Camelia, the Perl 6 bug

IRC log for #darcs-theory, 2008-10-02

| Channels | #darcs-theory index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:53 Igloo_ joined #darcs-theory
01:57 rwbarton joined #darcs-theory
08:02 kowey joined #darcs-theory
15:06 kowey joined #darcs-theory
17:19 arjanb joined #darcs-theory
18:55 rwbarton This might be a little off-topic, but...
18:56 rwbarton Is there any general theory for taking an algebraic data type F and producing a notion of "patches" for values of type F a in terms of patches for a?
18:56 rwbarton I'm thinking of something like the derivative, but a bit different
18:58 Igloo I didn't follow that. Is F a type or a type constructor?
18:59 Igloo Are you talking about fmap?
18:59 rwbarton F is a type constructor, say  data F a = Pair (a, a)
18:59 rwbarton I'd like to say that a patch to an F a is either a patch on the first half of the pair, or a patch on the second half of the pair,
19:00 rwbarton and that those two kinds of patches always commute.
19:00 Igloo I'm not aware of anything
19:01 rwbarton does darcs have a general notion that it applies to both files-in-a-directory and bytes-in-a-file?
19:01 rwbarton (Not that either of those is necessarily well-represented by an algebraic data type)
19:01 Igloo darcs primitive patch commute is written by hand
19:02 rwbarton Is there somewhere in the darcs source I can see what kinds of patches exist?
19:04 Igloo src/Darcs/Patch/Prim.lhs
19:06 rwbarton Ah, great
19:20 arjanb there's a draft paper trying to describe patches in terms of three sets: http://people.cs.uu.nl/andres/VersionControl.html
21:25 kowey joined #darcs-theory
21:26 lispy rwbarton: src/Darcs/Patch/Patchy.lhs is also good
21:27 lispy rwbarton: Prim tells you which kinds of patches the UI supports, and other places define type classes which can be instanced for different Haskell types of patches
21:27 lispy rwbarton: I'm not 100% sure which meaning of type you have there :)
21:32 rwbarton lispy: I think I managed to say "kind" everywhere I didn't mean type-as-in-Haskell. :)
21:33 rwbarton lispy: And yeah, I took a look at Patchy.lhs too.
21:35 rwbarton I'm curious about the claim on line Prim.lhs:807 that a merge cannot fail, when as far as I can tell merge (on Prim) roughly reduces to a commute that certainly could fail.
21:37 rwbarton (But it's not too important to me.)
21:38 Igloo Merging prim patches can fail
21:38 Igloo What are you doing, OOI?
21:39 rwbarton My ultimate project is some kind of Google Document-like application for real-time collaboration
21:39 Igloo Ah, cool
21:39 Igloo Are you coding from scratch?
21:40 rwbarton I will be once there is some architecture :)
21:40 rwbarton (It'll probably use an off-the-shelf messaging server)
21:41 Igloo What will you do if you hit a conflict?
21:42 Igloo (Apart from being generally interested, I'm wondering if it would make sense for you to use part of camp as a library)
21:42 rwbarton That's a good question.  The hope is that we'll have low latency so conflicts will be small and users won't mind possibly losing a few changes.
21:42 rwbarton i.e., we'll probably drop one of the patches and notify the user in a non-intrusive way.
21:42 rwbarton camp?
21:43 Igloo http://projects.haskell.org/camp/
21:43 rwbarton Ah, very interesting.
21:57 rwbarton Anyways, I can let you know if we ever have a demo or something.  (It's something some of my friends are working on in our spare time, so it may not actually happen.)

| Channels | #darcs-theory index | Today | | Search | Google Search | Plain-Text | summary