Ahh... Oscar. The conversation evolved.
Robert and I began by talking about methods an image search company could use to search for images that have been copied.
The two methods are being discussed. Here's are the methods in a hypothetical.
Suppose I have 10 images on my site. The company X is hired to by photographer Y to discover whether I am violating Y's copyright by hosting his image.
Method I: Company X can create a 'bot' (aka computer program or script), that comes to my site, tries to load every file on my server possibly copying each to its own server. This is generally called 'scraping'. When it finds an image, it compares that image to photographer Y's image. If there is a match, it reports the URI for that image to photographer Y.
Method II: Company X can create a bot that goes to
Google's image search page scours all those images, when it finds a match to photographer Y's image, the bot is programmed to click over to the offending site (in this case mine). After clicking over, it identifies the URI and reports that to photographer Y. This method will permit Company A to find images on my site
provided I have been permitting Google to crawl my images and save those to Google's server. (Note, I have changed verbs from "scrape" to "crawl". The two actions could be described as being 'exactly the same'; scraping could be defined as "unwelcome crawling". Sort of like "Stop your pawing!","Cara mia! That was a caress!")
Because Method I consumes lots of
my server resource (my $$), I prefer company X use Method II. That would mean they consume lots of
Google's server resources (Googles $$). As it happens Method II might also consume fewer of
company X's resources because they know that everything at Google images is an image and they don't have to crawl through lots of non-image material to find image links. So company X might want to do this. This method II is what Mulligan was suggesting when he kicked this off with
I wonder if this company's using Google's image search as the backbone? Is it possible to do that?Method II would be "using Google's image search as the backbone" of company X's system. But now lets turn to Mulligan's second question: Is it possible?
I'm pretty sure
Google wouldn't like company X to use Google's search in quite this way both because (a) it consumes Googles server resources and (b) crawlers don't click advertising links and so don't make Google any money.
So, Google would likely be motivated to take steps to make it difficult for companies of this nature from using Google search in quite this way. These steps might involve having its legal eagles write TOS that prohibit the behavior or it might involve using technology to notice the behavior and prevent or throttle it. The first would involve people like you writing a TOS, but it would be toothless if programmers and coders didn't do something to notice the behavior or collect evidence. And if they can notice or identify the behavior, they are likely to try to prevent it or throttle it. And Google is chockful of people who know how to code. So, I would bet they use code to inhibit Company A from using the image search as a backbone. That said: It's impossible to use technology
entirely prevent a determined party from scraping a public facing resource while still permitting public access, so its possible Company A does it nonetheless.
So, that was the first part of the conversation.
After that, Robert switched to: Who would block Google from crawling? And I told him some people might block Google from crawling
images and explained why they might do so. Robert likely understands my fuller point because he already knows that Google names their crawler and has different ones. Their image crawler is separate from their text crawler. So if I like, I can block Googles
image crawler but permit the
text crawler. One reason I might block the former and not the later is that I get very little 'good' traffic from image searches. I get lots of 'good' traffic from the text searches. (I'm the one who defines what's good from my point of view.)
So.. I think that's what the F we were talking about.
Topic: new image bot/spider/scraper (Read 10837 times)