Birgitte_sb@yahoo.com, 23/07/2012 19:27:
On Jul 23, 2012, at 7:42 AM, "Federico Leva (Nemo)" wrote:
Birgitte_sb@yahoo.com, 23/07/2012 14:28:
I am unaware of what the shortcomings of the current system are and where any improvements would be felt. This makes it a bit hard to have a firm opinion of the trade-offs involved with changing the system. So what exactly are the problems people are having with the current geolocation system?
As the page tries to prove, looks like the current system is completely unreliable and therefore useless for most geonotices in Italy and probably other places.
I think it would be useful to have a wider study of the accuracy of the current system. Privacy issues are always a concern. I am not certain I could support gathering more exact information on users who are well-served by the current system. It would be more supportable, I think, if there were a way to turn on the browser-based system only for those who are in areas that are known to be poorly served by the current system. Or if you were to ask those who geolocate to known ambiguous areas to opt-in to browser-based geolocation. There is obviously a benefit for some people, but a cost to everyone if we were to switch wholesale. Further study to determine exactly how widespread and how significant the benefit would be is something that I think might be useful.
What if the new system happened e.g. to be needed for geonotices (to distinguish regions within a country) but not fundraising (which so far cares only about country, for currency/language/payment/legal purposes)?
Nemo
On Jul 23, 2012, at 12:48 PM, "Federico Leva (Nemo)" nemowiki@gmail.com wrote:
Birgitte_sb@yahoo.com, 23/07/2012 19:27:
On Jul 23, 2012, at 7:42 AM, "Federico Leva (Nemo)" wrote:
Birgitte_sb@yahoo.com, 23/07/2012 14:28:
I am unaware of what the shortcomings of the current system are and where any improvements would be felt. This makes it a bit hard to have a firm opinion of the trade-offs involved with changing the system. So what exactly are the problems people are having with the current geolocation system?
As the page tries to prove, looks like the current system is completely unreliable and therefore useless for most geonotices in Italy and probably other places.
I think it would be useful to have a wider study of the accuracy of the current system. Privacy issues are always a concern. I am not certain I could support gathering more exact information on users who are well-served by the current system. It would be more supportable, I think, if there were a way to turn on the browser-based system only for those who are in areas that are known to be poorly served by the current system. Or if you were to ask those who geolocate to known ambiguous areas to opt-in to browser-based geolocation. There is obviously a benefit for some people, but a cost to everyone if we were to switch wholesale. Further study to determine exactly how widespread and how significant the benefit would be is something that I think might be useful.
What if the new system happened e.g. to be needed for geonotices (to distinguish regions within a country) but not fundraising (which so far cares only about country, for currency/language/payment/legal purposes)?
I already was thinking it was as you said. I can't see why I would feel any different about using it for fundraising purposes, and I think we already use the separate browser data rather than geolocation to identify language. If anything I might be inclined to think a person would find it more desirable to know it.WM is hosting an event in their city, than to learn that their money is wanted in more targeted way. Not to put down fundraising, but I think people really like to know about local events. I certainly enjoy these notices. Maybe big city folk are too jaded to feel this way, but imagine that many other people must enjoy this too. I know whenever I see a local event mentioned on some big website, I always think of Judy Garland (if you have ever seen the movie "Meet Me in St. Louis") saying, "I can't believe it. Right here were we live - right here in St. Louis!" I don't get that magical feeling from fundraisers! So I definitely believe what it.WM wants to do, to connect people with local events, has real value. And that it has value for the individual people just as much as for it.WM.
The main question is whether the benefit from being able to connect people with local events is worth the risk of collecting more personalized of their data than we are accustomed to handling. Maybe the benefit does win out for many people in Italy (I don't really understand enough about what degree of improvement you are anticipating to have a firm opinion). But it is certainly not worth the risk for people in areas that do not notice problems with the current system. This is why I am suggesting that the browser feature might only be limited to areas that are known to reach some pre-defined level of error under geolocation. Or else that it be made an opt-in feature (perhaps even advertised through the current geonotice in areas that are known to be a problem). However I don't believe that gathering more browser data for everyone everywhere is a likely to be good overall solution.
Birgitte SB
2012/7/24 Birgitte_sb@yahoo.com:
[...] So I definitely believe what it.WM wants to do, to connect people with local events, has real value. And that it has value for the individual people just as much as for it.WM.
strong +1.
The main question is whether the benefit from being able to connect people with local events is worth the risk of collecting more personalized of their data than we are accustomed to handling.
I could be wrong but I don't think we would "handle more data" than what we are doing now. We are not going to use that data and as far as I know that data "dies" in the moment the system has output the message.
Maybe the benefit does win out for many people in Italy (I don't really understand enough about what degree of improvement you are anticipating to have a firm opinion). But it is certainly not worth the risk for people in areas that do not notice problems with the current system.
I will say that this seems fair enough for me, from what I have seen I can say that the current system is working on wide flat areas with enough internet provider servers coverage, this could vary from country to country for a number of reasons (from morphology to broadband penetration, etc.).
This is why I am suggesting that the browser feature might only be limited to areas that are known to reach some pre-defined level of error under geolocation. Or else that it be made an opt-in feature (perhaps even advertised through the current geonotice in areas that are known to be a problem). However I don't believe that gathering more browser data for everyone everywhere is a likely to be good overall solution.
Yes, opt-in is a viable solution but I think it spoils the best from the system, in fact with opt-in we would only reach people that: * are registered users * are expert enough to know that such gadgets exist and know how to activate them (I didn't know this myself till some months ago although I'm a Wikipedia editor from 4+ years). and I expect this category of users are already participating or already know how to get informations about meet-ups and local events even if are buried in obscure pages.
Anyway, I would be very glad of seeing a new study, comprehending more countries but I don't have enough time (and probably all the necessary competences) needed for it, maybe we could some professor interested on doing such a study (I think it could well be a research work for an undergraduate student or a bachelor thesis).
Do you think this could work?
Cristian
On Jul 24, 2012, at 11:14 AM, Cristian Consonni kikkocristian@gmail.com wrote:
2012/7/24 Birgitte_sb@yahoo.com:
[...] So I definitely believe what it.WM wants to do, to connect people with local events, has real value. And that it has value for the individual people just as much as for it.WM.
strong +1.
The main question is whether the benefit from being able to connect people with local events is worth the risk of collecting more personalized of their data than we are accustomed to handling.
I could be wrong but I don't think we would "handle more data" than what we are doing now. We are not going to use that data and as far as I know that data "dies" in the moment the system has output the message.
Maybe this is the area that needs more study. And I am probably the wrong person to try and even formulate technical questions, but is there a way to make use of this data without storing it? Without even knowing who recieved what personalized messages, unless, of course,they choose to respond? There is still the "creep" factor, in that readers of the messages will not necessarily know that it was all handled blindly. However it should technically preserve the reader's privacy, if no person nor computer can recall what messages any person was presented with. And if an FAQ is well-linked, anyone who under the misperception that it was handled invasively can learn how serious we are about the subject.
If such a thing is even possible I suppoose it becomes a sort of philosophical question. If no one nor computer can know whether or not you recieved a personally targeted message is the targeting invasive of your privacy?
Birgitte SB
2012/7/24 Birgitte_sb@yahoo.com:
On Jul 24, 2012, at 11:14 AM, Cristian Consonni kikkocristian@gmail.com wrote:
2012/7/24 Birgitte_sb@yahoo.com:
[...] So I definitely believe what it.WM wants to do, to connect people with local events, has real value. And that it has value for the individual people just as much as for it.WM.
strong +1.
The main question is whether the benefit from being able to connect people with local events is worth the risk of collecting more personalized of their data than we are accustomed to handling.
I could be wrong but I don't think we would "handle more data" than what we are doing now. We are not going to use that data and as far as I know that data "dies" in the moment the system has output the message.
Maybe this is the area that needs more study. And I am probably the wrong person to try and even formulate technical questions, but is there a way to make use of this data without storing it? Without even knowing who recieved what personalized messages, unless, of course,they choose to respond?
I will forward these questions to wikitech-l.
There is still the "creep" factor, in that readers of the messages will not necessarily know that it was all handled blindly. However it should technically preserve the reader's privacy, if no person nor computer can recall what messages any person was presented with. And if an FAQ is well-linked, anyone who under the misperception that it was handled invasively can learn how serious we are about the subject.
A FAQ page would be useful, then I think we could decide that, for example: * the geonotice can be centered only on big cities (e.g. "Chef-lieus") * Messages should not make use of the geolocalization information itself, but only static test (i.e. NO "Hey you living in X, there a meetup in Milan" - YES for "There's meetup in Milan")
If such a thing is even possible I suppoose it becomes a sort of philosophical question. If no one nor computer can know whether or not you recieved a personally targeted message is the targeting invasive of your privacy?
yes, that's pretty much my point, I think nobody is interested to know if somebody has seen a geotarget message or not, my point is about making people from Milan (north of Italy) know that there is a local meetup without bothering people from Rome (center) or Palermo (south of Italy). (please note that with the current system we were not even able to accurately distinguish betwen macro-regions as north/center/south Italy)
Cristian
On Wed, Jul 25, 2012 at 4:20 PM, Cristian Consonni kikkocristian@gmail.comwrote:
2012/7/24 Birgitte_sb@yahoo.com:
On Jul 24, 2012, at 11:14 AM, Cristian Consonni kikkocristian@gmail.com
wrote:
2012/7/24 Birgitte_sb@yahoo.com:
[...] So I definitely believe what it.WM wants to do, to connect
people with local events, has real value. And that it has value for the individual people just as much as for it.WM.
strong +1.
The main question is whether the benefit from being able to connect
people with local events is worth the risk of collecting more personalized of their data than we are accustomed to handling.
I could be wrong but I don't think we would "handle more data" than what we are doing now. We are not going to use that data and as far as I know that data "dies" in the moment the system has output the message.
Maybe this is the area that needs more study. And I am probably the
wrong person to try and even formulate technical questions, but is there a way to make use of this data without storing it? Without even knowing who recieved what personalized messages, unless, of course,they choose to respond?
I will forward these questions to wikitech-l.
There is still the "creep" factor, in that readers of the messages will
not necessarily know that it was all handled blindly. However it should technically preserve the reader's privacy, if no person nor computer can recall what messages any person was presented with. And if an FAQ is well-linked, anyone who under the misperception that it was handled invasively can learn how serious we are about the subject.
A FAQ page would be useful, then I think we could decide that, for example:
- the geonotice can be centered only on big cities (e.g. "Chef-lieus")
- Messages should not make use of the geolocalization information
itself, but only static test (i.e. NO "Hey you living in X, there a meetup in Milan" - YES for "There's meetup in Milan")
If such a thing is even possible I suppoose it becomes a sort of
philosophical question. If no one nor computer can know whether or not you recieved a personally targeted message is the targeting invasive of your privacy?
yes, that's pretty much my point, I think nobody is interested to know if somebody has seen a geotarget message or not, my point is about making people from Milan (north of Italy) know that there is a local meetup without bothering people from Rome (center) or Palermo (south of Italy). (please note that with the current system we were not even able to accurately distinguish betwen macro-regions as north/center/south Italy)
There also are infoboxes and categories that people use to indicate where they live.
It could also be a preference, to make it private info. (and would be nice to get geonotices for "home" even while I am traveling, or if I use a proxy or something)
Cheers, Katie
Cristian
Wikimedia-l mailing list Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l
2012/7/25 aude aude.wiki@gmail.com:
There also are infoboxes and categories that people use to indicate where they live.
I think this solutions has the same drawbacks of using an opt-in gadget.
It could also be a preference, to make it private info. (and would be nice to get geonotices for "home" even while I am traveling, or if I use a proxy or something)
I think instead this could work better than the previous because on the first access one can review its preferences.
I also think there should be an optiont to say "please do not bother me with geotargeted messages" i.e. an opt-out option.
Cristian
wikimedia-l@lists.wikimedia.org