4 ms·
I disagree. I think the image and resource preloading done by Google is perfectly fine, but clicking actual links will mess up tons of systems (like, for exampl
by jeroenhd 5y ago
I disagree. I think the image and resource preloading done by Google is perfectly fine, but clicking actual links will mess up tons of systems (like, for example, links that are only supposed to be valid once, like in password reset emails).
There are good use cases for links with single-use tokens in them. Companies sad and desperate enough to suck as much data from you as humanly possible (i.e. every single news letter with tracking links) ruin these use cases for everyone, assuming they are indeed the reason Google is implementing this feature.
Given that many links (like, again, from password reset emails) give the person who clicks the link instant access to your account, I'd say this behaviour goes further than just filtering out trackers. Google has no business opening my Slack account or entering the change password page for the services I use. The stalking prevention they apply to external images and such is fine in my opinion, but links to external web pages should be left alone. You never know when Google accidentally clicks a link that says "confirm order" or "unsubscribe" because its magical AI misinterpreted the contents of an email.
- skybrian 5y agoFrom the very beginning of the web when HTTP was defined, there has been the rule that a GET request should never take an action on its own. Actions should be based on some other method like a POST request. Google is following the standard. People who take actions based on GET requests are not. Sure, mistakes happen out of ignorance, but they should be fixed.
- Silhouette 5y agoFrom the very beginning of the web when HTTP was defined, there has been the rule that a GET request should never take an action on its own. Actions should be based on some other method like a POST request. Maybe, but then several decades passed and now billions of people don't use online systems the same ways any more. It used to be that I could send a legitimate mail to a friend or family member and not worry that whatever mail system they use would refuse to deliver it to them because my system didn't jump through several not-quite-standard hoops that didn't exist when the email protocols were defined. Google seem fine with discarding the historical standards on which most of the Internet is built in that situation. In any case, if a communications service is going to snoop on your private communications and take actions that would be impossible without spying on you, I think the burden is 100% on them not to screw anything up for anyone, ever.
- meibo 5y ago"Things change as time goes on" is not an excuse to ignore valid standards. All big email providers know this and know that GETs should not affect their unsubscribe/confirmation links. It's part of their job.
- Silhouette 5y ago"Things change as time goes on" is not an excuse to ignore valid standards. You're talking about an idealised, theoretical world. I'm talking about the real one, the same one where big mail providers routinely ignore valid standards themselves in their efforts to fight real world problems like spam and identity theft. Again, if someone is going to help themselves to private information then the burden should be 100% on them not to screw anything up for anyone, whether or not that anyone was following any particular set of rules.
- skybrian 5y agoYes, people often don't follow standards, but they are sometimes useful in the real world for deciding who should fix their software, and for making breakage less likely due to whatever changes happen in the future. Google isn't the only vendor you need to worry about. Other mail agents, browsers, or proxies will sometimes prefetch URLs. On the web you often don't know what software your users are running or what it will do, and there's no way of knowing what they'll do in a year. Users have the right to run whatever mail software they want. But if you follow standards then you have a better chance. If you don't, it's a risk. Sometimes vendors do go the extra mile to tolerate other people's buggy code, but not always. You can decide to blame everyone else if you want, but if you use GET requests for user actions, something will likely break eventually.
- Silhouette 5y agoYes, I understand the reason for the standards, and other things being equal I am all for following them. I just think it's strange that an organisation spying on someone's communications and then doing anything it otherwise couldn't and breaking anything as a result should be getting such a free ride. There is no law that says any web application you write has to follow those standards, and the only reason not doing so is a problem in this case is because the likes of Google didn't mind their own business.