4 ms·
I really hate analogies in law because it always seems to provide very little support for whether a law is a good idea or not. Your analogy is to breaking and
by aero142 13y ago
I really hate analogies in law because it always seems to provide very little support for whether a law is a good idea or not. Your analogy is to breaking and entering a physical residence, but you skip over the fact that this is a publicly available website. I'd say it has more in common with photographing a public building. Lets say Dunkin' Donuts doesn't want me photographing their buildings and making a map of all their locations. They demand that I stop taking photo's from the street and put up a small tarp that covers the physical line of sight to their building that I used to photograph. I then come back and photograph their building from 2 feet to the right and they claim I bypassed their building security.
I don't think my analogy is any more correct than your, which is why I would rather we just decide what is best for society. Seems a lot more useful than trying to decide whether IP filtering is analogous to breaking a window.
- anigbrowl 13y agoDunkin' Donuts could exclude you from their premises if they saw fit. Your analogy is equivalent to 3taps making a list of Craigslist URLs and getting a front page screenshot for each one. Scraping is qualitatively different - you have go on into the website to analyze the content, so I suggest it's more like walking into the DD kitchen and trying to document their customer transactions.
- krichman 13y agoIt is nothing like walking into the DD kitchen and trying to document it. It's quite a bizarre fantasy to think that physical property rights are anything like client-server interactions.
- graycat 13y agoThe Web site sent a Web page to a user. What was sent was essentially just simple text although maybe with some JPG or PNG files. That's the way HTTP, HTML, and CSS work -- mostly just simple text. The Web site voluntarily sent this data, mostly just simple text. Now the user has the data that the Web site voluntarily sent. Commonly the user keeps the data; e.g., their Web browser commonly keeps a copy of this data to speed future accesses to it. Besides a Web browser is perfectly willing to write the data to files and a directory so that the Web browser can display the data again. Maybe the user will analyze the data, e.g., see what colors were used for the fonts. So what? Besides, there is no way for the Web site to tell what the user does with the data unless maybe the user republishes the data. Screen scraping is just using the data that the Web site freely sent. If the data is not republished, etc., then tough to claim that the user did anything wrong.
- mpyne 13y ago> The Web site voluntarily sent this data, mostly just simple text. If you know of a way to configure Apache to deliver web data to everyone but a certain subset of users without having to force the authorized subset to use authentication then the whole world is all ears. In fact, I'm sure HN could use this good news first so that they don't have to use such a non-specific ban system as IP bans... But until then 3taps had no question that they knew the web site operator did not want them to access their website at all, and in fact had to go out of their way to get around the IP ban, so let's not pretend like the court decision here is setting some kind of general precedent.
- graycat 13y agoNo, your argument is mixed up and confused. IP is irrelevant and not good evidence of anything. The IP address used by 3taps can be changed by the 3taps ISP at any time for any reason. IP can't be used for authentication. With public key cryptography and Kerberos, there are some excellent means of authentication. If Craigslist wants to use such authentication, fine, and then they can effectively and accurately block any given collection of users. But usually a Web site, e.g., HN, offers access to any IP address anonymously, without authentication. In that case, it's next absurd for the Web site to complain about some person when they have next to no good evidence on that person. Your "go out of their way" is wildly false; all that had to happen was just their electric company to drop power for one second. My electric company does this about once a week. Then the cable modem will forget its assigned IP address and, when electrical power is restored, request a new IP address. The 3taps people need not be aware of this at all. Moreover, the ISP can have assigned the banned IP address to someone else, Joe, not involved. Then Joe's usage of the Web site is no evidence against 3taps. Maybe 3taps knew that the Web site did not want them to use their site, but more importantly the site had no good evidence, at least not from IP address, if 3taps was using their site at all or not.
- mpyne 13y ago> Your "go out of their way" is wildly false; all that had to happen was just their electric company to drop power for one second. So you're saying that business networks typically have completely random IPs setup by their ISP? I would hope not, as that means SSL sites could not have worked at all for most people prior to Windows Vista. Not to mention the certain problem of how Google DNS is setup for people (Hint: It uses a static IP). In fact I think you might get even more disappointed if you consider the types of "proof" that are considered acceptable within the legal system, and commerce in general. For instance, completing a contract by faxing over a document, having it signed, and faxing it back. That has all the same theoretical issues associated with it as blocking static IP address and yet you don't see the entire edifice of the justice system or commerce falling to bits, now do you? > Moreover, the ISP can have assigned the banned IP address to someone else, Joe, not involved. Then Joe's usage of the Web site is no evidence against 3taps. Why are you speaking in terms of "maybes"? 3taps themselves admitted to using a proxy to evade the ban. They knew they were blocked, and they knew why. QED So while I would agree with you in general that an IP address is not a priori an identifier, that's not at issue in this specific case.