YOUR dream front end - Page 3

iamgnat · 02-12-2008, 12:08 PM

Quote: Originally Posted by Jirka Jirout

Bindings? KVO?

I believe KVO is there (isn't that a fundamental part of Cocoa?), I don't know about bindings though.

-dave

Creative0Mind · 02-12-2008, 12:10 PM

Subscribed.

Cybix · 02-12-2008, 04:54 PM

you should all collaborate on a project, then we'd have the ultimate FE. instead of a handful of un-finished and un-polished ones

NeonDev · 02-12-2008, 05:22 PM

Quote: Originally Posted by Cybix

you should all collaborate on a project, then we'd have the ultimate FE. instead of a handful of un-finished and un-polished ones

problem mostly seems to be that not every dev here will agree on how to tackle anything that has to do with displaying anything... like Jirka said, the only real possibility for collaboration is in code that can be share across all mac FEs

iamgnat · 02-12-2008, 05:54 PM

Quote: Originally Posted by NeonDev

problem mostly seems to be that not every dev here will agree on how to tackle anything that has to do with displaying anything... like Jirka said, the only real possibility for collaboration is in code that can be share across all mac FEs

Not only that, but organizing a project of the magnitude that a FE should be is no small feat. Especially if there is no pay check involved.

Furthermore, coordinating across time zones and life schedules is not conducive to easy rollout plans.

It's not that it can't be done, it's just a lot of work and has a tendency to strip out most/all of the fun which is what we are in it for. If you look at the most successful OSS projects going on today, they almost all of some kind of corporate sponsorship and staff that don't have to balance another job that puts food on the table with a project they are doing in their free time.

What really needs to happen is for the Mac car community to grow both in users and developers. As that happens, there will be a larger pool of people to focus on the projects that they like the most which will enable them to come into their own like the Windows FEs (and if you look at the Windows folks, they don't just have one FE and they have as many opinions as we do).

The good news is that we are growing (seems like we've at least doubled if not quadrupled the number of active (posts at least once a week) members in the last few years), but we will probably always be a niche group with all the issues associated to it.

The one thing we need to absolutely avoid is having the various projects get into "mine is better than yours" ****ing matches (which hasn't happened to date) since that helps no one. Good natured collaboration (even if it is just ideas) is what all of us need to deliver the best products we can.

-dave

Bugbyte · 02-12-2008, 08:35 PM

I don't think Neon Boombox looks 'primitive' or unpolished. That's kind of harsh. I think it looks pretty nice. Usable in the car? I'm not so certain about try to drive and drag that dial around at the same time, but I already put my order in for key bindings so you can drive it with a rotary controller.

I think all this arguing about architecture or user interface skirts the main issue - capabilities. Any FE should be first and foremost about [[[[[[helping one drive the car. In the same way that flight management systems are designed to help aircraft crews operate the aircraft easily, an FE should be about making it easier to drive.

Sure, music is number one. The issue is that with a car unit, you have so much music to select from and so many sources you can choose from (mp3, cd, xm, radio) that you need help managing it while driving safely.

I want information about where I am, where I'm going, how far I've either gone, how close the nearest gas station with cheap prices is, what's my average mpg, historical info on car operations, temps, trip speed, estimated time of arrivals, navigation, and so forth. It should be simple to set up, easy to use, and interactive if necessary.

And for heaven's sake, make it skinnable. NOBODY can agree on what the interface should look/sound like.

A dream FE should integrate calculations you have to make in your head while you're driving, inform you of things you didn't know about (temp/wx, traffic conditions, traffic cams, gas prices) do something you couldn't normally do (like transmit your location or read your your emails), and integrate stuff you normally have to fumble about the cabin for -like answering the phone or dialing.

It should help you out, not make your life more difficult. I haven't seen a single FE Windows, Mac, or otherwise that does this. They all fail to consider the driver's position and instead require more mental agility to deal with, not less.

Unlike FMS systems, you can't go 'heads down' in the cockpit in a car. You'll run off the road. A dream FE would anticipate, help out, and integrate, making your trip easier, not harder.

smilr · 02-26-2008, 10:54 PM

I've been developing a Front end for my own use. (what's that make, 4 or 5 mac FE's in progress now?) So far I've shied away from animations and flashy colors - focusing on readability in bad lighting conditions (for non transflective screen use) and the music engine.

For me, it is essential to keep some controls always available (volume, mute, pause, the ability to switch to another function (nav / music / gadgets / web) with a single button press) regardless of what function the system is in. Too many of the FEs I've seen have hidden away these controls when you're doing something else, just to reclaim a little screen real-estate.

Audio is the key function for this system, so I've been heading towards using CoreAudio as the music playback system (as opposed to Quicktime or controlling iTunes in the background). This will allow me (I hope) to do front/back fading, subwoofer level control, graphic EQ and time displacement - just like conventional car stereos.

It may hold all my music, it may be easy to use, it could even dance a jig - but if it doesn't SOUND better than a regular stereo, what's the point?

NeonDev · 02-27-2008, 11:37 AM

i feel that 2 button presses is acceptable for switching functions (from music to navigation etc). working on a small 800x480 screen that needs controls big enough and spaced enough to read/poke from a small distance doesn't leave enough room half a dozen buttons for function switching IMO.

I agree that mute and other such buttons should be easily accessible but i feel that it should be user definable which/how they are. ie a tray somewhere for "hot controls" or even leave quick access to those as voice commands (i know not everyone has a mic)

as far as audio processing I felt that the audio hardware would be able to handle EQ and other functions. but a core audio solution would be tops, however difficult to implement.

are you going to release your FE for others to use?

iamgnat · 02-27-2008, 12:50 PM

Quote: Originally Posted by NeonDev

i feel that 2 button presses is acceptable for switching functions (from music to navigation etc). working on a small 800x480 screen that needs controls big enough and spaced enough to read/poke from a small distance doesn't leave enough room half a dozen buttons for function switching IMO.

I believe the reality is somewhere in between ready access to controls (e.g. always visible) and on demand (e.g. auto-hiding toolbar/dock), but I agree that space is at a premium on screens that the "average" car PC will use (e.g. 7" to 8" running some 800xYYY resolution).

I think controls that are specific to a single function (e.g. music player) should only be visible on that functions screen(s). If it shows up elsewhere, it is that much less space that can be used (e.g. the "mythical" navigation maps would be smaller and therefore less useful). On the other hand, functions that are common to most/all features (e.g. volume control/mute) would require one-touch access IMHO.

I also think that many/most features don't need the whole screen as long as the normal/typical view is displaying the "critical" information (e.g. it would be cool to show all the tag info from a MP3, but it would clutter up a small screen to the point where the driver can't get the "critical" information at a glance) which leaves space for "one-touch" functionality.

On the flip side, you will have features like Video and Navigation that would benefit from using the full screen, but care should be taken to make it easy to get out and switch to another feature if needed.

My rule of thumb is that if it takes more than 1 command in a non-full screen feature to get to another common (allowing the user to define what is common), 2 commands from a full screen function, or an additional (e.g. 2 or 3 depending on screen usage) command for non-common functions, then the UI takes too much attention to navigate in the car environment.

-dave

ibmeruu · 03-02-2008, 04:30 PM

Neon Boom Box has allot of potential.

I would love a Speedometer in it because I am going to put a drum machine in my steering wheel.

That will block my analog speedo so I would like to just see one on my screen.

Or maybe just a speedo widget ?

smilr · 03-02-2008, 08:58 PM

Quote: Originally Posted by NeonDev

i feel that 2 button presses is acceptable for switching functions (from music to navigation etc). working on a small 800x480 screen that needs controls big enough and spaced enough to read/poke from a small distance doesn't leave enough room half a dozen buttons for function switching IMO.

I agree that mute and other such buttons should be easily accessible but i feel that it should be user definable which/how they are. ie a tray somewhere for "hot controls" or even leave quick access to those as voice commands (i know not everyone has a mic)

as far as audio processing I felt that the audio hardware would be able to handle EQ and other functions. but a core audio solution would be tops, however difficult to implement.

are you going to release your FE for others to use?

I've had to rethink my UI layout now that I have tried it on my actual touchscreen. Some of the buttons were too small, and indeed - these were the function switching buttons.

I do plan on releasing my FE. Although I've written out pretty much ALL of the code from AMP2 that I began with, it still technically has roots in that program, which was open-sourced, thus mine must be even if I hadn't wished for that.

The UI I will build is going to be far less flashy than those I've seen elsewhere for Mac / PC. I'm also not putting any effort at all into video playback. In the end, I suspect the only thing anyone else will want to use from it is the Coreaudio playback engine, assuming I ever finish the damn thing.

NeonDev · 03-03-2008, 05:54 PM

Quote: Originally Posted by ibmeruu

Neon Boom Box has allot of potential.

I would love a Speedometer in it because I am going to put a drum machine in my steering wheel.

That will block my analog speedo so I would like to just see one on my screen.

Or maybe just a speedo widget ?

before that happens I need to actually have a mini in my car. I don't actually have a carputer yet thats why development has been kinda slow on NBB.

but when I get one and get some way of hooking my mini up to the obd2 interface on my car im sure I can rig up something. if you are a developer at all I can help you write your own module for NBB there is an SDK but I haven't released anything but a verry rough beta that has changed a whole lot since.

NeonDev · 03-03-2008, 06:01 PM

Quote: Originally Posted by smilr

I've had to rethink my UI layout now that I have tried it on my actual touchscreen. Some of the buttons were too small, and indeed - these were the function switching buttons.

I do plan on releasing my FE. Although I've written out pretty much ALL of the code from AMP2 that I began with, it still technically has roots in that program, which was open-sourced, thus mine must be even if I hadn't wished for that.

The UI I will build is going to be far less flashy than those I've seen elsewhere for Mac / PC. I'm also not putting any effort at all into video playback. In the end, I suspect the only thing anyone else will want to use from it is the Coreaudio playback engine, assuming I ever finish the damn thing.

I don't think that you have to open source anything just because it is based off AMP. (what license is AMP under). Also AMP itself has a large amount of borrowed code from various places on the internet (don't know about AMP 2) so depending on what you kept from AMP you may need to open only part of your code...

anyhow how far have you gotten on your playback engine? I have barely used any core audio myself largely because it doesn't seem nearly as programmer friendly as quicktime...

lack of flash is fine with most people btw. NBB gets criticized for being too flashy I think... I have tried to cut back regardless because of performance constraints. there are alot of optimizations I need to make for release 6 but first i need to get 5 out! (any day now)

Bugbyte · 03-03-2008, 07:22 PM

So...I just took a five hour trip with the Mini and several front ends in the car. The upshot is that none of them *quite* work the way I'd like them to.

I'm mostly in need of music and navigation. The music come from both the Mini itself and XM. I buy music on iTunes, so I have music with DRM in it. I also have a dead touchscreen on an 8.4" screen that I haven't been able to find a replacement for. I use either the keyboard/mouse or a Griffin Powermate for control.

-Car Front End - Works fine, but is playlist based. Doesn't work with spaces, though. To access navigation, you have to quit the program.

-Neon Boombox - Lots of promise, but doesn't do playlists. You can use Spaces on it to reveal the nav program running in a different space, making it compatible with nav.

-FrontRow - Works best, includes video, photos, etc. When you get out to fuel the car and restart, picks up where it left off, like a radio. Can be controlled with Powermate. Can't be used with spaces, therefore no nav capability. Also, don't like the 'now playing' screen. Very murky in daylight, text too small.

None of the frontends I tested can play or control XM.

It'm very glad to see development in Mac front ends and grateful to have betas to test. It's difficult for someone like myself who knows more than the average user but less than the application programmers and hackers to cobble together a front end that works with just a few music inputs.

iamgnat · 03-03-2008, 08:13 PM

Quote: Originally Posted by Bugbyte

Car Front End - Works fine, but is playlist based. Doesn't work with spaces, though. To access navigation, you have to quit the program.

Technically you don't have to quit, you can hide it instead (standard cmd+H keys or the Hide button from the Menu screen).

As far as "doesn't work" with Spaces, I assume that you mean that your hot key commands to switch spaces doesn't work? I'm curious what NBB does (or doesn't do) that allows it to work. I have a couple ideas (i'm taking a rather heavy handed approach key bindings or maybe they aren't capturing the display (e.g. not a true full screen app)) that i'll look into.

EDIT: NBB must not be capturing the display. As soon as I stop capturing the display, Spaces works fine. The downside though is that when switching back to CFE's space it wasn't always getting focus (fine with a TS, but would suck without a mouse/pointing device of some sort). It would be nice if Apple gave us hooks into Spaces, but as far as I know they still haven't and no one has hacked it yet.

-dave