[Sugar-devel] [ANNOUNCE] Groupthink 0.1 pre-alpha

[Bobby Powers]

I stumbled across this google tech talk today on the same subject:
http://www.youtube.com/watch?v=S2Hp_1jqpY8

I haven't watched most of it yet, but thought others might be
interested as well.

bobby

On Thu, Jan 15, 2009 at 3:02 AM, Martin Edmund Sevior
<martines at unimelb.edu.au> wrote:
>
> Hi Benjamin,
>             Thanks very much! I'm very interested in looking at the source
> code and data structures.
>
> These kind of problems are exactly what abicollab (as used by Write) solves.
>
> I have given some detailed presentations about this. You can find one here:
>
> (the ogg version)
>
> http://mirror.linux.org.au/pub/linux.conf.au/2008/Wed/mel8-083.ogg
>
> (presentation odp)
>
> http://mirror.linux.org.au/pub/linux.conf.au/2008/slides/083-AbiCollab.odp
>
> and wiki write up:
>
> http://www.abisource.com/wiki/AbiCollab
>
> Maybe these ideas can help.
>
> Cheers!
>
> Martin
>
>
> -----Original Message-----
> From: sugar-devel-bounces at lists.sugarlabs.org on behalf of Benjamin M.
> Schwartz
> Sent: Thu 1/15/2009 3:25 PM
> To: Martin Edmund Sevior
> Cc: bens at alum.mit.edu; Sugar Devel; Chris Ball
> Subject: Re: [Sugar-devel] [ANNOUNCE] Groupthink 0.1 pre-alpha
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Martin Edmund Sevior wrote:
>>          How do you solve "internet lag".
>>
>> User A puts a character "A" in position 10, then before user B see sees
> this (because of the finite propagation time), he puts character "B" in
> position 10?
>>
>> Who wins? You just have to make sure the the document remains the same
> for both users.
>
> This is, indeed, the central problem.  For the moment, the answer is:
> Groupthink does not support full documents, only short snippets of text.
> I have prototyped a data structure that I believe can coherently resolve
> these sorts of edit conflicts in long documents without any negotiation,
> but it remains to be seen if the design will work.
>
> In general, Groupthink's approach (described at length in docstrings in
> the code) is to write each data structure in such a way that any two users
> who have observed the same set of messages will arrive at the same state,
> regardless of the order in which those messages are received.  The hard
> part is figuring out how to do this for each kind of data structure.
> However, once it is working, the code can be reused for many purposes
> without needing to understand how it works.
>
> - --Ben
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.9 (GNU/Linux)
>
> iEYEARECAAYFAkluuskACgkQUJT6e6HFtqQb2QCfXnSJHwZpn6Q/OsDSs8nJEb3x
> vWIAoIJYuSuEp8EFOeKynHDLctW2S5iq
> =Po5R
> -----END PGP SIGNATURE-----
> _______________________________________________
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel
>
>
> _______________________________________________
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel
>
>