On Tue, Aug 4, 2009 at 1:06 PM, Howard Lewis Ship<hls...@gmail.com> wrote: > > On Tue, Aug 4, 2009 at 10:03 AM, Rich Hickey<richhic...@gmail.com> wrote: >> >> >> >> On Aug 4, 12:35 pm, Howard Lewis Ship <hls...@gmail.com> wrote: >>> These aren't criticisms, just trying to understand this better. >>> >>> So assoc! etc. return a new instance of the transient? >> >> They return the next value of the transient. >> >>> Or do they >>> return the same transient? >> >> I'm not promising that they do. >> >>> If they return a new instance, what >>> exactly is the advantage (though, obviously there is one, from the >>> transcript). >> >> Right. >> >>> If they return the same instance do you prevent the >>> calling code from invoking further changes or not? >>> >> >> This one I don't understand, you can use the return value, don't reuse >> the argument. If your code is structured functionally and non- >> persistently that is how it will be anyway. >> >>> I like the idea of restricting the transient to just the originating thread. >>> >>> Does this mean we'll need a reduce! >> >> No, reduce farms out the work to the reducing function and doesn't >> create a composite data structure itself. The reducing fn might use a >> transient but that's an implementation detail of it. >> >>> and map!, etc? >> >> No, map returns a sequence. >> >>> Would there be an >>> advantage to those? Is there an intersection (or potential danger) >>> between transient structures and laziness? >>> >> >> Transient data structures should never be part of the public interface >> of anything. They are just an implementation detail, an optimization, >> e.g. into/vec/vector now use transients, but aren't now into!/vec!/ >> vector! - the transients are inside. >> >> Hope that helps, >> >> Rich > > > It does, pretty much what I expected. Care to summarize why they are > so much faster than persistent types (for those of us too lazy/busy to > read the source)? >
Fully persistent modifications require full path copies, while transient modifications copy nodes as needed, and once a copy has been made, subsequent modifications affecting that node can be made in place. Rich --~--~---------~--~----~------------~-------~--~----~ 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 your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~----------~----~----~----~------~----~------~--~---
