Open adl opened 10 years ago
I fully agree with the last sentence. I would postpone a support for weights to the following version of our format. (I think that a support for finite automata would be much easier and more desirable - and we should decide about that early as HOAF will be renamed to HAF if our decision is positive).
There are tons of extensions of automata, but weighted as well as probabililstic etc. are by far out of scope of our applications and intentions. I wouldn't consider them for v2 either
Should we make some room for representation of weighted omega-automata?
We could have some optional header
that somehow describes the semi-ring used for weights.
And allow initial weight to be specified with something like:
and transitions could have weights
Notice that the above suggestion is not compatible the change suggested for alternation (#2) since multiple initial states are now represented with a state formula such as
0|1
. It is also not clear to me how weights are used in alternating automata. I suspect that weights, if any, should appear in thestate-formula
syntax.None of our tools support weights, so working on this does not seem a priority at the moment.