A distinct type for output register automata #69
Replies: 1 comment
-
The main issue I have with the current implementation is that a lot of code uses Given that a lot of the code seems to be based on the work of [0,1], I assume @fhowar would be the best to comment on this. As far as I can tell, there is no inherent downside to modeling (instantiated) outputs as transition outputs rather than "special" transition inputs. Quite the contrary: we may get rid of some FIXMEs and have more consistent semantics. I have added some initial drafts which also include some of the view separation discussed in #70. [0] - https://link.springer.com/chapter/10.1007/978-3-642-34026-0_41 |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Currently, RALib supports the concept of outputs via a special type of
OutputTransition
. However, as far as I can tell, the concerned register automata are still acceptors and the transduction is actually implemented as a formal language of alternating input symbols and output symbols (so a transductioni_1 i_2
->o_1 o_2
is represented as an (accepted) wordi_1 o_1 i_2 o_2
).Would it be possible/reasonable to introduce a separate type (e.g., a
RegisterMealyMachine
) that actually supports transition outputs directly? This could be interesting from a performance point of view (requiring fewer locations to represent the same transduction) and could de-clutter some of the code by getting rid of "ioMode" checks which could instead be encapsulated in separate, type-specific implementations.Beta Was this translation helpful? Give feedback.
All reactions