[mythtv] New MythTV interface announced!!

[Xavier Hervy]

Richard Morton wrote:
> HI Xavier,
>
> Thanks for your reply. I just want to start this by asking have you 
> looked at the design document?
Yes and no, I took a glance at it (do you really expect me to read 25 
pages ;-) )
>
> The document at no point says that the solution would or would-not use 
> MythWeb as its basis. This solution doesn't compete with MythWeb 
> (except for the maybe "remote keyboard"); they aimed at different 
> purposes ad audiences.
>
> Have you used a Sonos?
>
> Imagine a Sonos remote controller for MythTV... now your getting close 
> to what we are suggesting... interface on teh remote device; playback 
> through the stereo & tv.
I do understand what you are trying to achieve
>
>
> > You are not answering to the essential question, why reinventing the 
> wheel ?
>
>
> The Telnet interface is simplistic, doesn't allow you to complete some 
> simple tasks; it could be extended to do what we want if the 
> mythfrontend developers were interested in helping us with just a few 
> extensions to the telnet interface.
At the moment there is 4 ways to communicate with mythtv (AFAIK)
- mythweb: backend & frontends through the web remote (via telnet)
- upnp: backend only (i.e no action)
- telnet: mainly frontend with key , jumpoint and access to the backend 
(recordings, listing) in a very slow manner (at least on my phone).
-mythlcdserver: does work roughly like telnet but it does allow to have 
some feedbacks from the frontend such as when the user is changing the 
current menu, what song is being played in an even driven manner

Unfortunally there is few drawbacks:

- the telnet interface use the keybindings instead of using the actions 
(mute, menu, info, back ...) which make all the client relying on the 
keybinding, if the keybinding change, all the telnet client will likely 
stop working properly. I did try to work on this issue before to allow 
the telnet interface to use the action instead, but i had few problems 
on the way and nobody to answer my interrogations.

- the telnet interface does not allow more than one connection at the 
same time. I send a patch to fix this few months ago, but I do not think 
it did make it way to trunk so far.

- the telnet interface does not allow event driven mechanism. it would 
be nice to be able to register event listener via telnet, ex:
#registerEventListener music.elaspedTime
and have a feedback such as
#music.elaspedTime 190

- finally the telnet interface does not allow plugin to register an 
interface (events, action ... available for a particular plugin)

Such mechanisms could allow to remove mythlcdserver (hack), and replace 
it by a telnet client to do the same job, this would resolve the 
complexity too.

>
>
> For a start; the interface is designed that there is no need for on 
> screen menus, there is no need for on screen text... The architecture 
> of the solution would be up to the developers of the solution - 
> afterall that is what they are good at...
>
> With a few extensions video playback could be completed via a very 
> minimalist (almost no) theme and OSD.
There is no need to change anything in the video playback, you can 
already play any videos (recording + mythvideos)  on mythfrontend via telnet
>
> Other media playback (i.e. music playlists, pictures etc) would still 
> need to be completed via a thirdparty application as mythmusic does 
> not have a telnet interface unless you go the keypress route of 
> navigating into mythmusic then requesting two presses down for the 3rd 
> album to be selected... this is not going to be very reliable...
Why not adding it then ? seems to be a quicker way to me.
>
> I will update the document
>
>  
>
>
>     Sending command through telnet is much quicker than sending it
>     over http. The key is to use telnet for what it is best for and
>     using http to retrieve data.
>
>
> I agree it is all about using the best technology to get the best 
> interface for users, which will drive adoption. The WAF would go 
> through the roof with this interface ;-)
>  
>
>
>     I personally  use a simple remote application for Symbian Phone
>     which does use telnet and I can not say it is slower than using a
>     standard remote (not mentionning that the range is much better
>     thanks to the wifi).
>
>
> and for some that is fine, but I have played with the available 
> remotes on touchscreen devices and I dont think they are where they 
> need to be. Touchscreens can be so much more than a representation of 
> a physical remote control - why not just use a device where you can 
> feel the buttons so you can /see/ the results as you can look at the TV...
>
> This solution allows you to look at the screen of the touchdevice... 
> so you can see where the controls are and read the information rather 
> than having to move your head back and forth between the two screens 
> (tablet & TV).
>
>
Should I understand that this proposal rely on the fact that the device 
have a web browser ?
If so, I still think you can already do most of this using a modified 
(templated) mythweb interface (using telnet in the background).
Xavier