un-NIH-ing accounts?

[Aaron J. Seigo]

On Monday, October 8, 2012 09:34:41 Mark wrote:
> On Mon, Oct 8, 2012 at 9:31 AM, Aaron J. Seigo <aseigo at kde.org> wrote:
> > On Monday, October 8, 2012 02:50:57 Alex Fiestas wrote:
> >> This project came out at the last Active sprint, at least 2 people there
> >> mentioned that the MeeGo system wasn't something we wanted to keep.
> > 
> > Could someone share what the reasoning was?
> 
> I could post the sections of the blog post, but that's not very nice i guess
> :) http://debarshiray.wordpress.com/2012/10/06/goa-why-it-is-the-way-it-is/
> read "The Split" section. It sums up a few reasons.

i'm more interested in the reasons raised at the Active sprint.

i've read debarshi's blog entry and the reasons given were:

* it has Qt in it (not a problem for us)
* it used libdbus-glib-1 and how it conflicts with GDBus. he wasn't even sure 
if it still used libdbus-glib-1, and I'm not sure any the parts of it we would 
use would use libdbus-glib-1 either. As Ubuntu is using it just fine with Qt 
code, I'm guessing it's not a real problem at this point.
* it does pass-through auth which is more difficult for applications. that seems 
to be mitigated, though, given this response:

	https://mail.gnome.org/archives/desktop-devel-list/2011-
April/msg00139.html

so he seems to have 2 indefensible and/or irrelevant points, and one that 
could use some more looking into.

tbh, my main question / concern is whether or not the accounts library 
requires the SSO daemon to be used in all cases. one use case requirement we 
have is a way to configure akonadi. assuming akonadi does not start using the 
SSO daemon, then the accounts management needs to be able to be made to 
configure akonadi. i doubt it will be the only example of this.

i've spoken with stskeeps from mer and he noted that it is currently used in 
nemo, and was looked at for inclusion in mer core but they passed in order to 
focus on stabalizing more core pieces while the SSO stabalized in nemo. so it 
could even, with more support, end up in mer core too.

what i'm really hoping is that there is someone with interest in this topic 
area that can offer a good synposes of the benefits and negatives of using the 
accounts-sso framework compared to writing it ourselves.

right now i have no great opinions one way or the other, just want to ensure 
we look at the whole field and make a decision from knowledge :)

-- 
Aaron J. Seigo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20121008/068cd3d1/attachment.sig>