Hi All
Following last threads on this ML, I see that the discussion became very technical and it got to the level of discussing the possible ways of implementing things. Thinking about having the MI in the upper menu there, warms my soul and helps me
dream about having lots of icons aggregated to one, keeping
cleaner and better overall looking desktop and still providing the required
functionality. But it currently still seems kind of tailored to "developers" needs, which is perfectly normal for this stage of a product.
There is still so much talk and discussions about what the MI should be. Should it be just a notification dialog? Should it be a launcher for specific app in MI, for example Evolution? Should I have applications items there only when the applications run, when they have a message, or always?
I just thought that as an end user I don't really care about applications. All I care about are my email messages, IM messages, twitter/
identi.ca messages, maybe RSS news items, calls I missed, etc. So why can't I see *that* in MI? In MI, I want to see my missed events, *not applications*.
So I have the following idea: there are libraries exist that can connect to almost any IM network and "listen" to messages that are directed to me (even by "pretending" I'm online), there are libraries for fetching email information, twitter information, jabber information, RSS information, I guess it is available also for SIP, and could be for many gazillions other message sources/channels.
MI should be able to use these libraries *without applications* to connect to messages source I would configure, and fetch the information. And then it would show this information, *regardless of the applications I may or may not use*, or run for that matter.
This would achieve the following goals:
1. MI will present me the simple *content*-based view of what I've missed; not applications. For example:
Email Messages
Inbox
Gmail
Other Mailboxes
Twitter Messages
alex@twitter
Identica Messages
alex@xxxxxxxxx Jabber Messages
alex@xxxxxxxxxxxxxxxxx
ICQ Messages
987654321 - alourie
News (RSS)
NY Times
Planet Ubuntu
Phone
Missed Calls
Missed Messages
2. MI would be able to *present* me any messages I have missed without me having application running (I only have IDE opened during coding session, but I still would like to get email notifications).
3. I would be able to click on the item in the list in MI, and then an application will be launched to handle that specific item.
4. It will keep MI as it is - an indicator/notificator, which is intuitively clean and clear concept.
Possible QA session with myself I could think of:
Q1. How would the generic service be defined?
A1. A system-configuration UI utility (similar to Preferred Applications) would offer configuring various "message channels/sources", and optionally applications that would handle them. The underlying infrastructure could be modal (plugin-based), where each component/plugin provides means (API) for connecting to a new kind of service and presenting the results in MI.
Q2. What will happen if "message channel/source" is defined, but no application for handling messages is installed?
A2. Upon "clicking" the message in MI, a notification will appear offering to install an appropriate application and even possibly automatically configuring it using the already provided information (from Q1).
Q3. How would MI know which application to use for specific item? (an expansion of Q1)
A3. It could be set by:
a) A system wide default application for specific message type/source, such as Evolution for all kinds of email (as in Q1)
b) MI-specific item handler, for example Firefox for Gmail items, Evolution for Inbox, Thunderbird for Other Mailboxes
c) MI-specific group of items handler, for example Gwibber for Identi.ca and Twitter messages
I'm not sure how different this is from the current design - probably very different, but it seemed to me that MI is being too focused on handling calls within specific application sets, while I feel it will be cooler focusing on messages I miss.
I'll be happy for feedback on viability of this and please don't kill this right away :-)
Alex.
P.S. Huge thanks to
mac_v for helping and raising tough questions.