Most of the major Internet companies use Quova (Google, Yahoo, MSN, etc.) for their IP geolocation data (linky.
You don't need to have your ISP contact them.
Just send an email to support@quova.com with your IP and physical location. They used to provide weekly data updates, so I imagine it is equivalent or more frequent now.
OK, there are several things that I've found are missing from this conversation:
Quality of underlying map data. Google only provides the lower quality Teleatlas data through their API (even though they use both Teledata and Navteq for their maps.google.com service). This has been well-discussed on O'Reilly's blog
. Microsoft appears to provide high quality Navteq data through their API (based on the license marks shown on sites using Microsoft's APIs).
Currency of map data. All of the providers receive quarterly updates from Navteq and Teleatlas. I looked at the currency of their data a year ago and found that Google and Microsoft were equally current, but they both were behind Yahoo (faster updates). It is probably worth doing a similar analysis and determining if one or the other is keeping their data more current.
Advertising. I'm more familiar with Google's license than Microsoft's, however the Google license does have a few points that people should be aware of. Most notably they may include advertising in your maps.
You cannot charge consumers
You may not use the API if you charge consumers for content displayed on the maps. (Think you might want to create a premium section? You can't use Google maps). I don't know MSFTs terms.
They are both powerful services, and I'm definitely more familiar with Google's service and terms, but you definitely need to consider a lot more than which API is easier to use.
Yeah, if they really want to anonymize the data, they need to publicize the techniques they use.
Also, people are still missing several key items:
- Search History is turned on by default. These records are NEVER purged. - As far as I can tell from the article, this only applies to Search. Adsense, Gmail, Maps, etc. may be excluded from these new policies.
- They are still retaining all historical searches. So, they will always be able to look back at stats such as search term frequency, etc. They are not losing much in the way of long-term analytics.
I think people are largely missing what Google currently allows/prevents, the point of the article and the significance of the actions.
1. Google already polices activities that are likely to be illegal. Just look at the long list of activities that Google does not accept advertising on: https://adwords.google.com/select/contentpolicy.ht ml
Just a few examples:
- prostitution
- weapons
- gambling
- counterfeiting
oh, and copywrighted works
2. An account rep assisted the site owners with selecting keywords to bid on. The account rep suggested terms like "bootleg movie download" and "pirated".
So, Google accepted websites into Adsense that violate their own content policies, extended them credit and then helped them promote the websites via Adwords.
3. Why does any of this matter? Google is struggling to strike deals with video copywright owners to keep this content on Youtube / Google video. A key component of all of these deals is Google's ability (and of course willingness) to filter illegal material.
Google loses a lot of credibility when it can't ensure its employees abide by its promises. It looks even worse because Google profited from it. Using the metaphor from above, this is equivalent to a TV store that sold billboard space above the store that said where to go to find illegally stolen TVs. If you're a TV manufacturer, you're not going to what your TVs in that store.
I think it would help us ALL (especially the media) to become more familliar with patent law. When you read patents, you should focus ENTIRELY on the claims. The description is useful to provide context, but what is patented is in the claims, NOT the description.
Cliff notes for reading claims. Look for:
- parent claims. These claims have no dependency on other claims. In the Amazon patent, the claims that stand on their own are 1, 12, 13, 26 & 27. These are the claims that can stand alone and where you should focus most of your energy. - subordinate claims. These often begin with "the method described in claim X, where". These claims modify the parent claims. However, and this is key, if the parent claim is invalid, then so are ALL of the subordinate claims.
The patent covers the method described in the parent claim fully. If there is an "and" in the claim, then there is an "and" in what is covered. That means for something to infringe on the patent, it must meet ALL of the "ands".
In the case of the first claim, the method that is being patented requires a system that: - "receives an Interent order from an eletronic catalog" AND - estimates the date that the person would have interacted with the product. WHERE - this estimate is based at least on the product type. AND - sends a follow up email with a link. AND - uses collaborative filtering AND that review to present the customer with other products he might like.
While not unique, this approach is fairly narrow. Any system that does not use ALL of these methods is not covered by the patent. Consumer reviews are not patented, just this VERY specific approach to consumer reviews and recommendations.
Slashdot Mirror
Most of the major Internet companies use Quova (Google, Yahoo, MSN, etc.) for their IP geolocation data (linky. You don't need to have your ISP contact them. Just send an email to support@quova.com with your IP and physical location. They used to provide weekly data updates, so I imagine it is equivalent or more frequent now.
- Quality of underlying map data. Google only provides the lower quality Teleatlas data through their API (even though they use both Teledata and Navteq for their maps.google.com service). This has been well-discussed on O'Reilly's blog
- . Microsoft appears to provide high quality Navteq data through their API (based on the license marks shown on sites using Microsoft's APIs).
- Currency of map data. All of the providers receive quarterly updates from Navteq and Teleatlas. I looked at the currency of their data a year ago and found that Google and Microsoft were equally current, but they both were behind Yahoo (faster updates). It is probably worth doing a similar analysis and determining if one or the other is keeping their data more current.
- Advertising. I'm more familiar with Google's license than Microsoft's, however the Google license does have a few points that people should be aware of. Most notably they may include advertising in your maps.
- You cannot charge consumers
- You may not use the API if you charge consumers for content displayed on the maps. (Think you might want to create a premium section? You can't use Google maps). I don't know MSFTs terms.
They are both powerful services, and I'm definitely more familiar with Google's service and terms, but you definitely need to consider a lot more than which API is easier to use.Yeah, if they really want to anonymize the data, they need to publicize the techniques they use.
Also, people are still missing several key items:
- Search History is turned on by default. These records are NEVER purged.
- As far as I can tell from the article, this only applies to Search. Adsense, Gmail, Maps, etc. may be excluded from these new policies.
- They are still retaining all historical searches. So, they will always be able to look back at stats such as search term frequency, etc. They are not losing much in the way of long-term analytics.
I think people are largely missing what Google currently allows/prevents, the point of the article and the significance of the actions. 1. Google already polices activities that are likely to be illegal. Just look at the long list of activities that Google does not accept advertising on: https://adwords.google.com/select/contentpolicy.ht ml
Just a few examples:
- prostitution
- weapons
- gambling
- counterfeiting
oh, and copywrighted works
2. An account rep assisted the site owners with selecting keywords to bid on. The account rep suggested terms like "bootleg movie download" and "pirated".
So, Google accepted websites into Adsense that violate their own content policies, extended them credit and then helped them promote the websites via Adwords.
3. Why does any of this matter? Google is struggling to strike deals with video copywright owners to keep this content on Youtube / Google video. A key component of all of these deals is Google's ability (and of course willingness) to filter illegal material.
Google loses a lot of credibility when it can't ensure its employees abide by its promises. It looks even worse because Google profited from it. Using the metaphor from above, this is equivalent to a TV store that sold billboard space above the store that said where to go to find illegally stolen TVs. If you're a TV manufacturer, you're not going to what your TVs in that store.
I think it would help us ALL (especially the media) to become more familliar with patent law. When you read patents, you should focus ENTIRELY on the claims. The description is useful to provide context, but what is patented is in the claims, NOT the description.
Cliff notes for reading claims. Look for:
- parent claims. These claims have no dependency on other claims. In the Amazon patent, the claims that stand on their own are 1, 12, 13, 26 & 27. These are the claims that can stand alone and where you should focus most of your energy.
- subordinate claims. These often begin with "the method described in claim X, where". These claims modify the parent claims. However, and this is key, if the parent claim is invalid, then so are ALL of the subordinate claims.
The patent covers the method described in the parent claim fully. If there is an "and" in the claim, then there is an "and" in what is covered. That means for something to infringe on the patent, it must meet ALL of the "ands".
In the case of the first claim, the method that is being patented requires a system that:
- "receives an Interent order from an eletronic catalog" AND
- estimates the date that the person would have interacted with the product. WHERE
- this estimate is based at least on the product type. AND
- sends a follow up email with a link. AND
- uses collaborative filtering AND that review to present the customer with other products he might like.
While not unique, this approach is fairly narrow. Any system that does not use ALL of these methods is not covered by the patent. Consumer reviews are not patented, just this VERY specific approach to consumer reviews and recommendations.
Carry on now. Nothing to see here.