Open Computing Solution -- OCS

I posted in different threads the basis for this idea, however, I'd like to pull those thoughts together here in one thread.

From my short experience here at mp3car I have realized that this market is wide open and severely immature, scattered and under developed. Meaning, the roads are open and the lights are green.

This goes for both hardware and software.

As a few people here have noticed, the days of simple stereo head units and car navigation systems are numbered. This can be attributed to what many pioneers here have worked hard to achieve and the simple everyday tool we carry in our pockets, the cell phone.

Hardware:
Many parties including mp3car have in many ways started in the right direction! bravo!!

Software:
Being a hack programmer myself, I can understand the problem. Actually Netscape/Mozilla comes to mind too. This whole segment of car-puting is still itself wide open to attack that WILL come from the big boys. Much like how M$ was able to dominate Apple for almost 20 years.

[let me ramble a bit more]
Recently, Apple did the right thing. dump a stale system for a better visionary (I still wonder about BeOS). Netscape was different, Mozilla dumped a stable, dated system for 10 years of pain and hard work which still needs to work out more bugs and problems. (firefox is a cpu / memory hog on my system).

Now, what I'm trying to get at is that the community needs to consolidate and get some project management into place. I realize that programmers are artist and like to work their own code and even like better to start from a fresh code base hoping to prevent problems that occur in current solutions. However this approach is time consuming and allows other teams to gain market share.

Only being Open Source doesn't mean to be OPEN in the sense I'm thinking. You need to get people to develop and sell what you are making. Money can be made much the same way as Mozilla.

Here is a simple example - Why don't car manufacturers agree on 1 color for each wire they use? Why not decide on 1 type of harness? Why not decide on 1 set of screws for the dash and the same for other parts of the car.

People are just too busy creating their own private solutions that are great but adoption is low because everybody is trying to do their own thing.

So, what I'd like to propose is a true "Open" Solution that is not just "Open" in the sense of open source.

1.) build a foundation of members (who should include competitors)
2.) decide on 1 OS -- some variation of *nix is probably better, Moblin/Android ???
3.) set an outline of APIs and create a SDK
4.) invite all competitors to join
5.) be Apache, MIT or LGPL based
6.) get hardware companies to build

The core I'm tending to think should be very, very cell phone friendly. We have to face it that if you want to grab a large share of the market then you need to realize that you phone is going to be just as powerful as what you have in the dash. So why not start now?

So, in a sense Android is compelling to use.

Mobile computing is not just on the road ... it's going to mean from the john in the morning to your desk at work.

Will this mean a fully cloud integration? Not, for another 10 + years, possibly never. People don't completely trust Goggle so why should they trust the cloud. Another problem is that people have private data ... And finally, the cloud is still not mature enough to support 100% coverage wherever you are. So, you still need to have "docking" ability with your own storage solution (home pc).

Sorry, for rambling. I wrote this a bit late at night and the small text box size ... lolo, should print out and edit :tape:

I think Android would be a terrible platform to use in a CarPC environment. Take a look at the android-x86 project and give their distro a whirl on a notebook or something... it'd take entirely too much work to make it a functional CarPC OS.

On Android.... Android is great for super low power devices, such as a phone. There is simply no reason to use an OS which is built from the ground up centered around the fact that it is going to be used on low power devices, on a high power device like a carpc. You can get much more speed and functionality out of a full operating system slimmed down for the carpc.

Also you'd have to write your own front end/system of applications for in-car use since the Android apps are centered around being used on a phone. They are NOT usable in a carpc where ease of use is a primary concern over functionality. You can't lean over the carpc screen to go through 10 screens to get to something like you can on a phone.


Another problem you're going to run into, is getting anyone to agree to any sort of set of standards. OSDash is attempting to do this for web services integration, and we'll see how that goes... but as for anything commercial like hardware? Forget it. It's not good business for a hardware company to make their stuff compatible with other peoples stuff, cause then you'll just go buy the other persons stuff. Better to have everything proprietary, so you have to stick with one company for everything. Just look at modern car audio systems (like BMW/Jag systems)

To be honest, I have never looked at Android. The idea of using that OS was just to have to the cross-over ability between phone and car.

Are you up on Moblin?

I agree on that manufacturers like to go proprietary, but that sounds like Apple. M$ was able to get installed on almost everything.

I also agree on the agreement part! :)

However, I am not really thinking of just re-writting a front-end. The idea is pull together everybody to pool everybody's efforts thus the idea of using a non-GPL license or using the commercial version LGPL. So, the hard part is to get multiple groups together behind one effort, like OpenOffice and Mozilla have been able to do.

I can understand how Android is lacking, which is why I threw out Moblin too. (Backend technologies here.) Having a larger base will draw venture funding, resource grants, and other monetary assistance. People working out of their rooms is great but it's sad when their efforts don't get their due deserves!

There are quite a few smart and creative people here at mp3car and I'd love to see their efforts be enjoyed by a larger crowd!

I may have missed a memo or something, but I thought the whole point of the open source movement was to provide low or no cost alternatives to proprietary applications. I am often amazed and impressed by the time, effort and quality of those results. But as far as I can tell, nobody is making any real money doing it.

The economics of the carPC are simple and deadly. It is a tiny market, the knowledge level required to participate is far outside the consumer norm, and the cost of entry is prohibitive. 99% of the people that drive cars have no idea what any of this stuff is and don't care. They just what a device that does everything and works every single time they turn the car on. Anything else is unacceptable. Configuration should consist of choosing between on and off.

As a hobby, or in some cases a holy quest, I thnk most of these open source projects are pretty cool. They very often lead to some truly innovative solutions to common problems. They are the true trailblazers. But its the guy that comes along afterwards and builds the road and puts up the toll booth that makes the money.

CarPC hardware and software WILL get standardized, but it will be Intel and Microsoft that do it. There isn't anybody else with the necessary industry clout, financial strength and inclination to do it. And when they do it, it will be in an all-in-one box in the dash, not in a cell phone.

But of course, I could be completely wrong.

VegasGuy

I agree with you especially on the end consumer opinion !

However, there are open source projects whose foundation gets by with funding, ie, Mozilla, OpenOffice, SugarCRM, etc ...
Other projects make their money on custom development for customers.

Problem is having the right angle and niche!
What you all here found is that niche! Looking through old threads, I'm just amazed by what all you folks have done!

I don't think M$ will compete yet. They are more of the buy you out or fund you so well that you are in bed with them.

In regards to the last comment about not in a cell phone, I think the sheeva plug guys are on the right track. Or at least heading in a direction which the future is going to be on too.

I personally would rather play the devils advocate and instigate you guys into building the foundation that will get funded or bought out rather than let M$ just walk in and buy centrafuse.

I don't think the future is that far off since now most car makers are supporting the iphone/ipods and bluetooth. For better or worse the iphone generation have changed phones are car audio ... to some extent home audio too!

Quote:

---

Originally Posted by VegasGuy

The economics of the carPC are simple and deadly. It is a tiny market, the knowledge level required to participate is far outside the consumer norm, and the cost of entry is prohibitive. 99% of the people that drive cars have no idea what any of this stuff is and don't care. They just what a device that does everything and works every single time they turn the car on. Anything else is unacceptable. Configuration should consist of choosing between on and off.

---

Boy, did you ever nail it! That's as succinct a summary of the issues as I've ever heard.

@SapporoGuy - take a look at the Linux ICE project and nGhost. Both are open, Linux based, and fit many of the needs you articulate except for the 'wildly' popular one. If you wanted to help, what you could do is demystify the installation of Linux ICE by figuring out how to package it up so all you have to do is boot it from a USB stick or something like that.

Malcom mentioned the OSDash project, which is currently moving from the idea stage to the launch stage. We'll see how that goes and we ought to be launching its own forum here in the next day or so. It is similar but not identical to what you're talking about.

the biggest issue that i have, as a hardware guy, is software-- give me any physical device, and i will figure out a way to modify it, install it and wire it up, but, give me software, i have no clue what i am doing, and because of that, linux, or other inegrated os's are a horrifying experience, because i have no clue how to even just install drivers(i know i need to log in as root, or super user in terminal, but that is about it)...

i think that to get the masses into carpc's, the installation would need to be extremely simplified from what it is. ie: to load the os/interface/drivers, use one disc, and it will automatically start to install when the system boots up. want to add music? just throw the disc/flash drive into the nearest slot, and a box will pop up asking if you want to add the music, or just play it...

i know that there is way more to this than what i make of it to program the sofware to do it, but the only way, IMO to get the masses interested in car computing is to take out the confusion of the software, and make the installation as mindless as possible to do.

i've been thinking about this alot myself.

I've begun to realize that pretty much every FE is lacking something. I think RR is the most customizable, and is probably going to be the way I go, but it still has quite a few "locked in" choices, that I will need to undo to make it what I want. And i have to admit.. all of that undoing is really really daunting.

The front ends are in my opinion.. trying to be too much. Really when it comes down to it, a good FE is nothing more than a skinnable launcher. Once preferences on which media player, or which codecs or what have you start piling on top of that you start running into the issue of my way or the loooong hiway to your way.

I'd like to see a front end that once installed comes up to an options screen on first boot. this options screen is nothing more than an "app store" that allows you to add plugins for whatever it is your after. as the FE creator you of course could write these plugins, but the real point is choice + survivbility. my first major hurdle with riderunner was trying to figure out.. wtf is phoco.. well now phoco is deader than dead, and even looking for info on it leads you in huge circles. it wasn't until very recently i found out, development stopped, and the guy doesn't even want the code handed out.

if phoco was just a plugin, in an "app store" it would still be there, or it could be removed and i never would have even knew it existed.

So in short, i think my biggest realization is that there are alot of projects that started properly, but now we've got ALOT of cruft, and not alot of "publication" for these apps.

this would be a HUGE opportunity for mp3car to open up an app store, and offer up an sdk to access that store.. expand that to begin offering paid and free apps and you're looking at "the next big thing" and the big boys will start to stand up and take notice, and hopefully start buying up these projects and getting these talented folks some recognition and funding to drive the projects higher.

/noob rant