Thread: The RR Replacements.

Right. The majority of people don't. Right now.

Just like we didn't used to have broadband or cell phones or even wireless phones.

What do we know about the likelihood of cheap, fast wireless? Here's how I see it:

1. Social networking demands always on connectivity.
2. Almost anyone with a cellphone has some form of net access.
3. Boost mobile phones have cheap, slow access.
4. iPhones, Android and Windows Mobile phones provide expensive net access.
5. The cost of that access has been falling.

Which begs the question - who is building the car PC of the future if everyone is building the car PC of today? If the future car PC is simply an evolution of today's PC, great. We don't have to worry about it.

But that's not how I see it. We can either invent the future or follow the inventors. Our choice.

2. Almost anyone with a cellphone has some form of net access.
3. Boost mobile phones have cheap, slow access.

2) I don't. I have a cellphone, but it can't do the internets, and I never text.
3) 50 bucks a month...
Over the time it takes me to pay for my current car, 50 bucks a month would come to more than 10% of the cost of my entire car. Or compared to my payments including interest, it will clock in at >20% of the cost of the entire car. [significantly higher than that the more time I own the car as its value depreciates, and the amount spent on internets can only go up]
That doesn't fit in my definition of "affordable", let alone "cheap".

Overall, I remain unconvinced of the value of internets in cars.


I guess another way of looking at it:

who is building the car PC of the future if everyone is building the car PC of today?

What is going to be the killer feature, or group of killer features, that will be so great that a car PC would make it worth paying a subscription for an in-car internet connection? Garmin have GPS units that understand traffic conditions without a subscription [according to a brief google search]. Radio is free.


Don't get me wrong, I'm not a total luddite; I even write some carputer software. I'm just waiting for someone to me show something that makes a subscription worth it, and I currently cannot perceive something that would make it worth it to me.

Social networking isn't it - especially for as long as "social networking" continues to be the ridiculous masturbatory high-on-my-own-ajax that is twitter. Keep in mind that social networking is a constant stream of information that you must read from a screen and devote attention to. This seems to have more potential for bad for drivers than for good - if someone cannot talk on the phone safely, then how can something demanding a richer interaction be safe, ever?

Gary (-;

Good points, all. And let's say at the outset that we're having an academic discussion here. I'm not trying to offend anyone by saying I think they're coding the 'wrong thing' or that they lack imagination.

To start with, I'm really not trying to make an argument that there is some killer app out there that makes turning your car into a network node so compelling that you'll part with $40 a month to do it. I'm advancing the notion that a car that is connected has greater potential value to the driver than an unconnected one and we aren't considering that.

Driver: Convergence

I base the proposition on what I perceive as the convergence of improvements in cost, coverage and speed for connectivity with cost, size and performance of smart devices. My thesis is that this convergence is and will continue to change mobile (walking, driving, riding) computing from a device centric (i.e. desktop, laptop, car PC) experience with discrete connectivity (at work, at home but not in the car or the sandwich shop) to a location and circumstance based experience that relies on connectivity to support that experience. The fact of the matter is that PC's are now so mobile they travel in your pocket and their cost is approaching the point where most everyone will be able to afford one.

Rather than the car sporting a device that travels around with it, the user now has a computer that travels with him/her to the car but also anywhere else they go. The present car PC experience follows this location/circumstance experience but it can only go so far. It removes the limitation of music confined to a single CD or iPod full of music and moves it up to a hard drive full of music. It takes personal navigation software and extends that by integrating it into your own customized and personal display front end.

That front end represents the doorway to a set of capabilities. Music, video, interface with the car's OBDII and so on. Yet it is generally a command and control experience. You command, it controls.

Possibility: Car as a network node

I'm simply applying standard network theory to the mobile environment and sort of doing it aloud in this discussion. Consider the value of a single telephone, which is useless if you are the only person that owns one. However, each person that connects a telephone to the network increases the value of the network exponentially. While it is clear that there is tremendous value to networking stationary PC's together on the net for the purposes of communication, data exchange, accessing apps and more, I don't see why that fails to apply to the car experience. Instead, I think the challenge is to address the captive attention requirement that stationary computing and smart devices impose on the user.

Here is where the current paradigm, built around the command and control of an isolated computing device, interferes with imagining possibility. What value could network access possibly add to the in car computing experience?

Thought experiment: An approach to imagining the car as a network node

I work in the technology industry with lots of engineers. They are really good at explaining why something cannot be accomplished. I've found it useful to ask the question the other way around -"What would it take to accomplish this?" Many times a solution is found. In that same spirit, I offer a question for those still reading this post:

"If you had infinite and free network connectivity in your car, what would you do with it?"

Some things right off the bat that I'd do...

1. I'd try to use it to make my driving life easier. I'd use it as a predictive tool that, if it knew the route I'd picked off of Google maps and knew my location, would monitor for traffic jams ahead and try to calculate ways around them.
   
2. I'd monitor trapster continuously and correlate that to my current position and notify me when I was close. Which is what trapster does except that I don't want to run a trapster app to do it.
   
3. I'd have it translate emails or IM's to me into audio files that it could read to me. I'd have it upload audio files of my response to a server that translated them into text and responded as if I'd typed them.
   
4. I'd use it to log OBDII data periodically and process it to see if there was any predictive maintenance I should do on my car.
   
5. I'd use it to record fuel tank levels and mileage automatically so it would guess when I filled up with gas and keep an automatic record of it. I might enter the prices at the point of sale or I might find that information on the web and automatically record it.
   
6. I'd have a server on the net monitor my location and decide if something needs to be done because of it. Send an email to my wife that I'm running late because I'm traveling slower than normal on a stretch of road I often frequent in the morning and afternoon? Tune the radio to traffic reports when the car starts? Bring up traffic cameras that monitor the route I normally travel to/from work if it looks like heavy congestion?
   
7. I'd use that connectivity to let the net do the heavy lifting and my car PC do mostly the display and local controls because I'm tired of endlessly reconfiguring my car PC when I tried to add a function that either broke another one or switched to a skin that supported a new thing but not an old thing. I'd rather connect to a server and run a web app that does what I want than try to figure out which .dll's I need to register.
   
8. I'd limit platform dependence and rely on the network to sort it out. Having a front end dependent on an OS used to be a requirement. I'd look for ways to reduce that dependence to as few regrettably necessary OS or hardware dependencies while letting apps on the web do as much heavy lifting as possible. That would allow more people with more devices participate. And they would need to know less and less technical knowledge to make it happen.
   
9. I'd make it open and easy for others to add to or modify

Lots of stuff *could* be done. Some of it is impractical, but I don't care. I'm a hobbyist. What would you do?

Originally Posted by Bugbyte

Some things right off the bat that I'd do...

1. I'd try to use it to make my driving life easier. I'd use it as a predictive tool that, if it knew the route I'd picked off of Google maps and knew my location, would monitor for traffic jams ahead and try to calculate ways around them.
   
2. I'd monitor trapster continuously and correlate that to my current position and notify me when I was close. Which is what trapster does except that I don't want to run a trapster app to do it.
   
3. I'd have it translate emails or IM's to me into audio files that it could read to me. I'd have it upload audio files of my response to a server that translated them into text and responded as if I'd typed them.
   
4. I'd use it to log OBDII data periodically and process it to see if there was any predictive maintenance I should do on my car.
   
5. I'd use it to record fuel tank levels and mileage automatically so it would guess when I filled up with gas and keep an automatic record of it. I might enter the prices at the point of sale or I might find that information on the web and automatically record it.
   
6. I'd have a server on the net monitor my location and decide if something needs to be done because of it. Send an email to my wife that I'm running late because I'm traveling slower than normal on a stretch of road I often frequent in the morning and afternoon? Tune the radio to traffic reports when the car starts? Bring up traffic cameras that monitor the route I normally travel to/from work if it looks like heavy congestion?
   
7. I'd use that connectivity to let the net do the heavy lifting and my car PC do mostly the display and local controls because I'm tired of endlessly reconfiguring my car PC when I tried to add a function that either broke another one or switched to a skin that supported a new thing but not an old thing. I'd rather connect to a server and run a web app that does what I want than try to figure out which .dll's I need to register.
   
8. I'd limit platform dependence and rely on the network to sort it out. Having a front end dependent on an OS used to be a requirement. I'd look for ways to reduce that dependence to as few regrettably necessary OS or hardware dependencies while letting apps on the web do as much heavy lifting as possible. That would allow more people with more devices participate. And they would need to know less and less technical knowledge to make it happen.
   
9. I'd make it open and easy for others to add to or modify

Lots of stuff *could* be done. Some of it is impractical, but I don't care. I'm a hobbyist. What would you do?

First off-excellent post....and I look forward to seeing where this conversation leads.
That said...Lets try to sort out what really needs the internet:

1. Traffic data-yea it could use the internet for ~$50 per month...or could use an msn direct for ~$50 per year.
2. Not sure about your area...but by me there are so many trapster alerts per mile that it becomes useless. That said-yea internet required.
3. Text via speech synthesis/recognition doesn't require an internet connection. Email could...or something like a blackberry could be used.
4. Check out the OpenOBD project I started up-it expands on this idea...and does so without requiring internet access. AFAIK there is no source for this info on the internet.
5. All of this is local stuff except fuel prices. For those, see my MSN direct comment above.
6. Yea if someone wants to host a server and write all the custom algorithms to do it. Traffic cams would be useless compared to traffic congestion feeds right into the nav engine.
7. Entirely a RR issue....better designed front ends don't have any of those problems.
8. Theres a giant gap between hardware and ajax apps...platform specific code would still be needed. Also a non-issue for platform independent front-ends.

So out of that whole list - internet would be beneficial for emails and vehicle tracking. I don't mean to shoot down some of your ideas I just think that in order to sell the idea-innovative and unique uses need to be thought of.

There is also another issue thats been overlooked here-Data Licensing. Just because the internet provides some data doesn't mean its legal or free for a front end to use it. Many RR apps are illegal/violate a websites terms of service. That may work for a small community project but certainly would never fly as the "future of car computing". The sad fact is that traffic, stocks, navigation data, etc. require licensing or subscription services in order to be legally used.

Sure, but that's stuff that just came off the top of my head while posting. Surely someone can come up with better and more imaginative uses. I hope I'm not misinterpreting your post but it sounds like as long as you have properly architected software apps that there's really nothing that the net can add to your in car experience.

Is that a correct interpretation?

Originally Posted by Bugbyte

Sure, but that's stuff that just came off the top of my head while posting. Surely someone can come up with better and more imaginative uses. I hope I'm not misinterpreting your post but it sounds like as long as you have properly architected software apps that there's really nothing that the net can add to your in car experience.

Is that a correct interpretation?

Well it certainly could be interpreted that way...but that wasn't the intention. More of playing devils advocate-trying to ask the questions your average user would ask when you tell them "this is better". I was trying not to make it about my personal opinion but there is an obvious bias that probably showed.

My personal opinion is that cloud computing has a great future for conventional PC's but will never have a real use in carpcs. The internet on the other hand could have some cool uses to complement conventional front ends but personally I don't believe it will ever replace it.

The reason comes down to this simple question-what happens in an area with no cell service?

Internet based navigation could never work in this scenario because the last thing you want is to make a wrong turn and have no cell signal to get you back on track. Quite a few other "mission critical" services would need to be available offline - and if they work offline whats the incentive for them to be online.

Long story short-I hear lots of comments along the lines of "there are so many reasons internet will change carputing" but have yet to hear an actual reason that makes me say "wow i need internet in my car". I wanna hear as many users as we can posting ideas and see if anything innovative turns up.

I'd have to agree with Chunky, there just isnt enough of a driving reason to have internet in the car at the moment. As bugbyte listed, there are huge possibilities the things that you can do with internet connectivity, but none of them really justify the cost to the everyday user. Perhaps one day when costs go down, and there becomes a reason to justify it people like Chunky and I will get internet in our car, but until then I truly believe that designing a frontend with a net-centric thought in mind isn't such a good idea unless you parallel develop it along with the standard offline view.

Don't get me wrong, I'm not saying it shouldn't be developed, I'm just saying don't focus on it until internet in-car is more of a mainstream thing. Just to throw an example out there, there is nothing wrong with a google maps plugin for instance... but it's 100% useless to the majority of users so I don't believe that it should be focused on as core functionality.

No question this is a non-starter for the moment. But I have to tell you, I've been experimenting with my own little CloudCar via the iPhone that mp3car have so kindly been paying for me to use and I have to say that even though I live in an Edge only area and often make drives where I lose connectivity, it isn't all that big a deal.

For example, I've been using the xGPS app and the AT&T navigator app. Both rely on a network connection to provide map info. Sometimes I don't have a net connection. However, I always have the purple guidance line, even when it is over an area that is blank. The nature of the signal loss is generally infrequent and the outages short. It's inconvenient, but only in a few cases is it catastrophic. Like when you need to start the trip but aren't in service range.

With ever increasing reliability of data networks, these outages will become less and less common. However, they will still occur and you all make excellent points about the comm failure modes that are necessary. I just don't see them as showstoppers.

Here's the type of thing I'm currently enjoying in my CloudCar via iPhone hack job:
1. No synching of music necessary. I stream it from my home computer over the web, keeping it up to date at all times. All playlists are available and dynamically updated. If I change it on my home system, it shows up on the iPhone because I access it through a web interface (product is called Dot.tunes)

2. Fallback for failure in streaming (and this does happen regularly) is local storage of some songs and local playback.

Why can't these things be combined? Obviously, I have some space on the iPhone for music. When I select a playlist and I have data connectivity, why shouldn't my phone take advantage of the connection and pre-load as many songs in the playlist as it can before it loses the connection? There's the issue of what happens if I switch playlists, but again, couldn't I load the first song in all my playlists? Couldn't it check to see if that song is resident on my iPhone and playback the local copy if it is there while it downloads the next song in the list? That way, if I drive out of range, I'll have a cache of songs to play until I drive back into range.

Why couldn't nav work the same way? xGPS includes a desktop app manager that lets you download a user-selectable amount of data to the phone so you don't have to rely on the network in, say, your home city. Why couldn't it work the same way as the music caching approach? When I plug in my nav start and end points, first download detailed data for the next couple of miles and then rough data for the balance of the trip. Then, download as much data as possible while the connection is good. That would probably work most of the time.

In fact, why wait? Why not download gps data as I drive along even if I'm not using the gps app? That way I have data when I need it, connection or not. Failure mode - occasionally.

Partial Connectivity and a Vision

Again, you make good points about cost and utility that aren't immediately realizable for most hobbyists. Let's turn towards how partially networking your car would do. We'll stick with WiFi for the moment. It will work only in your driveway or when at a hotspot.

Imagine a website - call it iDashboard. It allows you to log on with a user ID and gives you access to some control over your front end's configuration. You can find skins that others have published and 'load' them into your account, try them out to see how they look, then commit them to your computer. When you start your car and boot up it checks to see how it should be configured, then loads the configuration from the web site.

The web site also functions as a place to upload and store data like gps tracks, OBDII log data, playlists and so forth. The intent is to make it easier to configure and consolidate settings and functions on the car PC.

Later, as connectivity begins to improve and become more common, the site could function as a gateway for web apps and servers to offer services to the user. If it were a commercial site, I don't see why a company like Toyota or Nissan wouldn't want to offer extended warranties to customers who are willing to allow them access to OBDII data. They would start a relationship with them while at the same time plugging them into their dealer network and schedule them for maintenance. If they were smart, they'd allow anyone with a car to plug in and use the services so they could identify potential future customers, make them aware of sales promotions and so on.

Owners with a particular interest in a certain kind of car like a Corvette could swap and share themed skins for their car or download professionally designed ones by GM. Services to track 0-60 times, skidpad G info, autocross results or whatever, could be added by regular folks with an interest in it.

Hmmm. I think rather than hijacking Kev's thread, I probably ought to make my own.

Bugbyte's list is why I said "What is going to be the killer feature, or group of killer features"; No single item listed by bugbyte is worth it [yet, at least], but together they may add up eventually. The problem is that in bug's list, each single item can already be done "well enough" by something else.

The reason comes down to this simple question-what happens in an area with no cell service?

Yes. This is the other big thing. And as mentioned, the one time you really care is the one time that it's not there - by definition, because it's when I'm stuck in the boonies without internet that I really need internet :-)


So, another thing to ponder is mesh networks. Is there going to be value of mesh networks [each car is an ad-hoc wireless node] rather than an internet connection? Obviously there's a very clearly defined critical mass, below which it's [almost] worthless, and above which it's useful. But perhaps there are useful things in there;
1) You could be made aware of traffic conditions instantaneously further up the road, and get off much earlier. No internet connection and no fees beyond the initial hardware.
2) Central to obdgpslogger's original idea was being able to bind GPS data to your OBDII data, something that no-one else had done. What if you could measure the collective oxygen ratios of all the cars in an area, and use it to map out exact air conditions, or likely fuel efficiency things?


Also, the other common thread is one of privacy; I'm simply not ok with my car GPS and OBDII information being made publicly available as it's happening. I do fear that always-on internet in cars will lead to a catastrophic loss of privacy.


My inner nerd feels that there's something right, there that I just can't figure out, that would be some kind of killer app for car mesh networks.

Gary (-;