New Account Types and data equivalency (was post TB 3.1 mailnews backend plans)
asutherland at asutherland.org
Mon Jul 19 19:12:42 UTC 2010
On 07/18/2010 06:45 AM, Robert Kaiser wrote:
> Do you mean that gloda is about to be changed from a pure index to
> something that actually holds important data and is necessary for the
> app to function properly?
(responding to list, as I think this question is likely interesting to
others and it's not clear it was intended for this to be directed only
It depends on your definition of "important data" ;). Specifically, I
would expect the first new account types to be things like twitter,
facebook, mailing list archives, etc. In these cases, we could probably
claim that the authoritative store is twitter, facebook, the site that
holds the archives, etc. At some point I would eventually expect for
someone to create a new account type/message source whose authoritative
store (or a replicated but equally authoritative store) resides within
I am planning to add explicit support for marking the importance of data
so that a single mechanism can be on the hook for ensuring reliable
storage so that account types/message sources do not have to roll that
In terms of being required for the app to function properly, I think it
depends what "app" you are talking about. None of my changes will touch
the mailnews core (pretend gloda does not live under mailnews/).
SeaMonkey's current MailNews consumer will be unaffected. Likewise, the
current Thunderbird 3-pane view will not be affected.
However, all new UI I will be working on will be building on top of the
gloda layer. Given our add-on driven development style, these will
happen as extensions, but for the purposes of discussion you could
consider them an "app". I think this is roughly what rkent is referring
to with what he has dubbed gigabird. I'm not on board with that name
(mainly because it too easily plays into quips about Thunderbird's
memory usage :) and have been referring to the effort as "thunderbird
prime" because it captures the general concept, sounds cool, and does
not require as much explanation as a completely different name would.
However, that's not a remotely official thing.
Obviously, any existing app that wants to leverage the new work is going
to need to operate in the same domain as the new work. The good news is
that my UI efforts are very focused on extensibility so if seamonkey
wanted to transition from operating at a mailnews level to a gloda level
it should be easier to implement a different set of UX decisions without
having to reimplement all the functionality.
More information about the tb-planning