Thanks. Two months ago I bought that book. So I'll take a look at it !
--
You received this message because you are subscribed to the Google
Groups Clojure group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with
I think it is a bit dull that for every combination [type1 type2] you have
to define an equivalent method for [type2 type1]
Is there a way to address this problem? perhaps with sets?
i need to think about this one :D
On Monday, November 19, 2012 4:32:15 PM UTC+1, Brian Marick wrote:
Here's
On Nov 20, 2012, at 10:22 AM, Thomas Goossens wrote:
I think it is a bit dull that for every combination [type1 type2] you have to
define an equivalent method for [type2 type1]
Is there a way to address this problem? perhaps with sets?
According to http://clojure.org/multimethods only
Then probably something like this will help me out
(defn defmutualmethods [multifn dispatch-val fn-tail]
(defmethod multifn dispatch-val fn-tail)
(defmethod multifn (reverse dispatch-val) fn-tail))
Nevertheless i'm making things more complex now because it does two things
now! But it
In a java project last year we had to create a board game.
A Board had Piece objects.
Some pieces can share position. Some don't.
The way i solved this in Java was using double dispatch.
http://pastebin.com/brrRtqsS
(I just extracted the methods out of their respective classes)
But i think it
A long time ago I posted some macros to do double-dispatch with protocols
and records. The link was http://paste.lisp.org/+2023 but it no longer
works.
The basic idea is you dispatch on the first argument to an intermediate
type that then dispatches on the second argument. It's complicated,
Here's an example of using multimethods. (Note: I overrode the Clojure
defmulti/defmethod macros with my own that play more nicely with the repl, and
for other purposes. It's a straightforward translation.) You can see all the
code here:
Yes indeed. with multimethods you can doue multi-argument dispatch.
But that means that for every possible collision. I have to specify a
method?
Would it be a good idea to introduce taxonomies here?
On Monday, November 19, 2012 4:02:18 PM UTC+1, Stuart Sierra wrote:
A long time ago I
for what its worth, I'm building a highly polymorphic board-game engine
[1] and I've stayed away from multimethods...I've used every single
polymorphism capabillity that clojure provides (records/protocols,
map-based, HOFs etc) except multi-methods. Performance is acritical
matter for this
On Nov 19, 2012, at 9:32 AM, Thomas Goossens wrote:
Yes indeed. with multimethods you can doue multi-argument dispatch.
But that means that for every possible collision. I have to specify a method?
If you read along in `asteroids.clj` file I showed, you'll see an example of a
taxonomy. A
10 matches
Mail list logo