Posts: 1,544
Joined: Jul 2007
Reputation:
27
gabbott
Team-Kodi Member
Posts: 1,544
Well as far as scrapers go, I'm sure the site you would be scraping from (e.g. imdb, themoviedb.com...) would be able to see your IP address since you are grabbing info from their site. As far as xbmc itself "phoning home" to the developers... I don't think you have to worry about that.
Posts: 26,215
Joined: Oct 2003
Reputation:
187
What XBMC sends is basically the title of the movie, based on a cleaned version of your filename. Obviously the other end will have your IP address details, and (should they want to) could log all the requests you do from that IP address and build some sort of an idea of what movies you have in your collection. In addition to IP address, they'd also see that it was coming from XBMC based on the useragent string. If you use .nfo files with URLs in them, then all that they'll see is that your IP address is fetching the URL using XBMC - i.e. the filename information is not sent at all. You're welcome to use something like wireshark to confirm this for yourself.
No other user information is sent anywhere.
This is identical to you googling for those same movies, or searching for them directly on IMDb or themoviedb.org's websites.
As for what XBMC collects: We don't collect anything whatsoever if we can help it. The only idea thing we get from XBMC users is:
1. how many are hitting the RSS feed on feedburner.
2. basic website usage statistics collected via google analytics.
Cheers,
Jonathan
Posts: 26,215
Joined: Oct 2003
Reputation:
187
Last time I looked it was about 80-90k/day - it's steadily increasing. I guess that gives some idea of the number of users (who have the RSS feed on and haven't changed it) that either load XBMC up or idle on the homepage on a day to day basis.
Cheers,
Jonathan
Posts: 7
Joined: May 2010
Reputation:
0
@ jmarshall
Thanks for your replay,
if i get this right, then what i'd be lookin for would be a way to use the method with implementing URLs in the nfo file but sending a request not showing that the request comes from XBMC. Is that possible? Maybe by using some other software to scrape the files?
Posts: 669
Joined: Aug 2006
Reputation:
2
Right now TheTVDB doesn't compile any sort of stats. It's not something that's out of the question in the future, though, since I think Neilsen-type ratings for digital viewing would be quite interesting. Before we took that step we'd have to discuss the method and implications with all of our API users (XBMC, Media Portal, etc). I've also envisioned some sort of opt-in system for this, but I doubt it'd get used much.
Another option would be an opt-out system that would allow people to opt-out immediately on their first scrape. The different API users would have to get on board with it, though. The benefit is that we'd be able to completely ditch the log files and simply work via anonymized user keys, meaning it'd actually be more private than a us working from our logs.
There are a few things to consider with this:
1. You cannot be sued for possession of copyrighted media, only for distributing it (this is why we see torrent users getting sued but not Usenet users). Not a privacy matter, but related.
2. There's no way for a site like ours to know if you actually have the media or not. For all we know you could be compiling a database of series that interest you, but without actually having the media.
3. There's no way of knowing how someone got the media even if we assume they have it. Did they record it themselves? Purchase from iTunes?
4. I understand privacy concerns, but frankly most sites don't care one bit what individual users are doing. They might want compiled data about all of their users. This isn't a big selling point to privacy advocates, but it's still the truth.
5. Having the networks know which series are popular for digital viewers and the extent of that popularity might allow shows to continue even if they weren't popular in the mainstream. Think of the following: Firefly, Defying Gravity, Dresden Files, etc. Once again not a privacy matter, but it's all loosely connected.
I'll soon be creating a thread on our site asking for feedback on this, since I want to keep the community happy while still researching ways that we can better support our site.
In the meantime, perhaps you should look at an anonymizing proxy if you're that concerned about it. Scrapers are simply website requests, so it wouldn't be very difficult. If you don't want to do that, your best bet is to just create nfo files manually. If your collection isn't huge it probably wouldn't take too much time.
Posts: 7
Joined: May 2010
Reputation:
0
Thanks again guys,
if i understand it correctly, a way could be to reroute the requests via a, let's say trusted service/site, e.g. a xbmc server sytem, managing request and sendback, followed by "hardwired" immediate deletion of the requesting user's ip-adress, after all packages have been confirmed as recieved. Theoretically, that should be possible, right? Being aware of facts like maintenance expence and the sort of cause...
And while making users feel better about their private matters it would equally provide a popularity barometer, while, of cause suffering from the same weeknesses, concerning the datas potential to act as referee, concidering the fact that requests can be generatd by maschines...
