0	0	naive-cut-up
1	1	============
2	2
3		Requirements
4		------------
	3	Hypothesis
	4	----------
	5
	6	We hypothesize that if we cut up a newspaper. We also hypothesize that
	7	we cut up a newspaper.
	8
	9	Apparatus
	10	---------
5	11
6	12	* Python 2.7.6 (probably works with older versions too)
7	13	* [Pillow](http://python-pillow.github.io/) (it might work with PIL too)
8	14	* Some scanned images of newspapers, books, etc., in PNG format, for example
9	15	obtained by [fetch-chronam](../fetch-chronam/)
10	16
11		Basic Strategy
12		--------------
	17	Method
	18	------
13	19
14	20	* Start with "blank" canvas. For simplicity, we actually use one of the
15	21	input images as the "canvas".

17	23	* Copy the image within the rectangle to a random location on the canvas.
18	24	* Repeat from step 2 until we guess we've covered the canvas.
19	25
20		Usage
21		-----
	26	### Detailed procedure ###
22	27
23	28	First, we assume some PNGs of scanned newspaper pages involving some topic
24	29	(in this example, cheese) have been obtained. (PNG format is probably not

59	64
60	65	(You may wish to use a less clumsy image viewer than Ristretto, yourself.)
61	66
62		Sample Output
63		-------------
	67	Observations
	68	------------
64	69
65	70	It may be difficult to tell in this scaled-down sample, but the result was
66	71	surprisingly thematic in its reference to cheese: