/usr/share/doc/prover9-doc/html/output.html is in prover9-doc 0.0.200902a-2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 | <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN">
<html>
<head>
<title>Prover9 Manual: Output Files</title>
<link rel="stylesheet" href="manual.css">
</head>
<body>
<hr>
<table width="100%">
<tr>
<colgroup>
<col width="33%">
<col width="34%">
<col width="33%">
</colgroup>
<td align="left"><i>Prover9 Manual</i>
<td align="center"><img src="prover9-5a-256t.gif">
<td align="right"><i>Version 2009-02A</i>
</table>
<hr>
<!-- Main content -->
<h1>Output Files</h1>
Even when Prover9 fails to find a proof, its output file
usually has lots of valuable information about the search.
The output file can suggest many ways of improving the
search for subsequent jobs as in the following examples.
<ul>
<li>The output shows how equalities are oriented; different
<a href="term-order.html">term ordering parameters</a>
may give better or more intuitive orientations.
<li>If Prover9 focused the search on uninteresting clauses
(see the sequence of given clauses), different
<a href="inf-rules.html">inference rules</a>, a different
<a href="select.html#pick_given_ratio"><tt><b>pick_given_ratio</b></tt></a>, or a specialized
<a href="weight.html">weighting function</a>
can be used.
<li>If Prover9 ran out of time or memory with a huge <tt>sos</tt> list
and small <tt>usable</tt> list (i.e., few given clauses were used),
the <a href="limits.html#sos_limit"><tt><b>sos_limit</b></tt></a> should be reduced.
</ul>
<h2>Basic Structure of Output Files</h2>
Prover9 output files are divided into sections and subsections
so the users (people and programs) can find what they are looking
for. The delimiters are self-explanatory. A few comments about
the sections are given here. For a specific example, see
the output file <a href="../examples/subset_trans.out">subset_trans.out</a>.
<pre class="my_file">
============================== Prover9 ===============================
<font color="#D10000"><i>Version, date, host computer, command</a>.</i></font>
============================== end of head ===========================
============================== INPUT =================================
<font color="#D10000"><i>Echo of the input. Everything in this section that is not
in the input is commented with "%", so copy-and-paste can be
done on this section to create a new input file.</i></font>
============================== end of input ==========================
============================== PROCESS GOALS =========================
<font color="#D10000"><i>The search is always by refutation, and this section shows
how goals are negated in preparation for the search.</i></font>
============================== end of process goals ==================
============================== PROCESS INITIAL CLAUSES ===============
<font color="#D10000"><i>This section shows the starting clauses (after Skomemization,
if applicable) and then some of what Prover9 does in preparation
for the search. This includes <a href="more-prep.html#predicate_elim"><tt><b>predicate_elim</b></tt></a>, term ordering
decisions, and <a href="auto.html#auto_inference"><tt><b>auto_inference</b></tt></a> settings. At this stage, clauses
may be deleted by subsumption and equations may be copied to the
list <tt>demodulators</tt>. See the flag <a href="more-prep.html#process_initial_sos"><tt><b>process_initial_sos</b></tt></a>.</i></font>
============================== end of process initial clauses ========
============================== CLAUSES FOR SEARCH ====================
<font color="#D10000"><i>This section shows the clauses just before the start of the
search, that is, just before selection of the first given clause.</i></font>
============================== end of clauses for search =============
============================== SEARCH ================================
<font color="#D10000"><i>This section typically shows the sequence of given clauses,
and it may also include PROOF and STATISTICS sections.</i></font>
============================== PROOF =================================
<font color="#D10000"><i>A proof in standard form.</i></font>
============================== end of proof ==========================
============================== STATISTICS ============================
<font color="#D10000"><i>We encourage users to look at statistics!</i></font>
============================== end of statistics =====================
============================== end of search =========================
</pre>
<a name="just">
<h2>Clause Justifications</h2>
After the initial stage of the output, each clause in the file has an
integer identifier (ID) and a justification that may refer to IDs
of other clauses. A justification is a list consisting of one
primary step and some number of secondary steps. Most primary steps
are inference rules applied to given clauses, and most secondary
steps consist of simplification, rewriting, or orienting equalities.
<p>
Many of the types of step refer to positions of literals or terms
in the parent clauses. Literals are identified by the characters
'a' (first literal), 'b' (second literal), etc. Terms are identified
by the literal identifier followed by a sequence of integers giving
the position of the term within the literal. For example, the
position 'c,1,3,2' means third literal, first argument, third
argument, second argument. Negation signs on literals are not
included in the sequence.
<p>
Primary Steps.
<ul>
<li><tt>assumption</tt> -- input formula.
<li><tt>clausify</tt> -- from CNF translation of a non-clausal assumption.
<li><tt>goal</tt> -- input formula.
<li><tt>deny</tt> -- from CNF translation of the negation of a goal.
<li><tt>resolve(59,b,47,c)</tt> -- resolve the second literal of clause 59
with the third literal of clause 47.
<li><tt>hyper(59, b,47,a, c,38,a)</tt> -- hyperresolution; interpret the
list as a clause ID followed by a sequence of triples,
<literal,clause-ID,literal> the inference is presented as a sequence
of binary resolution steps. In the example shown, start with clause 59;
then resolve literal b with clause 47 on literal a; with the result of
the first step, resolve literal c with clause 38 on literal a.
The special case "<tt>xx</tt>" means resolution with <tt>x=x</tt>.
<li><tt>ur(39, a,48,a, b,88,a, c,87,a, d,86,a)</tt> -- unit-resulting
resolution; the list is interpreted as in hyperresolution.
<li><tt>para(47(a,1),28(a,1,2,2,1))</tt> -- paramodulate from the clause 47
into clause 28 at the positions shown.
<li><tt>copy(59)</tt> -- copy clause 59.
<li><tt>back_rewite(59)</tt> -- copy clause 59.
<li><tt>back_unit_del(59)</tt> -- copy clause 59.
<li><tt>new_symbol(59)</tt> -- introduce a new constant (see parameter
<a href="inf-rules.html#new_constants"><tt><b>new_constants</b></tt></a>).
<li><tt>factor(59,b,c)</tt> -- factor clause 59 by unifying the second
and third literals.
<li><tt>xx_res(59,b)</tt> -- resolve the second literal of clause 59 with
<tt>x=x</tt>.
<li><tt>propositional</tt> -- not used in standard proofs.
<li><tt>instantiate</tt> -- not used in standard proofs.
<li><tt>ivy</tt> -- not used in standard proofs.
</ul>
<p>
Secondary Steps (each assumes a working clause, which is either the result of a
primary step or a previous secondary step).
<ul>
<li><tt>rewrite([38(5,R),47(5),59(6,R)])</tt> -- rewriting (demodulation)
with equations
38, 47, then 59; the arguments (5), (5), and (6) identify the positions
of the rewritten subterms (in an obscure way), and the argument R
indicates that the demodulator is used backward (right-to-left).
<li><tt>flip(c)</tt> -- the third literal is an equality that has been
flipped by the term ordering. This does not necessarily mean that
the equality is orientable by the primary term ordering, e.g., KBO.
<li><tt>merge(d)</tt> -- the fourth literal has been removed because it
was identical to a preceding literal.
<li><tt>unit_del(b,38)</tt> -- the second literal has been removed because
it was an instance of the negation clause 38 (which is a unit clause).
<li><tt>xx(b)</tt> -- the second literal has been removed because it was
an instance of <tt>x!=x</tt>.
</ul>
<h2>Standard Proofs</h2>
Prover9 proofs may be transformed by separate programs,
e.g., by <a href="prooftrans.html">Prooftrans</a>.
<h2>Options That Say What Goes To the Output File </h2>
<!-- start option echo_input -->
<a name="echo_input">
<pre class="my_option">
set(echo_input). % default set
clear(echo_input).
</pre>
<blockquote>
Clearing this flag suppresses printing of clauses, formulas,
weighting rules (and everything else that ends with <tt>end_of_list</tt>)
that would ordinarily appear in the <tt>INPUT</tt> section of the
output file.
</blockquote>
<!-- end option -->
<!-- start option quiet -->
<a name="quiet">
<pre class="my_option">
set(quiet).
clear(quiet). % default clear
</pre>
<blockquote>
Setting this flag causes most messages to the standard error file
(usually the user's screen) to be suppressed.
These messages include
notifications about proofs and statistics reports, and
warnings about demodulation limits.
Setting this flag also suppresses several messages to the ordinary
output file, and it clears the <a href="output.html#bell"><tt><b>bell</b></tt></a> flag.
</blockquote>
<!-- end option -->
<!-- start option print_initial_clauses -->
<a name="print_initial_clauses">
<pre class="my_option">
set(print_initial_clauses). % default set
clear(print_initial_clauses).
</pre>
<blockquote>
If this flag is set, clauses are printed in the
<tt>PROCESS INITIAL CLAUSES</tt> and
<tt>CLAUSES FOR SEARCH</tt> sections of the output file.
</blockquote>
<!-- end option -->
<!-- start option print_given -->
<a name="print_given">
<pre class="my_option">
set(print_given). % default set
clear(print_given).
</pre>
<blockquote>
Clearing this flag prevents given clauses from being printed to
the output file.
</blockquote>
<!-- end option -->
<!-- start option print_gen -->
<a name="print_gen">
<pre class="my_option">
set(print_gen).
clear(print_gen). % default clear
</pre>
<blockquote>
Setting this flag causes all generated clauses to be printed to
the the output file. In addition, some other information about the
processing of each generated clause is printed.
This flag can be output files to be really huge.
</blockquote>
<!-- end option -->
<!-- start option print_kept -->
<a name="print_kept">
<pre class="my_option">
set(print_kept).
clear(print_kept). % default clear
</pre>
<blockquote>
Setting this flag causes all kept clauses to be printed to the the output file.
In addition, some other information on the processing of kept clauses is
printed.
</blockquote>
<!-- end option -->
<!-- start option print_labeled -->
<a name="print_labeled">
<pre class="my_option">
set(print_labeled).
clear(print_labeled). % default clear
</pre>
<blockquote>
Setting this flag causes kept clauses containing label attributes
to be printed, even when the flag <a href="output.html#print_kept"><tt><b>print_kept</b></tt></a> is clear.
This flag is useful when using <a href="hints.html">the hints strategy</a>,
because when a clause matches a hint containing a label, the
label is copied to the clause. That is, clauses matching labeled hints
will be printed.
</blockquote>
<!-- end option -->
<!-- start option print_clause_properties -->
<a name="print_clause_properties">
<pre class="my_option">
set(print_clause_properties).
clear(print_clause_properties). % default clear
</pre>
<blockquote>
Setting this flag causes several properties of clauses to
be printed as "props" attributes on the clauses. The properties
include which literals are maximal (counting from 1), which
literals are maximal among literals of the same sign, and which
literals are selected for application of inference rules.
</blockquote>
<!-- end option -->
<!-- start option print_proofs -->
<a name="print_proofs">
<pre class="my_option">
set(print_proofs). % default set
clear(print_proofs).
</pre>
<blockquote>
Clearing this flag prevents proofs from being printed to
the output file. The proof message still goes to the standard
error file (usually the user's screen), unless the flag
<a href="output.html#quiet"><tt><b>quiet</b></tt></a> has been set.
</blockquote>
<!-- end option -->
<!-- start option default_output -->
<a name="default_output">
<pre class="my_option">
set(default_output). % default set
clear(default_output).
</pre>
<blockquote>
Setting this flag restores most of the output flags and parameters
to their default values.
Clearing this flag does nothing.
</blockquote>
<!-- end option -->
<!-- start option report -->
<a name="report">
<pre class="my_option">
assign(report, <i>n</i>). % default <i>n</i>=-1, range [-1 .. <tt>INT_MAX</tt>]
</pre>
<blockquote>
If <i>n</i> > 0, statistics are sent to the output
file approximately every <i>n</i> seconds.
(On Unix-like systems, one can also tell Prover9 to print statistics to the
output file by sending the signal <tt>USR1</tt> to a
running Prover9 process, e.g., <tt>kill -USR1 4223</tt>.)
</blockquote>
<!-- end option -->
<!-- start option stats -->
<a name="stats">
<pre class="my_option">
assign(stats, <i>string</i>). % default <i>string</i>=lots, range [none,some,lots,all]
</pre>
<blockquote>
This parameter determines how many statistics are sent to the output file.
</blockquote>
<!-- end option -->
<!-- start option clocks -->
<a name="clocks">
<pre class="my_option">
set(clocks).
clear(clocks). % default clear
</pre>
<blockquote>
If this flag is set, various operations during the Prover9 job
are timed (e.g., inference, demodulation, and subsumption),
and timing reports are sent to the output file.
<p>
Timing the operations can be expensive, especially
in Solaris and Macintosh systems. On Linux systems,
<tt>set(clocks)</tt> typically adds 5% -- 10% to the run time.
</blockquote>
<!-- end option -->
<!-- start option bell -->
<a name="bell">
<pre class="my_option">
set(bell). % default set
clear(bell).
</pre>
<blockquote>
If this flag is set, Prover9 beeps when important things happen,
such as proofs and warnings. Some users run searches
that find hundreds of proofs, and they clear this flag
to prevent all of the beeping.
</blockquote>
<!-- end option -->
<hr>
Next Section:
<a href="weight.html">Weighting</a>
</body>
</html>
|