Yeah, I found that one. I'll still go for the Java implementation to start with (also makes it easier to put in clj-ds). Then gvec as a starting point for a Clojure version (if I ever get there :). One thing I was wondering about: there would be quite a bit of code I could reuse verbatim from gvec - but I obviously don't want to copy it. But I'm unsure how to share that code with gvec and still have full platform speed.
/Karl On 13/11/2011, at 18.48, David Nolen wrote: > You should definitely look at gvec.clj. It's a PersistentVector > implementation in Clojure but for the various primitive types. > > David > > On Sun, Nov 13, 2011 at 12:38 PM, Karl Krukow <[email protected]> wrote: > > On 13/11/2011, at 18.14, David Nolen wrote: > >> This would be awesome. From his talk it sounds like it shouldn't replace >> PersistentVector at all, in fact you should be able to share structure with >> PersistentVector right? >> >> David >> > > Yes, I think so. If I understand correctly, one cool thing is that the set of > valid Clojure PersistentVector structures is actually a subset of the RRB > structures (hence the word "relaxed"). This means that without breaking the > contract of IPersistentVector one could transparently move from a Clojure > vector to a RRB vector, say when concat'ed with another large vector. I > haven't thought this completely through, but I think it would work. > > I'm, working at it right now, actually. I decided to go with a Java impl for > round one. This is easier for me (yes, easy not simple :) since I've studied > the persistent data structures in Java before, and I'm still not sure what > the best way to implement low-level stuff like this is in Clojure. > > Once I get my head around the details of the algorithm, and I have a working > Java version, I'll try and port it to Clojure (and I'll probably ask here for > advice). > > /Karl > > >> On Sun, Nov 13, 2011 at 10:02 AM, Krukow <[email protected]> wrote: >> By the way, I don''t see it as a replacement for PersistentVector, but >> rather as an additional data structure, that could perhaps be put >> somewhere in contrib if it turns out well. >> >> /Karl >> >> On Nov 13, 9:54 am, Karl Krukow <[email protected]> wrote: >> > Hi, >> > Yes, Phil Bagwell presented that at the Conj. >> > >> > I'm really interested in this stuff, so as an exercise in Clojure, I've >> > started thinking about how to implement it. At first I'll try and do it >> > without looking at Scala - if I get stuck I might peek a bit :) >> > >> > If you're interested, I'll put it online once it is in a state that is >> > fairly reasonable. >> > >> > /Karl >> > On 12/11/2011, at 21.18, bernardH wrote: >> > >> > >> > >> > >> > >> > >> > >> > > hi, >> > >> > > I just stumbled upon a paper on en enhanced immutable data structure >> > > [0] that could be useful for Clojure implementation. >> > >> > > Could it enable batter parallel programming semantics as discussed by >> > > Guy Steel in this presentation [1] ? >> > >> > > I'd have to look deeper into it, but I'd be very interested in what >> > > other Clojurian think of it. >> > >> > > Best Regards, >> > >> > > B. >> > >> > > [0]http://infoscience.epfl.ch/record/169879/files/RMTrees.pdf >> > >> > > Abstract. Immutable vectors are a convenient data structure for func- >> > > tional programming and part of the standard library of modern >> > > languages >> > > like Clojure and Scala. The common implementation is based on wide >> > > trees with a fixed number of children per node, which allows fast in- >> > > dexed lookup and update operations. In this paper we extend the vector >> > > data type with a new underlying data structure, Relaxed Radix Balanced >> > > Trees (RRB-Trees), and show how this structure allows immutable vector >> > > concatenation, insert-at and splits in O(logN ) time while maintaining >> > > the >> > > index, update and iteration speeds of the original vector data >> > > structure. >> > >> > > [1]http://research.sun.com/projects/plrg/Publications/ICFPAugust2009Stee... >> > >> > > -- >> > > You received this message because you are subscribed to the Google >> > > Groups "Clojure" group. >> > > To post to this group, send email to [email protected] >> > > Note that posts from new members are moderated - please be patient with >> > > your first post. >> > > To unsubscribe from this group, send email to >> > > [email protected] >> > > For more options, visit this group at >> > >http://groups.google.com/group/clojure?hl=en >> >> -- >> You received this message because you are subscribed to the Google >> Groups "Clojure" group. >> To post to this group, send email to [email protected] >> Note that posts from new members are moderated - please be patient with your >> first post. >> To unsubscribe from this group, send email to >> [email protected] >> For more options, visit this group at >> http://groups.google.com/group/clojure?hl=en >> >> >> -- >> You received this message because you are subscribed to the Google >> Groups "Clojure" group. >> To post to this group, send email to [email protected] >> Note that posts from new members are moderated - please be patient with your >> first post. >> To unsubscribe from this group, send email to >> [email protected] >> For more options, visit this group at >> http://groups.google.com/group/clojure?hl=en > > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to [email protected] > Note that posts from new members are moderated - please be patient with your > first post. > To unsubscribe from this group, send email to > [email protected] > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to [email protected] > Note that posts from new members are moderated - please be patient with your > first post. > To unsubscribe from this group, send email to > [email protected] > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to [email protected] Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/clojure?hl=en
