ubuntu-phone team mailing list archive

[Jonathan Meek <shrouded.cloud@xxxxxxxxx>]

First and foremost, I would like to ask if there are still plans on the
table to make search universal in the new, converged Ubuntu? Will there be
that search item in the menu bar at all times like we've seen in early
designs? Is it there in the current implementation? I don't know because
I'm honestly just waiting for the time when there is a phone on sale so I
don't have to worry about channels and such... I'd just like a simple
experience that I can report back on and help improve. The rest of this
email will be comprised primarily of my views on why search should not only
be leveraged, it is critical. Ignore this if the plans still is in place
for universal search.


 Alright, let's begin: the toolkit is in place to provide developers
pre-made blocks with which to develop their applications. In this day and
age, I cannot think of very many applications that would be out of place
with a search function. No need to keep duplicating code when we could
implement it in the toolkit and developers can re-use these same search
boxes.


 As it stands, from my testing of the core apps on the desktop, most apps
have reimplemented search by borrowing code. (Please correct me if I am
wrong about this fact. Also, side note: why can't we agree on where
delete/close/clear will go? Why's it on the left for apps, but right for
search/text boxes?) So, let's just move that into the toolkit and make it
easy to reuse.


 But, that only brings things up to par. For this platform to be
successful, it can't just match others, we have to do better. Thus, I
propose that search be able to hook into other parts of the applications.
Indeed that it should eventually default to searching all non-user data
(with the option to search user data-- such as music metadata-- left up to
the developer, with a corresponding apparmor policy). That the left side of
the menu-bar could become a finder function. I'm in clock app and search
for “add city” and it parses the model names in the QML applications or HUD
functions, offers to take me out to the dash to some related scope...


 Or if I find myself looking for San Francisco, the same application
presents a visually appealing clock for that area as well as offering to
open a weather scope. I guess the point of this is to bring some of the
dash's beauty to apps. Unity has become very focused on search... So has
Android. Siri and Cortana, search.


 While the dash by and large supports this, we shouldn't limit it simply
because that's what's currently being done. We need smarter search. And
search cannot be smart without access to all the things that actually use
the data. We need search to bring the whole system together, not have to
drill down into everything to get to relevant data.


 So, let's sort of sum up what I propose:

1. Take search and put it in the toolkit with various ways of presenting
list of information

2. Beef it up and let search hook into all parts of the toolkit (developer
opt-in)

3. Transition to search being on by default for all non-user data with the
option for user data being searchable, but flagging these scenarios as
being potential security risks (user opt-in)

4. Enhance search results visual appeal and integrate HUD with it.

5. Integrate Scope Scope results (I guess, still iffy on this one, but it
is a first draft, why not go for the moon and let someone else shoot me
down?)

6. Remove the need for search even needing to be added to the application.
It simply will be. It will always be there for the user to access in the
menu-bar.


 That's the gist of it. I'm not too concerned with privacy as my
understanding is that things can't properly be smart without data. However,
I'm sure the privacy-minded will chime in and help shape the idea in that
light. I'm also sure that security won't be nearly as simple as I've laid
out, but I'm an idea man, not security ;)


 Essentially we want to take Finder, take the Dash and mash them together.
We don't have Siri, we don't have Now, we don't have Cortana, but this
proposal will be all but necessary to lay the framework for an open
personal assistant. Let's be cautious, but not stop moving forward.