Thread: wth are there so many frontends?

Originally Posted by CarComp

I'm not making a commitment here, but i'm playing around with fullscreen directx. I've always wanted a 3d accelerated front end. MediaEngine seems to work pretty good. Maybe its time I created something totally different than everything.

kinda like SD (to it's advantage AND detriment) is already written using directX? heh, nGhost is also written taking advantage of OpenGL. I'm sure others may also.

Were you just being sarcastic?

I think what would be helpful (and I've suggested this before) is a section in the wiki that charts a list of frontends and features that they have. That way when someone says, "Frontend A doesn't do XYZ, I'll just write my own frontend that does XYZ" they'll be able to see that there are a bunch of other frontends that also do XYZ AND frontend B has a plugin API so I can just add what I want to that.

If anyone wants to start that wiki section, I'll be sure to add stuff to it about nGhost and LinuxICE. I'm sure other frontend devs would also do the same.

Originally Posted by CarComp

I'm not making a commitment here, but i'm playing around with fullscreen directx. I've always wanted a 3d accelerated front end. MediaEngine seems to work pretty good. Maybe its time I created something totally different than everything.

You would need a frontend with "intelligent skins" = bitmaps with drawing code. Then you can replace the drawing code as I describe here:
http://www.mp3car.com/vbulletin/cf-p...-concepts.html


The idea which it is based on is:
If I were a MediaPlayer component (an isolated piece of code), knowing how to browse and find titles in a database, controling the audioplayer component, What do I want from the GUI and what do I send to it ? Answer: I define that I need a button "+" and a textbox where I can put text into and I get signals back "Button forward pressed" and "Button forward released". I am not interested how it is drawn.

I've already written a frontend once. It still stands as the only one with its type of functionality (when it comes to menus and skins... It doesn't follow the "one on and one off" dual huge picture design), unless you consider the Cobra series of applications where I got some of the methods. DirectX and customization don't really go hand in hand, but i'm going to be working that out. It all comes down to being able to create a good set of rules for skins as well as being able to implement new ones without much effort. An example would be treating components in a skin (like the Pause button). You would have to define its properties, and its behavior in the skin. This is just pseudo code...

Obj btPause
.IMG = "\bluemax\pause.jpg"
.TOP = 26.17
.LEFT = 30.21
.FUNC = PAUSE
.ACTIVEMETHOD = SPECULARIZE & HIGHLIGHT
END btPause

Then, somewhere down in the main skin's subprogram...

Sub OnShow()
(some other code)
Select Case CurrentScreen
...
Case MUSIC
...
(Would determine the method of showing just the pause button)
(Others would be similar so all buttons would have the same effect)
btPause FADEIN & SPECULARIZE & WIGGLE

End select
...
End Sub


Like I said this would have to be a somewhat complex skin language of sorts but it wouldn't be the first time I wrote a parsing engine. I think it could be done as well as standardized.

When it comes down to the original post, why there are so many frontends, I think that what happened is in 2004 there were pretty much NO frontends. There was Cobra3, then what was going to be carplayer never materialized. It seems like 4 or 5 people all started writing a frontend at the same time.. Me and phat_b's MediaEngine, Coyote's MediaCar, Magnetik's AutoPlay, Jy's Pycar, and MediaCruiser. Development was slower than dirt cause no one knew how to program yet very well for these kinds of apps. This cause a whole slew of me too people to write applications.. The V.O.I.C.E.S. app came up, then Frodoplayer, then many others in the last 2 years.

The fact with the chart that tracks app features is that you'd never be able to create a complete one without allowing developers to just add their own checkmarks. A single entitiy couldn't keep up with the change. Its been tried.

Thats pretty much why there are so many. I think there are enough similar ones now. There needs to be something that hasn't been done yet.

I realize the list of features could get out of hand. I would limit it to the basic features of:

Code:

Developer | Code Language | audio backend | video backend | skinable | Plugin Interface y/n | Open Source? | Price | dead/alive |

...maybe a few more.

I would also have allow a paragraph for each frontend to describe (in summery form) what makes it different from the rest of the pile. Also a paragraph for a brief history may be useful also.

This kind of stuff would be really interesting to me. Maybe it would help the potential frontend writer focus on an existing frontend rather than spawning off a new one.

There's so many just because different people want different things, I know a lot use RR but after installing it it still needed skins before it would work and I downloaded a skin and couldn't get it to work still so I gave up, the install is just to complicated so I use Frodoplayer since when I installed it it worked and then skinned it how I wanted within the capabilities of the program, though as far as I know there's still no frontend I would consider my idea one.

Ideally i'd actualy like something more similar to the iPhone/iPod touch's interface, while I dislike Apple products on principle that interface would be damn near perfect for in car use imo.

CarComp hit the nail on the head... as someone mentioned iPhone... the idea of the interface design and functionality needs to be completely rethought. Don't try to make a touchscreen representation of a car stereo or big button touchscreen computer. A different HMI design is needed now... one that will make use simply and fairly intuitive. Installation and configuration needs to be pretty simple and straight forward... yes... even at the cost of some customizability. It doesn't need to do everything for everyone... it just needs to contain a core set of functions and do them very well. Everything else can be add-on plugins for people to pick and choose from... and the installation/configuration of them needs to be simple as well...

Exactly! Thats what phat_b and I were attempting with MediaEngine. At the cost of super duper customization, it will install, run, and work right out of the box for most people.

How would you define an interface? Right now we have gesture recognition, and touchscreen buttons etc. My favorite was and will always be the IRman and a remote control. Still there are huge areas for improvement. Its VERY hard to change a mentality about how things should work, and its even harder if you've used those things for a long time.

One of the things I want to try in the future is a core setup like you said, but its just a base. Alone it will do nothing, and EVERYTHING is a plug in. If you want just GPS and movies, hey go right ahead. If you just want a player, it will do that too. etc. And NO, I know girder probably does this, but you have to have a PhD in girderology to figure it out.

Gesture recognition seems like Palm Pilot writing recognition... sounds cooler than it really is in practice. So it just turns out to be a PIA and you end up using OSK/buttons instead.

I'm not much of an IR fan... RF or wired is my preference, but yeah... a nice little hard interface is nice to have... too bad there aren't any remotes that are small, but contain what we essentially need. Changing people's mentality IS tough... but if you do it right... they'll come (like with the iPod and iTouch)...

Your idea doesn't sound bad... as along as some control is exerted over the plug-ins to ensure they work....

Well elaborate? I'm open to suggestions.