2015-05-11, 21:43
I could not find any specific discussion of this earlier and has to do with treatment of User Agents by mobile carriers.
The standard practice in the existing scrapers appear to be to use a standard desktop web browser user agent so it will work on all devices where the code runs. This way one does not have to deal with sites that may return different formats depending on user agents.
I noticed on the T-Mobile forums a post that seems to suggest that T-Mobile uses the User Agent string to decide whether to consider the data usage as use of tethered devices (non mobile agents are counted as tethered data users even if running on the app). The responses to that post suggest that changing user agent string is just bad programming.
What is Kodi community position on this?
I was able to verify the claim on my own T-Mobile phone by changing the user agent on Firefox web browser running on the phone. As soon as you change to a non-mobile agent, any HTTP (not HTTPS) traffic the app generates counts against the tethering limit. If the tethering limit is reached, T-Mobile returns a redirect page asking people to sign up for hotspot service or to purchase more data. Clearly, all metadata scrapers will stop functioning in such a situation as if the site being scraped is not available with no further information to the user as to what the problem is.
This is very bad for users of Kodi on cellular devices (at least on T-Mobile) with any such meta scrapers in use. People who use these legal plugins have no way of knowing that some plugins may result in their tethering data limits being used up (as opposed to just be counted as mobile data which is irrelevant for users with unlimited data or much higher data limits than typical tethering limits) with no expectation that such a thing will happen even if they are not tethering. They expect it to count against their data plan but not against tethering limits or restrictions.
Of course, developers can point fingers at T-Mobile and T-Mobile can point fingers at developers but this does not help any users enjoying Kodi as their media app on mobile devices.
I looked at some of the meta data scrapers out there and most seem to use Windows FireFox string and so will trigger tether data usage on T-Mobile when Kodi is installed on a T-mobile device (and other carriers doing similar).
Is there a guideline that can be provided to developers so they do not create this problem for users of their scrapers? Should Kodi community take any responsibility for working around this?
Should the scrapers provide a clear notice that its use on mobile devices may use up tethering data and will not function if tethering capacity is not available on some carriers? Most users aren't tech savvy enough to deduce this on their own.
Should the scrapers be written to handle responses for both mobile agents and desktop agents from sites they are scraping as best practice (though it seems onerous)?
Just posing open questions for discussion not suggesting any solution or blaming anybody.
Thanks
The standard practice in the existing scrapers appear to be to use a standard desktop web browser user agent so it will work on all devices where the code runs. This way one does not have to deal with sites that may return different formats depending on user agents.
I noticed on the T-Mobile forums a post that seems to suggest that T-Mobile uses the User Agent string to decide whether to consider the data usage as use of tethered devices (non mobile agents are counted as tethered data users even if running on the app). The responses to that post suggest that changing user agent string is just bad programming.
What is Kodi community position on this?
I was able to verify the claim on my own T-Mobile phone by changing the user agent on Firefox web browser running on the phone. As soon as you change to a non-mobile agent, any HTTP (not HTTPS) traffic the app generates counts against the tethering limit. If the tethering limit is reached, T-Mobile returns a redirect page asking people to sign up for hotspot service or to purchase more data. Clearly, all metadata scrapers will stop functioning in such a situation as if the site being scraped is not available with no further information to the user as to what the problem is.
This is very bad for users of Kodi on cellular devices (at least on T-Mobile) with any such meta scrapers in use. People who use these legal plugins have no way of knowing that some plugins may result in their tethering data limits being used up (as opposed to just be counted as mobile data which is irrelevant for users with unlimited data or much higher data limits than typical tethering limits) with no expectation that such a thing will happen even if they are not tethering. They expect it to count against their data plan but not against tethering limits or restrictions.
Of course, developers can point fingers at T-Mobile and T-Mobile can point fingers at developers but this does not help any users enjoying Kodi as their media app on mobile devices.
I looked at some of the meta data scrapers out there and most seem to use Windows FireFox string and so will trigger tether data usage on T-Mobile when Kodi is installed on a T-mobile device (and other carriers doing similar).
Is there a guideline that can be provided to developers so they do not create this problem for users of their scrapers? Should Kodi community take any responsibility for working around this?
Should the scrapers provide a clear notice that its use on mobile devices may use up tethering data and will not function if tethering capacity is not available on some carriers? Most users aren't tech savvy enough to deduce this on their own.
Should the scrapers be written to handle responses for both mobile agents and desktop agents from sites they are scraping as best practice (though it seems onerous)?
Just posing open questions for discussion not suggesting any solution or blaming anybody.
Thanks
