The actual ASP.NET code in the web service is virtually zero. I have no idea what the web frontend will be written in. Python or PHP would be a good choice.Quote:
Originally Posted by n2k
justchat_t: I agree with your list and I would add to the list whatever Matt just committed that's not on the list :P. I know he committed some mysql stuff and a web service definition or two. I haven't looked at the code yet.
Code:svn co https://openicevis.svn.sourceforge.net/svnroot/openicevis openicevis
If I can just make a recommendation, you may want to just start by listing the resources you want to expose. Kev started that in post 4 I think. Then list the actions you want to allow on each resource. From there, you can start modeling the service contracts you'll need (most times these will be typical CRUD ops). For example:
Playlist
- Create
- Read
- Update (e.g. title)
- Delete
- Add Song
- Remove Song
For a RESTful architecture you can then figure out what your endpoints will be, based on the resources (e.g. /openvis/playlist, /openvis/media, /openvis/config, etc.). Your operations will then use HTTP methods (GET, POST, PUT, DELETE) and the data will either be in the URL or as the HTTP payload for POST and PUT ops.
From there you can then actually worry about how to implement those services in C#, Java, PHP, Python, etc.
Interesting concept!
Just skimmed rather quickly so ...
+1 on why asp/.net ... good stuff by proprietary based. Python/Ruby much better choice.
php is ... well, I'm try to get away from that world. Enough is enough already.
Now, why I'm really replying.
About 6 years ago I had similar ideas. At the time, no clue of the carpc world nor the iphone. I was thinking more along the lines of memory sticks ... ahem!
Now, web 2.0 is well along why don't you consider the web 3. approach?
I don't want to pooh pooh on the idea but why are you guys trying to create the wheel on your own?
Offer an "open" welcome to all front-ends,
use a LGPL type of licensing so that each and every one of you can charge services based off the code,
work towards building a system to control the development process,
set up the standards and API base,
get "real" project management in place,
and then get paid funding.
I read a few posts back about the worry of google going $$$ but even open source projects turn to $$$ to continue their work.
I'm also sitting here and wondering why not hit up the google chrome/droid folks up for help, funding / buy out ?
However, like in a similar thread regarding the death of the pc / long live iphone-ness.
You should really consider the phone/head-unit tie-up for hardware. Also, another reason to go *nix/chrome.
Also, like the post that mentioned getting user stats for app usage you really should ask for what the end user wants/needs. ie, I gave up on the idea of buying a cheapy Car Navi system because my iphone was better and because I'm better at finding short cuts than the dang machine. However, a carpc/ PHU (Phone Head Unit) is really want I want. Bluetooth phone service, a few mp3s, a movie or 2 and OBD since I'm a geek. However, my mum would love the gas mileage reports!
Sorry, it's late over here in sonyland and I'm rambling.
Good Luck guys !!! you definitely are headed in the right direction!
Forgot there was a svn..... just took a look. Looks like WPF is the way its going, and he has some service started called promixity that does file syncing?Quote:
Originally Posted by kev000
Yup post 4 was where kev listed a lot of services, I added the DTC and vin lookup service and then we decided to only focus on the top few to start (the 4 that don't exist anywhere else). I kind of combined your second and third step into one but thats pretty much where were at now. Definitions of the first four services and their functions.Quote:
Originally Posted by SFiorito
Your HTTP ops idea was what I had originally thought of too because of how easy it is to work with and how universal it is but I think WCF is already in svn.
Right. There is going to have to be some learning on my part and possibly on Matt's part if we are going to use RESTful services. Know on any good docs on how one uses REST with WCF (or without)? On the client side, it's super easy for me to generate code from a WSDL and then run... I'll google around and see how easy it will be for RESTful services as well.
Talking with Matt, the proximity service is a frontend to OpenICE's Proximity which does full folder/file syncing among other things.
Justchat, let me know your sf.net username and I'll add give you commit permissions if you want. Maybe you could write a simple example of what you'd like to see and upload it so Matt can emulate.
This is a good primer on REST, and then goes into writing RESTful services with WCF: http://msdn.microsoft.com/en-us/library/dd203052.aspx
Essentially, you need to use the WebGet and WebInvoke attributes on your services operations. UriTemplates define the parameters in your URL:
[WebGet(UriTemplate = "playlists/{name}")]
GetPlaylist(string name);
[WebInvoke(Method = "Post", UriTemplate = "playlists/{name}")]
UpdatePlaylist(string name, Playlist updatedPlaylist);
And so on.... Every language has their own framework to support RESTful services, but they all follow the same concepts. One thing to keep in mind is that WSDL is only for SOAP services. REST is meant to be very simple, so there is no meta data format for it. It's just HTTP ops and XML (or JSON for AJAX apps).
Ok question..... Can we somehow pull off push notifications? (i know theres 2 different ways its typically done)
The whole send address to car feature could be pretty cool... or should we just poll for it?
push can be done, but it's probably easier to poll for it.Quote:
Originally Posted by justchat_1
Hi guys,
Great concept!
I read all the posts from the begining but ended up with a few questions...
As far as I got the consept it's all about having different servers running / providing different services to clients based on interaction between the server and the client. Or is it just one relaying and passing on data and requests to other servers?
But would the serverside software be something that an end user could run on his own server at his home without having to have intimate knowledge about a full server setup? After all setting up and running a complete webserver is not always the easiest thing in the world...
It was mentioned in this thread earlier that some of the interesting services would be gps tracking but that users don't trust a server with these data. I'll assume that users would be way happier with uploading these type of data to their own server, rather than some other server somewhere (I know I would at least).
And another tought, don't limit this to the carpc world. This could work just as good for a smarthouse and smartphone interaction as it does for the carpc. It might even be times were it could be interesting to communicate between your home and your car.
Also think about an "offline mode", incase someone drives of into an area without net coverage and don't want to loose their tracking data.
Yep, there's no reason why you can't cherry pick services to run on your home server. It is open source after all. Many of the crowd sourcing services are only useful with a large community providing data (otherwise your traffic data is going to show your traffic condition which isn't very useful).
The Consumer applications that use the services will have to work out how to do syncing when offline. It shouldn't be that hard to do.