alarsyo/spot
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

rename is_deterministic to is_universal

Browse source 
For #212.

* spot/twa/twa.hh: Rename prop_deterministic() as prop_universal(),
and keep the old name as deprecated.
* spot/twaalgos/isdet.cc, spot/twaalgos/isdet.hh: Rename
is_deterministic() as is_universal(), and add a new function
for is_deterministic().
* doc/org/concepts.org, doc/org/hoa.org, doc/org/tut21.org,
spot/tl/hierarchy.cc, spot/twa/twagraph.cc,
spot/twaalgos/are_isomorphic.cc, spot/twaalgos/determinize.cc,
spot/twaalgos/dtbasat.cc, spot/twaalgos/dtwasat.cc,
spot/twaalgos/hoa.cc, spot/twaalgos/minimize.cc,
spot/twaalgos/postproc.cc, spot/twaalgos/product.cc,
spot/twaalgos/randomgraph.cc, spot/twaalgos/remfin.cc,
spot/twaalgos/simulation.cc, spot/twaalgos/totgba.cc,
spot/twaalgos/word.cc, tests/python/product.ipynb,
tests/python/remfin.py: Adjust.
* NEWS: Mention the change.
This commit is contained in:
Alexandre Duret-Lutz 2017-03-27 19:19:52 +02:00
parent 4518724a5b
commit 4a5d7a3978
24 changed files with 181 additions and 180 deletions
Download patch file Download diff file Expand all files Collapse all files

38
doc/org/concepts.org
View file

@ -1054,18 +1054,18 @@ better choices.
There are actually several property flags that are stored into each
automaton, and that can be queried or set by algorithms:
| flag name | meaning when =true= |
|----------------------+----------------------------------------------------------------------------------------------|
| =state_acc= | automaton should be considered as having state-based acceptance |
| =inherently_weak= | accepting and rejecting cycles cannot be mixed in the same SCC |
| =weak= | transitions of an SCC all belong to the same acceptance sets |
| =very-weak= | weak automaton where all SCCs have size 1 |
| =terminal= | automaton is weak, accepting SCCs are complete, accepting edges may not go to rejecting SCCs |
| =complete= | it is always possible to move the automaton forward, using any letter |
| =deterministic= | there is at most one run *recognizing* a word, but not necessarily accepting it |
| =semi-deterministic= | any nondeterminism occurs before entering an accepting SCC |
| =unambiguous= | there is at most one run *accepting* a word (but it might be recognized several time) |
| =stutter_invariant= | the property recognized by the automaton is [[https://www.lrde.epita.fr/~adl/dl/adl/michaud.15.spin.pdf][stutter-invariant]] |
| flag name | meaning when =true= |
|----------------------+-------------------------------------------------------------------------------------------------------------------------------------------|
| =state_acc= | automaton should be considered as having state-based acceptance |
| =inherently_weak= | accepting and rejecting cycles cannot be mixed in the same SCC |
| =weak= | transitions of an SCC all belong to the same acceptance sets |
| =very-weak= | weak automaton where all SCCs have size 1 |
| =terminal= | automaton is weak, accepting SCCs are complete, accepting edges may not go to rejecting SCCs |
| =complete= | it is always possible to move the automaton forward, using any letter |
| =universal= | there is no non-determinism branching in the automaton (hence each word is *recognized* by at most one run, but not necessarily accepted) |
| =semi-deterministic= | any nondeterminism occurs before entering an accepting SCC |
| =unambiguous= | there is at most one run *accepting* a word (but it might be recognized several time) |
| =stutter_invariant= | the property recognized by the automaton is [[https://www.lrde.epita.fr/~adl/dl/adl/michaud.15.spin.pdf][stutter-invariant]] |
For each flag =flagname=, the =twa= class has a method
=prop_flagname()= that returns the value of the flag as an instance of
@ -1083,12 +1083,14 @@ operator =X= does not prevent the formula from being
stutter-invariant, but it would require additional work to check.
As another example, if you write an algorithm that must check whether
an automaton is deterministic, do not call the
=twa::prop_deterministic()= method, because that might return
=trival::maybe=. Instead, call =spot::is_deterministic(...)=: that
will respond in constant time if the =deterministic= property flag was
either =true= or =false=, otherwise it will actually explore the
automaton to decide its determinism.
an automaton is universal, do not call the =twa::prop_universal()=
method, because that might return =trival::maybe=. Instead, call
=spot::is_universal(...)=: that will respond in constant time if the
=universal= property flag was either =true= or =false=, otherwise it
will actually explore the automaton to decide its determinism. Note
that there is also a =spot::is_deterministic(...)= function, which is
equivalent to testing that the automaton is both universal and
existential.
These automata properties are encoded into the [[file:hoa.org::#property-bits][HOA format]], so they can
be preserved when building a processing pipeline using the shell.