[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: CoreLocation clone

From: Ivan Vučica
Subject: Re: CoreLocation clone
Date: Tue, 4 Jun 2013 13:14:06 +0200


My GSoC project for this year is implementing Opal (that is, Core Graphics) 
backend for GNUstep-GUI, so that part should not be a problem.

Right now I'm not very familiar with current drawing API (I'm about to become 
VERY familiar with it) so I'm not sure, but it might also be enough to draw 

I'd also suggest looking into implementing CATiledLayer in QuartzCore :-)


Ivan Vučica
via phone

On 4. 6. 2013., at 08:27, Maxthon Chan <address@hidden> wrote:

> Here is a list of tasks to finish to make this into being:
> 1) Enhance CoreBase, probably fuse it with Base or move a significant portion 
> of code from Base, so that CoreGraphics will work properly;
> 2) Implement CoreGraphics on top of Back and link against CoreBase;
> 3) Rewrite GUI to employ the new CoreGraphics API;
> 4) Implement CoreLocation. This may spin off a subproject to maintain an open 
> WLAN AP geolocating server using data from OBM
> 5) Implement MapKit that will link against GUI, CoreLocat;ion and 
> CoreGraphins. CG is used to generate maps.
> Tasks 4 can run parallel with tasks 1, 2 and 3.
> 在 2013-6-4,下午2:06,Sebastian Reitenbach <address@hidden> 写道:
>> On Tuesday, June 4, 2013 02:45 CEST, Ivan Vučica <address@hidden> wrote:
>>> On 4. 6. 2013., at 01:02, Maxthon Chan <address@hidden> wrote:
>>>> From what I know, you can use Android location services without Google 
>>>> Maps app as tried on a home-brew Android device - this meant that probably 
>>>> the locating API is not in the Google Maps app but in the system itself, 
>>>> and can be safely extracted.
>>> Indeed? What backend does that use? :-)
>>> If the homebrew device has GPS on it, then the location API's functionality 
>>> does not surprise me. If it, however, does not, then listening to traffic 
>>> with Wireshark is in order to figure out whose terms of use you want to 
>>> study before reversing the protocol :-)
>>> From what I can tell, you can't actually check this on a retail device 
>>> which comes with Google services, because these services can also provide 
>>> information to the "android.location" API, not just the "Google Location 
>>> Services for Android":
>>> If it turns out it is based on WLAN stations, it still almost certainly 
>>> uses an API endpoint that comes with terms of service. And if it's Google 
>>> Maps Location API, then all I can say is -- just because a blind eye is 
>>> turned when a Google-backed open source project (that is, the "Android Open 
>>> Source Project" as it's called) uses the API - well, that doesn't mean 
>>> they'll like GNUstep using it.
>>>> For OpenStreepMap and OpenBmap, this would require a cooperation between 
>>>> projects - they should be willing to back up our CoreLocation and MapKit 
>>>> implementation.
>>> Yes. From what little I know about OSM, it's primarily a data source; so we 
>>> would also need someone to provide tiles (the rendered images themselves). 
>>> An iOS app I know offers a choice between about 8 different tiling 
>>> providers; it may be worth looking into that.
>> I don't think you'd need an extra tiles provider. OSM offers an API to query 
>> its database, and you get back an XML snippet describing the map. You then 
>> can just render it on your own.
>> I use qlandkartegt a lot, which is doing exactly this.
>> cheers
>> Sebastian
>>> By the way -- I've just remembered MySTEP also has a MapKit implementation 
>>> which happens to use OpenStreetMap:
>>> Now, OpenBmap. OpenBmap doesn't actually provide a complete service; it 
>>> primarily provides a database of MAC-to-geolocation mappings. This is great 
>>> for what is needed for Core Location; one just needs to develop a good 
>>> triangulation algorithm based on relative signal strengths of multiple 
>>> MACs. Upload a list of MACs and signal strengths, get back the geolocation. 
>>> :-)
>>>> Have you ever noticed that there are lots of 3G network adapters for 
>>>> computers that plugs onto a computers USB or ExpressCard port? For some 
>>>> computer models (like Dell Latitude D620, as I have a really old one still 
>>>> alive, kicking and serves my blog - shameless plug, 
>>>> - to everyone over the Internet) there are even 
>>>> built-in ones?
>>> Sure. Can you issue an AT command that will list all the cell tower 
>>> identifiers along with signal strength toward that particular tower? Can 
>>> you issue an AT command to get an identifier of just the tower you're 
>>> connected to, so you can combine that with the current connection signal 
>>> strength? :-)
>>> That's the problem: you may be able to find the AT command on a particular 
>>> dongle, but it's bound to be nonstandard and probably won't exist on 
>>> whatever garbage operators around the world tend to sell. (For example, I 
>>> had some... 'experience' with a dongle that came with OS X 10.5 drivers. 
>>> These included libcurl... which, when overwritten on 10.6, broke 
>>> Not a nice situation to be in. Even worse is spending a 
>>> few hours on diagnosing that this is, indeed, the problem. *sigh*)
>>>> The MapKit for OS X in question seemed used only public Google Maps API. 
>>>> It is freely licensed as well, no key required and the author himself is 
>>>> not in trouble as well.
>>> Did you read Google Maps API terms of service and, more importantly, the 
>>> billing information? Aside from decision to go "thermonuclear", financials 
>>> are considered one of the reasons why Apple opted for rolling the infamous 
>>> Apple Maps service.
>>> Regarding the author of MapKit for OS X being sued: just because a tiger 
>>> turns a blind eye doesn't mean you should go poking him with a stick :-)
>>> Plus, aside from Google Maps API, you need to consider Google Maps Location 
>>> API, which has no free tier apart from 100 requests a day. Hardly enough 
>>> for any serious use.
>>> On 4. 6. 2013., at 01:07, Maxthon Chan <address@hidden> wrote:
>>>> And by what I say as crowd-sourcing, I mean ask users that are willing to 
>>>> share for information. What we are asking for here is several numbers: AP 
>>>> BSSID, transmission power, and its geological location (latitude and 
>>>> longitude, as well as height). The request is done by the device uploading 
>>>> a list of BSSIDs received and its signal strength, hence BSSID never 
>>>> leaves our servers, and thus privacy issue avoided.
>>> Sure; so the only people violating the privacy are -- we ourselves? :)
>>> Plus, what happens when someone queries for location of MAC "XYZ"? We 
>>> reveal the location -- which is the purpose of the API. Sure, that happens 
>>> already: Google's doing it, Apple's doing it, Skyhook's doing it, OpenBmap 
>>> people seem to be doing it. I'm just saying that it's private information, 
>>> in a way. :-)
>>> from the things-we-unknowningly-reveal dept:
>>> I really liked it when, at my former workplace, we received a crash report 
>>> for a game on OS X. Something ticked my curiosity and I playfully passed 
>>> into Skyhook the several listed AP MAC addresses that OS X dutifully 
>>> included in "full system report". Imagine my surprise when it turned out 
>>> that Skyhook claimed the location was just across Apple's campus...
>>> ...and yes, it did turn out it was a crash report from an Apple employee.
>>> --
>>> Ivan Vučica
>>> address@hidden -
> _______________________________________________
> Gnustep-dev mailing list
> address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]