6 ms·

> So it appears this is happening with Stack Overflow knowing about it and approving it, after all — they implemented schema.org. Not really. As far as I know,

by kaoD 12y ago

> So it appears this is happening with Stack Overflow knowing about it and approving it, after all — they implemented schema.org. Not really. As far as I know, implementing schema.org means your data is structured, not that anyone can do whatever they want with it. The real reason Google can do it is because Stack Overflow user submitted content is licensed under CC BY-SA 3.0. Unfortunately Google is not complying with the license since it requires providing a link to it and indicating which changes were made[0]. There is a mechanism in schema.org to specify licenses[1]. I couldn't find it in SO answers' attributes, but Google shouldn't assume that means they can do whatever they want with the content! In fact, wouldn't that mean the content is under copyright (unless specified otherwise) and therefore not remixable/shareable? As far as I can tell, even if those are excerpts, Google's use does not fall under fair use. Anyways, props to SO for choosing CC BY-SA for their user submitted content. I think it's fair (after all SO feeds from their users) and, even if detrimental to their interests in the short term, in the long term builds trust among their users. [0] https://creativecommons.org/licenses/by-sa/3.0/ https://creativecommons.org/licenses/by-sa/3.0/ [1] https://schema.org/license https://schema.org/license

coldpie 12y ago
I agree with your first paragraph, but I don't agree that google is violating CC BY-SA. They do provide a link back to the original source, very clearly shown in the picture in the article. This seems sufficient for attribution, though perhaps they should go further and actually cite the answerer's username. You could argue that combining two answers counts as modification, but that seems more of a nit-pick than enforcing the spirit of the license. Otherwise the text is unchanged.
- kaoD 12y ago
  From the license: > Attribution — You must give appropriate credit, _provide a link to the license_ [...] There is no link to the license. There isn't even a mention of the license! For me this is the most important point, therefore Google is not complying with the license. --- > You could argue that combining two answers counts as modification, but that seems more of a nit-pick than enforcing the spirit of the license. This is a minor point for me, but I still disagree. Google's work is definitely a remix and the license is very explicit about it. We're not the ones meant to interpret what's the spirit of the license or when it's fine to violate its terms. I guess (IANAL) even non-remixed excerpts (i.e. removing content) should be marked as excerpts to fully comply. Here's example of why it might not be just a nitpick: answers are sometimes conflicting, and their meaning might change radically when automated excerpts are amalgamated. Imagine a right answer turned wrong by the automated merge. This could damage SO's reputation, driving away visits which would otherwise come back to SO as a reliable source. If I were to remix them by hand, I would definitely mark it as a derived work.
  - Navarr 12y ago
    If this was a concern than Search Engines would be worthless! They couldn't embed any snippets from any webpage ever, as all materials are by default copyright their author with no license for reproduction. I'm very sure there have been several legal cases involving search engines extracting information from the website. Spain's news organizations come to mind - the end result being that not licensing Google was bad for business.
bryanlarsen 12y ago
CC BY-SA is not the license that Google is using. Stack Overflow added metadata to their HTML to enable Google to use the answers in their response box. This is the primary (and probably sole) reason for this metadata, therefore it constitutes an implicit license grant to Google. IANAL.
- kaoD 12y ago
  > CC BY-SA is not the license that Google is using. Exactly, so their license is not compatible with SO's (which requires Share-Alike) and therefore they're violating SO's license until they explicitly comply with CC BY-SA in those excerpts. What matters here is which license the source content is distributed under. > Stack Overflow added metadata to their HTML to enable Google to use the answers in their response box. This is the primary (and probably sole) reason for this metadata SO added metadata to enable anyone to use their data in any way (as long as they comply with the license). I'm very grateful for having structured data. Even if schema.org is a Google (+ Microsoft + Yahoo + Yandex) initiative, the idea is to structure the content, not to give it for free. > therefore it constitutes an implicit license grant to Google. I don't think it works that way. I'm not sure if "implicit licensing" will hold in court (after all, there is a explicit license contradicting it), but even if it would, the only thing schema.org implies is that the data is structured. Here are schema.org terms and conditions: https://schema.org/docs/terms.html https://schema.org/docs/terms.html It never says you give them your content for free. > IANAL. IANAL either :)
  - annnnd 12y ago
    Not sure why you got downvoted - here, have +1. I agree that they probably don't comply to the license by not linking to the license itself, however, I doubt anyone will take them to court for this. Also, this could be seen as quoting and is in some jurisdictions allowed (with due restrictions). So maybe they don't care about license... Aren't they doing the same for ages, when they show relevant portion of the page with the keywords?
    - kaoD 12y ago

by kaoD 12y ago

coldpie 12y ago

I agree with your first paragraph, but I don't agree that google is violating CC BY-SA. They do provide a link back to the original source, very clearly shown in the picture in the article. This seems sufficient for attribution, though perhaps they should go further and actually cite the answerer's username. You could argue that combining two answers counts as modification, but that seems more of a nit-pick than enforcing the spirit of the license. Otherwise the text is unchanged.

kaoD 12y ago
From the license: > Attribution — You must give appropriate credit, _provide a link to the license_ [...] There is no link to the license. There isn't even a mention of the license! For me this is the most important point, therefore Google is not complying with the license. --- > You could argue that combining two answers counts as modification, but that seems more of a nit-pick than enforcing the spirit of the license. This is a minor point for me, but I still disagree. Google's work is definitely a remix and the license is very explicit about it. We're not the ones meant to interpret what's the spirit of the license or when it's fine to violate its terms. I guess (IANAL) even non-remixed excerpts (i.e. removing content) should be marked as excerpts to fully comply. Here's example of why it might not be just a nitpick: answers are sometimes conflicting, and their meaning might change radically when automated excerpts are amalgamated. Imagine a right answer turned wrong by the automated merge. This could damage SO's reputation, driving away visits which would otherwise come back to SO as a reliable source. If I were to remix them by hand, I would definitely mark it as a derived work.
- Navarr 12y ago
  If this was a concern than Search Engines would be worthless! They couldn't embed any snippets from any webpage ever, as all materials are by default copyright their author with no license for reproduction. I'm very sure there have been several legal cases involving search engines extracting information from the website. Spain's news organizations come to mind - the end result being that not licensing Google was bad for business.

bryanlarsen 12y ago

CC BY-SA is not the license that Google is using. Stack Overflow added metadata to their HTML to enable Google to use the answers in their response box. This is the primary (and probably sole) reason for this metadata, therefore it constitutes an implicit license grant to Google. IANAL.

kaoD 12y ago
> CC BY-SA is not the license that Google is using. Exactly, so their license is not compatible with SO's (which requires Share-Alike) and therefore they're violating SO's license until they explicitly comply with CC BY-SA in those excerpts. What matters here is which license the source content is distributed under. > Stack Overflow added metadata to their HTML to enable Google to use the answers in their response box. This is the primary (and probably sole) reason for this metadata SO added metadata to enable anyone to use their data in any way (as long as they comply with the license). I'm very grateful for having structured data. Even if schema.org is a Google (+ Microsoft + Yahoo + Yandex) initiative, the idea is to structure the content, not to give it for free. > therefore it constitutes an implicit license grant to Google. I don't think it works that way. I'm not sure if "implicit licensing" will hold in court (after all, there is a explicit license contradicting it), but even if it would, the only thing schema.org implies is that the data is structured. Here are schema.org terms and conditions: https://schema.org/docs/terms.html https://schema.org/docs/terms.html It never says you give them your content for free. > IANAL. IANAL either :)
- annnnd 12y ago
  Not sure why you got downvoted - here, have +1. I agree that they probably don't comply to the license by not linking to the license itself, however, I doubt anyone will take them to court for this. Also, this could be seen as quoting and is in some jurisdictions allowed (with due restrictions). So maybe they don't care about license... Aren't they doing the same for ages, when they show relevant portion of the page with the keywords?
  - kaoD 12y ago