reposting a private message to anathema to kick off the public dialogue on this and open up discussion...



****************

Richard,



i drafted a brief summary of the opportunity as I see it here.



As far as a feature spec, my weekend is pretty shot with a wedding today and a bunch of other crap tomorrow so I won't have anything for you before next week. If you want to take a crack at building a simple app for the phone that allows you to extract the location data and send as XML or some other neatly-packaged format to a server, that's really all the device piece needs to do at this point to prove the concept. We'll need to think a bit about what that message format should look like since it should ideally be standardized across all devices and i have no idea how these things send location data (ie is it some representation of strength of signal between cell towers? if so it's gotta be converted to a gps coordinate at some point to be useful...). Anyways, I'm assuming you have some experience with all this.





FYI: in all likelihood I will not be the day-to-day Idea Champion on this one if it takes off. I run two tech startups right now and that eclipses most of my time. if this concept works and is big as I imagine it could be, it will need somebody with more daily bandwidth to drive it. My goal is to help put some forward motion to it initially, see what type of reception it gets and if it looks promising, to then find that person and take more of a strategic hands-off role than a day-to-day project manager role. Just wanted to let you know my situation so you can decide whether it's still something you're interested in working on it. I have no idea how the compensation system in CH works these days but my philosophy is that compensation/ownership should be directly tied to value of contribution- so however the best way to achieve that is how it will be set up.



Anyways, lemmeknow if you get something together and I'll try and round up the other technical people that I believe need to be involved to get the server piece. I didn't set up a Trac/SVN instance- you can use w/ the CH one for now and if enough people get involved we can probably convince CH to integrate Trac or move it to an external server to get the advantage of the ticket system.



sean

Posted: April 21, 2007, 12:44 pm

On April 23, 2007, 9:08 am anathema said:



Hi Sean,



I have spent a few hours today researching the availability of positional services to cellphone applications.



In Java as far as I can ascertain it is not possible to obtain positional data other than from a GPS device, either on board or over bluetooth. Of course Java could interface to native code but it would probably be as easy to implement the whole thing in native code.



In Symbian S60 it is possible to obtain cell data. Four values are obtained:



MCC: Mobile country code

MNC: Mobile network code

LAC: Location area code

CELLID: Based on the mast currently in use



There are various databases available which can translate these parameters into a geographical location. I have not fully investigated these yet but I will.



On S60 3rd edition which is the device which I favour there are security certification issues regarding access to positional data. While I have a rudimentary application which accesses this data running on my phone it will not run on your phone without developer certification specifically for your IMEI number. The way around this eventually is to apply for a commercial certification from Symbian



I have not yet researched other platforms such as BREW and Windows mobile.





Best regards,







Richard

Posted: April 24, 2007, 6:08 am

Sean,

I guess you don't have much time to devote to this at the moment but it has been continuing to bounce around my head and I feel like I need to get feedback on some ideas:

i) The application could integrate with twitter using its API to perform many of the tasks that would be required such as finding friends etc. No sense re-inventing the wheel and twitter has an already established user base to which we could market it.

ii) The range of bluetooth is too short to be really useful for locating people. Most times if they are in bluetooth range then they are already visible.

iii) I propose setting up a database which will give meaningful names to the Cellid data which can easily be obtained. Each user would have a home, maybe work and some other commonly visited places as would each group. The database would be small enough to be hosted locally on each user's phone with some sort of synchronisation being done over bluetooth when within range or possibly over the network.

iv) Use should be made of GPS features where they are available.

I don't know any more whether my ideas coincide with yours at all but I think they are now coherent enough to put together a prototype. Correct me if I'm wrong. I also have ideas for various other twitter related mashups, mobile and otherwise.

Hope to hear from you soon,


Richard

Posted: April 27, 2007, 4:02 am

Richard,
I believe you're right on the bluetooth signal not being strong enough for it to be useful- i just tried sync'ing my treo to my laptop from across my house and it didn't locate it. It would have to have at least 20yrds of range for this service to be meaningful if it's running over local bluetooth. You're probably right that it would need to utilize the LBS method.

If you can put together a functional prototype, go nuts with it. My vision was a bit different for how the service would work but this issue of the bluetooth signal dashes that idea so run with your ideas if you can make something worthwhile.

Piggybacking the Twitter user base is an excellent idea- like you said, no sense in trying to start from scratch and knowing Evan Williams' philosophy, I bet he'd be thrilled to have services built on top of Twitter.

As far as putting the db on the phone- I'm not sure I understand the significance of storing localities like "work, home, etc." The idea (as you know) is to have it work anywhere when you're out and about allowing you to discover when acquaintances are nearby and facilitate spontaneous encounters. If what you're proposing helps achieve that, go for it.

Let me know if you put something together. Again, I believe this is one that is fundable if it works. I envisioned the free version allowing you to discover only people that were explicitly in your network, while the premium version would allow you to discover strangers when both parties were running Feelrz in "permiscous" mode. The free version is valuable enough to download to connect with friends and there's an enticing-enough carrot to get some of those to convert to paying. I'm still concerned about the carrier fees for the LBS data under this scenario. what do you think?

sean

Posted: May 6, 2007, 7:34 pm

Carrier fees only apply, as far as I am aware,to actual meaningful location data: latitude/longitude etc . The point of the location tags as I see it is to lend meaning to the Cell ID data which is available for the cost of transferring a few bytes over the ether. (at least it is on my device and carrier. Testing would be required to assess its portability). Devices on different networks say would not know they were in close proximity due to the Cell ID data being different.

Posted: May 8, 2007, 2:06 am