A purely hypothetical question

[Robert Krausankas (BuddhaPi)]

Again to much time on my hands, I should be registering my works so I can make some fast cash!! just kidding...

So I have a site and saw this really cool image of the Nebraska skyline, I truly love this image but am to cheap to purchase it, so instead I just link directly to the image from the photographers blog, cause the one he has there doesn't have the ugly watermark... So now the image "appears" on my site, but is not on my servers, I'm guessing I could still be on the hook??

I think i saw a similair thread somewhere, where someone had an ad banner being dynamically served to his site and the banner had a portion of one of getty's images on it..They didn't seem to care that the actual file was hosted elsewhere, and proceeded to go after him anyway..

You'd think if picscout was scanning the server for images in question, it would not "see" this image as it's only a link, and doesn't it have to read the code...last i knew bots/crawlers didn't have eyes in which to actually view an image online..

okay back to your regular scheduled programming unless you feel the need to discuss this which would be nice

[SoylentGreen]

No matter what, the likes of MF and Getty will say that the "end user" is on the hook.   You can bet your life on it.  But, there's no specific provision in copyright law for that, however.

My research shows me that PicScout spiders all links and pages, drilling down further and further.
A PicScout report is then viewed by a human, who 'looks' at a site visually, to determine at what page level the infringing images appear 'visually'.
I say 'visually', as a single page may be made up of many nested frames for example, but 'visually' the pages need to be noted as first page, secondary page, etc.
"Visually" is the key here.  MF and Getty aren't concerned with where the content comes from, only that it appears 'visually' on the site in question.
Then, "screen captures" are taken.  All this follows their "end user" pays policy; they're not concerned with how the content got there.  Only that they can see it there.
By now, perhaps some of this has been automated?

I think that if a site owner linked directly to an image elsewhere on the 'net, thereby "pulling" the image into the site, this would likely constitute a (wilful) infringement.
However, the opposite is clearly a different condition.  If a third party "pushes" infringing content onto your site (even if you set up the scripts to do so) this may not be considered an infringement being made by the site owner.  This would invoke protections for the site owner under the DMCA.

Again, MF and Getty will insist that no matter what, you've infringed and that the "end user" must pay.
But that's a construct that they've invented, and it's not copyright "law".
They'll threaten the end user, and may even go to court.  But, there's no provision in copyright law as per Internet technology as I spoke of here.
So, the argument would be a purely logical one to a judge.  That's in addition to any previous precedents.  Refer to the push/pull argument above.

Google's Image Search has always been an interesting case.  It's case for legality is probably beyond the scope of this forum, but it's protected by the DMCA.
I've heard that it has to do with the way that Google stores (or doesn't store) the thumbnails, and links dynamically into the actual images or pages.
It also provides a meaningful service.

It would be interesting to see cases wherein a defendant successfully draws an analogy between Google Image Search and their own site.

S.G.

[Robert Krausankas (BuddhaPi)]

Thanx for the detailed feedback SG!

It was the visual part that got me to thinking about it.. I would think whomevers is manning the picsout division, they they would bother to have someone actually visually inspect the domains in questions, but it's certainly plausable.

And I would have to agree that "pulling" an image would be willful infringement.. On the other hand if a company is "pushing" the image in question I would have to think and hope that any judge would toss this out without batting an eye. I hope the person that first mentioned this comes  back here with an update..

What are you thoughts on whether when picscout does it drilling if it can get into password protected directory's?? I would think not, but if they did there could be some case against them, and in like breaking and entering for lack of a better term. The reason why I bring this up, is when developing sites for my clients I almost always upload the projects on the backend of my domain behind a password protected directory, so the client can see the development as it progresses.. as I've said before I'm almost certain that some client supply me with images they grabbed off the net..any thoughts?

[SoylentGreen]

There's a sequence of events that happens when PicScout is used to find infringements.
From what I've seen, the web logs of alleged infringers indicate multiple visits from Israel corresponding to an IP address associated with PicScout.
This appears to be an automated scan, spidering all content in the site. 
If infringing content has been found by PicScout, multiple visits will occur from an IP address from the likes of Getty, MF or Corbis after a short period of time.
This appears to be a visual inspection by a human; someone visits using a web browser, following manual 'clicks' through the pages of the site.
Note that things may have changed recently, as Getty has bought PicScout.  I'm not even sure if PicScout's operated from Israel anymore.

I doubt that a situation involving an image that is "pushed" would end up in court.
While MF or Getty will certainly intimidate everyone they can in order to make them pay, I don't think that they'd waste their money actually going to court over an issue that comes down to the DMCA.
The DMCA is fairly mature and well-understood in my opinion.
But, they tell everyone that they'll sue, regardless of what they actually intend to do.

I don't think that PicScout could spider password-protected directories.
If it did, that would cause some serious legal issues for them, and would invalidate any results gained through such actions.
I have heard of people accidentally leaving content in a non-protected area, and then getting caught.  So it's important to keep things in their proper place.

In addition, I should mention that some stock-image houses traditionally allowed the use of low-res versions of their images for prototyping web sites, or print articles.
Some may remember the CD-ROMS that were sent around by MF. 
If this type of fair use was discouraged by over-zealous searching by PicScout and the corresponding threats by a given stock image company, then it would discourage the use of said stock photos by designers.
Maybe it's happening already?

One almost feels like having a stock image company's content anywhere on your network(s)/device(s) is almost like having illegal porn, and the likes of Getty/MF/etc. are the cops coming to get you.
It's a wonder if they actually sell much to anyone.

S.G.

[Robert Krausankas (BuddhaPi)]

I'm hoping picscout is still based out of israel as I've blocked every IP from there via htaccess, sine we know picscout doesn't adhere to robot.txt. Could you possibly post a screenshot of a portion of a log showing piscouts spider, I'd be curious as to if there are any other telltale signs that may be worth adding to my file.

[SoylentGreen]

The logs contain a lot of info that could cause some issues if I post it.
If I cut parts out, then it loses context and meaning.

But, suffice to say that the 'signs' are/were Israel, domain name "BEZEQINT-HOSTING" (among others) and IP addresses known to be associated with PicScout.

S.G.

[Oscar Michelen]

buddhapi:  Now that PicScout has been bought by Getty, the company's servers location may be changing from Israel.

[SoylentGreen]

Here's quite an interesting tidbit from US law (as explained on Wikipedia):

Definition of "copy"
 
Several important rights exist under the United States copyright law only for “copies” of works — material objects in which the work is embodied.[13] Section 106(1) prohibits the reproduction only of copies of works, and section 106(3) prohibits the distribution only of copies of works.[14] Thus, as the Ninth Circuit held in Perfect 10, Inc. v. Amazon.com, Inc., a link (even a deep link or inline link) to an image does not involve reproduction of a copy of the image by the person on whose web page the link appears.[15] An instruction to a browser to jump to an URL is not a reproduction or distribution of a copy of what is at the referenced URL.[16]

http://en.wikipedia.org/wiki/Copyright_law_of_the_United_States

I think that this means that actual copying and distribution of content are infringements.  But, content that is only linked to on the Web is not considered an actual copy, nor is it a form of distribution.

S.G.

[Robert Krausankas (BuddhaPi)]

hmmm interesting, if in fact it means how we're reading it, the poster whom had an served from anoutside source would be off the hook as it were... Which also means that a big portion of the blog-o-sphere would also be off the hook, as mort times they're using images from other blogs, and only linking to them or pointing the URL to them..worst thing going on there is a little theft of bandwidth, which in these days amounts to nothing anymore...

I could very easily link to someones image on flickr and that image could also be in Getty's library, GI sends me a letter, and I have some more scratch paper...

another good find SG (sans avatar)

[Ozzie1]

PicScout bot or any bot can be blocked regardless of the IP address. You have to have Administrator access on Windows or Root on Unix. In any case, you block the bot by user agent and a number of other filters. We do it all the time to control TQ or traffic quality on some of our servers. Might be able to post some generic code but you have to have systems administrators customize it for your particular environment. Problem is if your site is part of a content display network and you are an ad sharing affiliate this works if the bot comes to your site first because it can't "crawl" to the ad. However, if it goes to the ad origin website first it could potentially, depending on timing, end up on your site via the backdoor so to speak.

[Robert Krausankas (BuddhaPi)]

the problem here is that piscout is masking the user-agent, it just come across looking like a browser, hence blacklisting the IP range..if you know the user-agent, please do share.. I'm fairly confident that piscout won't be getting very far on my servers, between the IP block and the bot trap I have set-up

[Ozzie1]

I will take a look at it next week. We identify and block hundreds of bots and have never seen one that we couldn't identify as such. That doesn't mean it is not possible. There are just so many client-side factors that make it difficult to mask a bot because search engine click fraud software has become very sophisticated due to the class actions against the search engines about six years ago.