[MojoMojo] Ideas about notification

[Marcus Ramberg]

On Wed, Nov 25, 2009 at 1:33 PM, hoelzro <rob at hoelzro.net> wrote:
>
> Hello, MojoMojo!
>
> I'm an ambitious young developer who's been using MojoMojo for quite some
> time now, and I figured that I should put my Perl and Catalyst skills to
> good use outside the workplace and contribute to a project as great as
> MojoMojo.

That's excellent news for us :-)

> Now that my introduction is out of the way, onto my ideas about
> notification.  I was looking at the issue pertaining to notification
> (http://github.com/marcusramberg/mojomojo/issues#issue/49), and I like the
> "More Difficult" option, where anyone can subscribe to a page's changes.  I
> have some ideas on how to expand upon that:
>
> - How about a recursive watch, so a user can monitor all page changes under
> a certain namespace?  This would allow admins to monitor the whole MM
> instance without much hassle.
How would this differ from an RSS feed of recent page edits?
> - I think a checkbox on page creation/edit to subscribe to that page's
> changes would be cool.
I like this idea, even though I worry about clutter on the editing
interface. Not sure how we make such things scale well.

> - A user setting to automatically subscribe to pages that that user creates
> or edits would also be nice.
This might be a better approach to the problem

> - Some sort of subscription management UI in the user's profile/preferences
> page would be nice.
Agree
> - As far as the actual implementation of the notification system goes, I
> propose a generic notification mechanism that various output modules could
> subscribe to.  A page (or something else in the future) would publish a
> changed event, and any installed output modules would broadcast the contents
> of the message to all interested parties.  That way, we could implement
> output modules for e-mail, XMPP, or anything else that comes to mind in the
> future.  When users subscribe, they can specify how they want to receive
> notifications.

Might not want to overengineer this. Let's try to start reasonably
simple with a DBIC class for notifications

> Granted, several of these ideas would take a lot of work, and some may be
> difficult to implement efficiently.  Please let me know what you think of
> them.

In general I think starting simple and then seeing from experience
what more we need is a better approach than implementing something
complicated out of the box.


With regards
Marcus Ramberg