Re: A mockup for an alternate conduit GUI.

["John Stowers" <john stowers gmail com>]

On Wed, Jul 23, 2008 at 1:04 AM, Martin Owens <doctormo gmail com> wrote:
>> There are still open UI questions - where does one put configuration
>> for SyncML phones?, other phones/devices?. We live in an imperfect
>> world, where not all required information about a device can be
>> queried and an automatic configuration guessed. The risk is that in
>> optimizing for the zero configuration case one makes the other cases
>> exponentially more painful. It does not have to be this way, only
>> through discussion, experimentation and mockups will and answer reveal
>> itself.
>
> Some of that kind of hardware information might find a nice home in
> HAL via dbus with more hardware detection type scripts and programs.
>

Definately, synce is moving that way, and I suspect the same can be
acomplished with Blackberry type devices.

With regard to Nokia type phones (i.e. sync via AT, or
obex-over-bluetooth/air/serial) I agree with Bastian. its not possible
to know with a useful level of certainty whether a phone supports
feature X/Y/Z, and to what extent their particular implementation of
$PROTOCOL is broken. There is also the difficulty of what *level* of
detail do they support - complete VCARD, or just name+number, etc.

The safest thing to do (in the vanilla phone case) appears to be
1) establish a connection to the phone via serial (over bluetooth),
using AT commands get the phone details
2) Either/or look up the phone capabilities/quirks in a DB based on
model number, or query phone capability directly (do you have a tasks
function, etc).
3) Try to read/write a vcard to the phone to see what data you get
back (to assess level of detail).

I suspect 1+2+3 is too intrusive to accomplish in a hal callout type
implementation.

John