Remove hint.html and index.html
As per discussions on GitHub the index.html and hint.html files should be deleted in favour of README.md files (from any .text/.txt/.markdown/.md files which will be turned into README.md in an upcoming commit).
This commit is contained in:
Cody Boone Ferguson committed
1 parent
ee8a36310f
commit
77fb0f11a2
143 files changed
-28606
No files matched your search
@@ -1,308 +0,0 @@
|
||||
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
|
||||
<html>
|
||||
<head>
|
||||
<meta name="GENERATOR" content="mkd2html 2.1.5a DL=DISCOUNT">
|
||||
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=utf-8">
|
||||
<link rel="stylesheet"
|
||||
type="text/css"
|
||||
href="../hint.css">
|
||||
<title>Most likely to be awarded</title>
|
||||
</head>
|
||||
<body>
|
||||
<h1>Most likely to be awarded</h1>
|
||||
|
||||
<p>Marcin Ciura <a href="mailto:mciura@gmail.com">mciura@gmail.com</a><br>
|
||||
Twitter: @mciura</p>
|
||||
|
||||
<h2>Judges' comments:</h2>
|
||||
|
||||
<h3>To use:</h3>
|
||||
|
||||
<pre><code>make
|
||||
|
||||
./prog < text
|
||||
</code></pre>
|
||||
|
||||
<h3>Try:</h3>
|
||||
|
||||
<pre><code>./prog < Makefile
|
||||
./prog < hint.text
|
||||
</code></pre>
|
||||
|
||||
<h3>Selected Judges Remarks:</h3>
|
||||
|
||||
<p>This text was processed by prog. You may get confused.
|
||||
We’re not really sure: prog.c wasn’t commented.
|
||||
Who has been thoroughly puzzled by prog? Also how obfuscated is prog?
|
||||
Having been written in C, how large of a vocabulary has it?</p>
|
||||
|
||||
<h2>Author’s comments:</h2>
|
||||
|
||||
<h1>The Program</h1>
|
||||
|
||||
<p>“The Elements of Style” by prof. William Strunk, Jr. and E. B. White
|
||||
is well known for its dislike of the passive voice. Direct, vigorous,
|
||||
active constructions were instead recommended by the authors.
|
||||
The development of this program has been driven by this tenet.</p>
|
||||
|
||||
<p>Only passive sentences from the standard input are output. The offending
|
||||
part(s) of each sentence will be highlighted. Several kinds of passive
|
||||
constructions are recognized:</p>
|
||||
|
||||
<ul>
|
||||
<li>Plain passive voice.
|
||||
Foo was bazzed by bar.</li>
|
||||
<li>Passive voice with “get”.
|
||||
I got almost quuxed.</li>
|
||||
<li>Contractions.
|
||||
I’m not really quuxed: foo wasn’t bazzed.</li>
|
||||
<li>Questions.
|
||||
Who has been quuxed? Was foo indeed bazzed? How was it bazzed?</li>
|
||||
<li>Clause-initial inversion.
|
||||
Also bazzed is bar.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>Sentences are never broken by periods that follow abbreviations and
|
||||
initials. Both regular and irregular past participles are properly
|
||||
dealt with. Note that some words are red herrings: even though their
|
||||
ending is -ed, they are not past participles.</p>
|
||||
|
||||
<h1>The Big String</h1>
|
||||
|
||||
<p>Here is the purpose of every byte inside the <strong><code>c[]</code></strong> string:</p>
|
||||
|
||||
<ul>
|
||||
<li>0-9: highlighting passive constructions in the output;</li>
|
||||
<li>10-50: classifying characters - called by <strong><code>V(4)</code></strong>;</li>
|
||||
<li>52: outputting the newline character;</li>
|
||||
<li>51-187: tokenizing text and detecting passive constructions - two
|
||||
type A state machines, called by <strong><code>V(2)</code></strong> and <strong><code>V(3)</code></strong>;</li>
|
||||
<li>186-2786: determining word categories (past participles, present
|
||||
simple and past simple forms of “to be”, other forms of “to be”,
|
||||
adverbs, wh-words, and other words) and recognizing abbreviations -
|
||||
two type B state machines, called by <strong><code>V(808)</code></strong> and <strong><code>V(205)</code></strong>.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<h1>The State Machines</h1>
|
||||
|
||||
<p>I heard you don’t like state machines so I put a state machine in your
|
||||
state machine so you can reject this entry while you reject this entry.</p>
|
||||
|
||||
<p>Both types of state machines in the program are Mealy machines: we can
|
||||
think of their transitions as triples (<strong><code>transition_label</code></strong>,
|
||||
<strong><code>next_state</code></strong>, <strong><code>output</code></strong>).</p>
|
||||
|
||||
<h2>Type A</h2>
|
||||
|
||||
<p>A state of type A state machine consists of eight transitions whose
|
||||
representations occupy consecutive bytes. The byte
|
||||
<strong><code>c[3 + 8*state + transition_label]</code></strong> contains
|
||||
<strong><code>(unsigned char)(30 + 5*next_state + output)</code></strong>.</p>
|
||||
|
||||
<p>The <code>transition_label</code> is either one of character classes:</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>nil</code></strong> = 0,</li>
|
||||
<li><strong><code>space</code></strong> = 1,</li>
|
||||
<li><strong><code>new_line</code></strong> = 2,</li>
|
||||
<li><strong><code>uppercase</code></strong> = 3,</li>
|
||||
<li><strong><code>lowercase</code></strong> = 4,</li>
|
||||
<li><strong><code>period</code></strong> = 5,</li>
|
||||
<li><strong><code>bang_question</code></strong> = 6,</li>
|
||||
<li><strong><code>punctuation</code></strong> = 7,</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>or one of word classes:</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>adverb</code></strong> = 0,</li>
|
||||
<li><strong><code>other</code></strong> = 2,</li>
|
||||
<li><strong><code>past_participle</code></strong> = 3,</li>
|
||||
<li><strong><code>to_be</code></strong> = 4,</li>
|
||||
<li><strong><code>are_were</code></strong> = 5,</li>
|
||||
<li><strong><code>wh_word</code></strong> = 6 (cf. the <strong><code>output</code></strong> of type B state machines).</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>The <strong><code>output</code></strong> is either one of:</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>end_of_text</code></strong> = 0,</li>
|
||||
<li><strong><code>sentence_start</code></strong> = 1,</li>
|
||||
<li><strong><code>sentence_end</code></strong> = 2,</li>
|
||||
<li><strong><code>token_end</code></strong> = 3,</li>
|
||||
<li><strong><code>ignore</code></strong> = 4,</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>or one of</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>passive_construction_start</code></strong> = 1,</li>
|
||||
<li><strong><code>ignore</code></strong> = 2,</li>
|
||||
<li><strong><code>passive_construction_end</code></strong> = 3.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>Type A state machines run continuously. Their state is kept in
|
||||
<strong><code>k[2]</code></strong> and <strong><code>k[3]</code></strong>.</p>
|
||||
|
||||
<p>Below are sample steps of scanning a sequence of characters with
|
||||
classes <strong><code>uppercase</code></strong>, <strong><code>lowercase</code></strong>, <strong><code>period</code></strong>, <strong><code>space</code></strong>,
|
||||
and <strong><code>uppercase</code></strong>.</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>state</code></strong> = 7, <strong><code>transition_label</code></strong> = 3 (<strong><code>uppercase</code></strong>), <strong><code>output</code></strong> = <strong><code>sentence_start</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 14, <strong><code>transition_label</code></strong> = 4 (<strong><code>lowercase</code></strong>), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 15, <strong><code>transition_label</code></strong> = 5 (<strong><code>period</code></strong>), <strong><code>output</code></strong> = <strong><code>token_end</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 13, <strong><code>transition_label</code></strong> = 1 (<strong><code>space</code></strong>), <strong><code>output</code></strong> = <strong><code>token_end</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 18, <strong><code>transition_label</code></strong> = 3 (<strong><code>uppercase</code></strong>), <strong><code>output</code></strong> = <strong><code>sentence_end</code></strong></li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>And here are sample steps of scanning tokens with classes <strong><code>to_be</code></strong>,
|
||||
<strong><code>adverb</code></strong>, <strong><code>past_participle</code></strong>.</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>state</code></strong> = 8, <strong><code>transition_label</code></strong> = 4 (<strong><code>to_be</code></strong>), <strong><code>output</code></strong> = <strong><code>passive_construction_start</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 22, <strong><code>transition_label</code></strong> = 0 (<strong><code>adverb</code></strong>), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 22, <strong><code>transition_label</code></strong> = 3 (<strong><code>past_participle</code></strong>), <strong><code>output</code></strong> = <strong><code>passive_construction_end</code></strong></li>
|
||||
</ul>
|
||||
|
||||
|
||||
<h2>Type B</h2>
|
||||
|
||||
<p>Type B state machines are built by a variant of the algorithm described
|
||||
in the paper
|
||||
<em><a href="http://sun.aei.polsl.pl/~mciura/publikacje/lexicon.pdf">How to squeeze a lexicon</a></em>.
|
||||
Their states are represented as interlaced sparse arrays of transitions.</p>
|
||||
|
||||
<p>The <strong><code>transition_label</code></strong> is one of:</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>word_start</code></strong> = 0,</li>
|
||||
<li><strong><code>period_or_apostrophe</code></strong> = 1,</li>
|
||||
<li><strong><code>a_z_letter</code></strong> = 2..27.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>The <code>output</code> is either one of word classes:</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>adverb</code></strong> = 0,</li>
|
||||
<li><strong><code>ignore</code></strong> = 1,</li>
|
||||
<li><strong><code>other</code></strong> = 2,</li>
|
||||
<li><strong><code>past_participle</code></strong> = 3,</li>
|
||||
<li><strong><code>to_be</code></strong> = 4,</li>
|
||||
<li><strong><code>are_were</code></strong> = 5,</li>
|
||||
<li><strong><code>wh_word</code></strong> = 6,</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>or</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>ignore</code></strong> = 1,</li>
|
||||
<li><strong><code>abbreviation</code></strong> = 2.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>The byte <strong><code>c[186 + (state + transition_label)] = (unsigned char)(2 + transition_label + 28 * output)</code></strong>.</p>
|
||||
|
||||
<p>The byte <strong><code>c[836 + 2*(state + transition_label)] =</code></strong></p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>' '</code></strong> when <strong><code>next_state / 376 == 0</code></strong>,</li>
|
||||
<li><strong><code>';'</code></strong> when <strong><code>next_state / 376 == 1</code></strong>.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>The byte <strong><code>c[837 + 2*(state + transition_label)] =</code></strong></p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>' '</code></strong> when <strong><code>next_state / 94 % 4 == 0</code></strong>,</li>
|
||||
<li><strong><code>';'</code></strong> when <strong><code>next_state / 94 % 4 == 1</code></strong>,</li>
|
||||
<li><strong><code>'{'</code></strong> when <strong><code>next_state / 94 % 4 == 2</code></strong>,</li>
|
||||
<li><strong><code>'}'</code></strong> when <strong><code>next_state / 94 % 4 == 3</code></strong>.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>The byte <strong><code>c[2136 + (state + transition_label)] = (unsigned char)(32 + next_state % 94)</code></strong>.</p>
|
||||
|
||||
<p>A type B state machine starts in a given <strong><code>state</code></strong> with <strong><code>output</code></strong>
|
||||
set to <strong><code>ignore</code></strong>. It scans a word backwards, converting character by
|
||||
character to <strong><code>transition_label</code></strong>. When it meets <strong><code>word_start</code></strong>, it
|
||||
returns, leaving the most recently stored <strong><code>output</code></strong> in <strong><code>v</code></strong>.
|
||||
It also returns when
|
||||
<strong><code>c[186 + (state + transition_label)] != (unsigned char)(2 + transition_label + 28 * output)</code></strong>,
|
||||
i.e. when the <strong><code>transition_label</code></strong> in the byte mismatches the
|
||||
<strong><code>transition_label</code></strong> used to compute the index. Otherwise, if the
|
||||
<strong><code>output</code></strong> in this byte differs from <strong><code>ignore</code></strong>, the state machine
|
||||
stores it in <strong><code>v</code></strong>. Regardless of the value of the <strong><code>output</code></strong>,
|
||||
it also assigns <strong><code>next_state</code></strong> to <strong><code>state</code></strong> and proceeds to scan
|
||||
the previous character. The state machine that determines the word
|
||||
class is set up so that it never returns <strong><code>ignore</code></strong>.</p>
|
||||
|
||||
<p>Below are sample steps of scanning the word “chiefly”.</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>state</code></strong> = 623, <strong><code>transition_label</code></strong> = 26 (‘y’), <strong><code>output</code></strong> = <strong><code>other</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 602, <strong><code>transition_label</code></strong> = 13 (‘l’), <strong><code>output</code></strong> = <strong><code>adverb</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 594, <strong><code>transition_label</code></strong> = 7 (‘f’), <strong><code>output</code></strong> = <strong><code>other</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 549, <strong><code>transition_label</code></strong> = 6 (‘e’), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 543, <strong><code>transition_label</code></strong> = 10 (‘i’), <strong><code>output</code></strong> = <strong><code>adverb</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 0, <strong><code>transition_label</code></strong> = 9 (‘h’) mismatches the value stored
|
||||
in the byte. Return <strong><code>adverb</code></strong>.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>And here are sample steps of scanning the string “i.e.”:</p>
|
||||
|
||||
<ul>
|
||||
<li><strong><code>state</code></strong> = 19, <strong><code>transition_label</code></strong> = 1 (‘.’), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 24, <strong><code>transition_label</code></strong> = 6 (‘e’), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 3, <strong><code>transition_label</code></strong> = 1 (‘.’), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 2, <strong><code>transition_label</code></strong> = 10 (‘i’), <strong><code>output</code></strong> = <strong><code>ignore</code></strong></li>
|
||||
<li><strong><code>state</code></strong> = 1, <strong><code>transition_label</code></strong> = 0 (‘\0’), <strong><code>output</code></strong> = <strong><code>abbreviation</code></strong>.
|
||||
Return <strong><code>abbreviation</code></strong>.</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<h1>Remarks</h1>
|
||||
|
||||
<p>The judges found a bug in the handling of contractions: <strong><code>prog.orig.c</code></strong>
|
||||
outputs “foo wasn'<strong>[t bazzed]</strong>”. I fixed the bug in <strong><code>prog.c</code></strong>, making
|
||||
it output “foo <strong>[wasn’t bazzed]</strong>” instead, and took the liberty to merge
|
||||
one statement into a <strong><code>for</code></strong> loop inside its <strong><code>main()</code></strong> function.</p>
|
||||
|
||||
<p>The program uses a 9,437,187-byte buffer so it probably will not run
|
||||
on 16-bit and smaller machines without changing the size of array <strong><code>C[]</code></strong>
|
||||
and the expression <strong><code>9<<20</code></strong> inside the call to <strong><code>read()</code></strong>.</p>
|
||||
|
||||
<p>I have tested it on:</p>
|
||||
|
||||
<ul>
|
||||
<li>Linux with gcc version 5.4.0 20160609,</li>
|
||||
<li>Linux with clang version 3.4-1ubuntu3,</li>
|
||||
<li>OS/X with Apple LLVM version 6.0 (clang-600.0.57).</li>
|
||||
</ul>
|
||||
|
||||
|
||||
<p>For a clean compile under c89 and c90, add <strong><code>-Woverlength-strings</code></strong>
|
||||
to <strong><code>CSILENCE</code></strong>.</p>
|
||||
|
||||
<hr />
|
||||
|
||||
<TABLE><TR>
|
||||
<TD><a rel="license" href="http://creativecommons.org/licenses/by-sa/3.0/"><img alt="Creative Commons License" style="border-width:0" src="http://i.creativecommons.org/l/by-sa/3.0/88x31.png" /></a></TD>
|
||||
<TD><P>© Copyright 1984-2018,
|
||||
<A HREF="/judges.html">Leo Broukhis, Simon Cooper, Landon Curt Noll</A>
|
||||
- All rights reserved<BR>
|
||||
This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/3.0/">Creative Commons Attribution-ShareAlike 3.0 Unported License</a>.</P></TD>
|
||||
<TD> <!--<a href="http://validator.w3.org/check?uri=referer"><img src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01 Transitional" height="31" width="88"></a>--></TD>
|
||||
</TR></TABLE>
|
||||
</body>
|
||||
</html>
|
||||
Reference in new issue
Block a user