Portability | portable |
---|---|
Stability | provisional |
Maintainer | nilsson@cs.yale.edu |
Safe Haskell | Safe-Inferred |
Definition of Yampa Event type.
Note on naming conventions used in this module.
Names here might have to be rethought. It's really a bit messy.
In general, the aim has been short and convenient names (like tag
,
attach
, lMerge
) and thus we have tried to stay away from suffixing/
prefixing conventions. E.g. Event
as a common suffix would be very
verbose.
However, part of the names come from a desire to stay close to similar
functions for the Maybe type. e.g. event
, fromEvent
, isEvent
.
In many cases, this use of Event
can could understood to refer to the
constructor Event
, not to the type name Event
. Thus this use of
event should not be seen as a suffixing-with-type-name convention. But
that is obviously not easy to see, and, more over, interpreting Event
as the name of the type might make equally good or better sense. E.g.
fromEvent
can also be seen as a function taking an event signal,
which is a partial function on time, to a normal signal. The latter is
then undefined when the source event function is undefined.
In other cases, it has been necessary to somehow stay out of the way of
names used by the prelude or other commonly imported modules/modules
which could be expected to be used heavily in Yampa code. In those cases
a suffix E
have been added. Examples are filterE
(exists in Prelude)
and joinE
(exists in Monad). Maybe the suffix isn't necessary in the
last case.
Some functions (actually only one currently, mapFilterE
) have got an E
suffix just because they're closely related (by name or semantics) to one
which already has an E
suffix. Another candidate would be splitE
to
complement joinE
. But events carrying pairs could obviously have other
sources than a joinE
, so currently it is called split
.
2003-05-19: Actually, have now changed to splitE
to avoid a clash
with the method split
in the class RandomGen.
2003-05-19: What about gate
? Stands out compared to e.g. filterE
.
Currently the E
suffix is considered an exception. Maybe we should use
completely different names to avoid the E
suffix. If the functions
are not used that often, Event
might be approriate. Alternatively the
suffix E
should be adopted globaly (except if the name already contains
event
in some form?).
Arguably, having both a type Event
and a constructor Event
is confusing
since there are more than one constructor. But the name Event
for the
constructor is quite apt. It's really the type name that is wrong. But
no one has found a better name, and changing it would be a really major
undertaking. Yes, the constructor Event
is not exported, but we still
need to talk conceptually about them. On the other hand, if we consider
Event-signals as partial functions on time, maybe it isn't so confusing:
they just don't have a value between events, so NoEvent
does not really
exist conceptually.
ToDo:
- Either: reveal NoEvent and Event
or: introcuce 'event = Event', call what's now event
fromEvent
,
and call what's now called fromEvent
something else, like
unsafeFromEvent
??? Better, dump it! After all, using current
names, 'fromEvent = event undefined'!
- data Event a
- noEvent :: Event a
- noEventFst :: (Event a, b) -> (Event c, b)
- noEventSnd :: (a, Event b) -> (a, Event c)
- maybeToEvent :: Maybe a -> Event a
- event :: a -> (b -> a) -> Event b -> a
- fromEvent :: Event a -> a
- isEvent :: Event a -> Bool
- isNoEvent :: Event a -> Bool
- tag :: Event a -> b -> Event b
- tagWith :: b -> Event a -> Event b
- attach :: Event a -> b -> Event (a, b)
- lMerge :: Event a -> Event a -> Event a
- rMerge :: Event a -> Event a -> Event a
- merge :: Event a -> Event a -> Event a
- mergeBy :: (a -> a -> a) -> Event a -> Event a -> Event a
- mapMerge :: (a -> c) -> (b -> c) -> (a -> b -> c) -> Event a -> Event b -> Event c
- mergeEvents :: [Event a] -> Event a
- catEvents :: [Event a] -> Event [a]
- joinE :: Event a -> Event b -> Event (a, b)
- splitE :: Event (a, b) -> (Event a, Event b)
- filterE :: (a -> Bool) -> Event a -> Event a
- mapFilterE :: (a -> Maybe b) -> Event a -> Event b
- gate :: Event a -> Bool -> Event a
Documentation
A single possible event occurrence, that is, a value that may or may not occur. Events are used to represent values that are not produced continuously, such as mouse clicks (only produced when the mouse is clicked, as opposed to mouse positions, which are always defined).
Make the NoEvent constructor available. Useful e.g. for initialization, ((-->) & friends), and it's easily available anyway (e.g. mergeEvents []).
noEventFst :: (Event a, b) -> (Event c, b)Source
Suppress any event in the first component of a pair.
noEventSnd :: (a, Event b) -> (a, Event c)Source
Suppress any event in the second component of a pair.
maybeToEvent :: Maybe a -> Event aSource
tag :: Event a -> b -> Event bSource
Tags an (occurring) event with a value (replacing the old value).
attach :: Event a -> b -> Event (a, b)Source
Attaches an extra value to the value of an occurring event.
lMerge :: Event a -> Event a -> Event aSource
Left-biased event merge (always prefer left event, if present).
rMerge :: Event a -> Event a -> Event aSource
Right-biased event merge (always prefer right event, if present).
merge :: Event a -> Event a -> Event aSource
Unbiased event merge: simultaneous occurrence is an error.
mergeBy :: (a -> a -> a) -> Event a -> Event a -> Event aSource
Event merge parameterized by a conflict resolution function.
mergeEvents :: [Event a] -> Event aSource
Merge a list of events; foremost event has priority.
joinE :: Event a -> Event b -> Event (a, b)Source
Join (conjunction) of two events. Only produces an event if both events exist.
filterE :: (a -> Bool) -> Event a -> Event aSource
Filter out events that don't satisfy some predicate.
mapFilterE :: (a -> Maybe b) -> Event a -> Event bSource