[sugar] Develop Activity

[Andrew Clunis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, Dec 20, 2006 at 01:28:01AM +0100, Ivan Krsti? wrote:
> Eben Eliason wrote:
> > One editing session may change a dozen different source files, yet 
> > result in only one new journaled version.
> 
> The develop activity works on files directly on the actual
> filesystem, and as such, the journal won't automatically track what it
> does. That means the activity itself gets to choose what to journal, and
> we have full flexibility there. It also means the bulk of the
> activity development can be completed without worrying about the journal.
> 
> A number of people in the thread have speculated on a unification of the
> journal and the filesystem, and I'd have to think that through, but I
> get a bad vibe from it. Filesystems are well suited to things like
> source code, and I'm not really comfortable trying to shoehorn bunches
> of source in the journal, at least until we have a strong grip on how
> we're doing general grouping in the journal (and can decide whether the
> two models are compatible.)

Yeah, that's right.  Especially considering the problem I mentioned
elsewhere on this thread about versioning the version control (".bzr"
getting verisoned by journal, etc.).

I was about to say on IRC just now that "the version control, including
branching, merging, etc. should not be done by Journal, because we don't
want to reinvent Yet Another Distributed Version Control System."

Then it occured to me: maybe we do.  The whole concept of people sharing,
modifiying things they've made in activities (read: files), and merging
changes back is a problem common to all the activities that persist
works, not just Develop.

Is Journal/Docstore/Yellow doing anything about those issues yet?

Everything just got harder.  I hate that. :(
> 
> Until then, it's important to keep in mind that the journal is not a
> catch-all solution: it's a fantastically useful abstraction over the fs
> for most user documents, but source code doesn't quite match that
> definition.

Yeah, I'm still going on this mystical version in my head.  Not to nag,
but I eagerly await the Journal stuff arriving in the HIG.
> 
> -- 
> Ivan Krsti? <krstic at solarsail.hcs.harvard.edu> | GPG: 0x147C722D
>
- --
Regards,
Andrew Clunis
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQFFiPEDALkUMXSNow8RAhCcAJ9NgasJdxKeBjFfhOujHutbBrXHpwCfQkdI
Ssgeal1mF3s7qAmodCzaex8=
=8WmW
-----END PGP SIGNATURE-----