4 ms·

Real use case for this is so Google can poke into context for links shared via services they have no ownership. They got used to know exactly what people share

by gcb0 8y ago

Real use case for this is so Google can poke into context for links shared via services they have no ownership. They got used to know exactly what people share and, much more importantly, what people click on email, thanks to gmail rewriting all links to a google track url (while misleading the user by a fake status text on mouse over). When they sum the info of 1. who send it, 2. who clicked on it, 3. what was the email context, they can further classify the page content. (for ads, search ranking, whatever they do today) But they can't do that on whatsapp, facebook, etc. With this "feature", now they have at least a modicum of context when the page impression show up on their Google Analytics trackers. Did you visit that wikipedia page because of recent news events, or doing homework? Or, we could just go with occam's razor and believe that a bunch of bored engineers at google decided to simply add a feature completely disconnected from any of their business plans, and spent a few months tweaking a git repo with nothing but a README file... while the actual code review has follow up questions that were left open and merged anyway: https://chromium-review.googlesource.com/c/chromium/src/+/1434215/17/third_party/blink/renderer/core/page/scrolling/text_fragment_finder.cc#57 https://chromium-review.googlesource.com/c/chromium/src/+/14... ¯\_(ツ)_/¯

nerdponx 8y ago
It's a shame because it is actually a good idea, but now Google has first-mover advantage, so any other implementations will just be a clone of what Google already did.
- dmurray 8y ago
  Isn't that...better than every browser having an incompatible implementation of it, so Firefox users could only share links with Firefox users, etc? I get the argument that it's better for these things to be done by consensus among the browser makers. Is it in scope for WHATWG?
- Izkata 8y ago
  http://archive.is/ http://archive.is/ got there first, sort of. If you archive a page then highlight some text, the URL changes to a direct shareable link to that text. Bonus: being its own site, and implemented in javascript, it's already cross-browser.
dvfjsdhgfv 8y ago
> what people click on email, thanks to gmail rewriting all links to a google track url (while misleading the user by a fake status text on mouse over) They use the same trick for Google Search in a way that is transparent to the user - as long as you have JS turned on. When you switch it off, you realize instead of clicking on HNs URL, you actually click https://www.google.com/url?q=https://news.ycombinator.com/&sa=U&ved=0ahUKEwjMxpegp73gAhXvlIsKHcxcDV0QFggZMAA&usg=AOvVaw23bpdDl4nmvkO_L9fANoX7 https://www.google.com/url?q=https://news.ycombinator.com/&s... Now, they want to move a step further. I wonder when they're going to stop.
- k_bx 8y ago
  This technique (redirecting through some dummy ?url=... page) is a common way of doing an external redirect on any service you're building. It's a simple way of making it impossible to see the actual Referer to external website, which is more secure to your data and the service you're building (e.g. Gmail).
  - z3t4 8y ago
    But browsers seem to have stopped giving out referrer some years back.
  - ajnin 8y ago
    There's a standard way to do that, which is to add a rel="noreferrer" tag to your outgoing links.
    - k_bx 8y ago
      Code usually doesn't carry final HTML in its data structures, whereas converting a URL in a single place where it's extracted is easy. In other words, making sure you didn't miss to put rel=noreferrer everywhere is way harder.
    - detaro 8y ago
      And it still makes somewhat sense to not entirely rely on it, given people still use browsers that do not support it.
dmurray 8y ago
> Or, we could just go with occam's razor and believe that a bunch of bored engineers at google decided to simply add a feature completely disconnected from any of their business plans I understand the skepticism about Google's motivations but "making Chrome better and more feature rich so people will want to use it" is completely in line with Google's business plans.
- gcb0 8y ago
  I believe you. That's what the fine folks at the IE team did with ajax too, for example. And I am not being ironic, but the line is so thin that I have to point it out either way.

by gcb0 8y ago

nerdponx 8y ago

It's a shame because it is actually a good idea, but now Google has first-mover advantage, so any other implementations will just be a clone of what Google already did.

dmurray 8y ago
Isn't that...better than every browser having an incompatible implementation of it, so Firefox users could only share links with Firefox users, etc? I get the argument that it's better for these things to be done by consensus among the browser makers. Is it in scope for WHATWG?
Izkata 8y ago
http://archive.is/ http://archive.is/ got there first, sort of. If you archive a page then highlight some text, the URL changes to a direct shareable link to that text. Bonus: being its own site, and implemented in javascript, it's already cross-browser.

dvfjsdhgfv 8y ago

> what people click on email, thanks to gmail rewriting all links to a google track url (while misleading the user by a fake status text on mouse over) They use the same trick for Google Search in a way that is transparent to the user - as long as you have JS turned on. When you switch it off, you realize instead of clicking on HNs URL, you actually click https://www.google.com/url?q=https://news.ycombinator.com/&sa=U&ved=0ahUKEwjMxpegp73gAhXvlIsKHcxcDV0QFggZMAA&usg=AOvVaw23bpdDl4nmvkO_L9fANoX7 https://www.google.com/url?q=https://news.ycombinator.com/&s... Now, they want to move a step further. I wonder when they're going to stop.

k_bx 8y ago
This technique (redirecting through some dummy ?url=... page) is a common way of doing an external redirect on any service you're building. It's a simple way of making it impossible to see the actual Referer to external website, which is more secure to your data and the service you're building (e.g. Gmail).
- z3t4 8y ago
  But browsers seem to have stopped giving out referrer some years back.
- ajnin 8y ago
  There's a standard way to do that, which is to add a rel="noreferrer" tag to your outgoing links.
  - k_bx 8y ago
    Code usually doesn't carry final HTML in its data structures, whereas converting a URL in a single place where it's extracted is easy. In other words, making sure you didn't miss to put rel=noreferrer everywhere is way harder.
  - detaro 8y ago
    And it still makes somewhat sense to not entirely rely on it, given people still use browsers that do not support it.

dmurray 8y ago

> Or, we could just go with occam's razor and believe that a bunch of bored engineers at google decided to simply add a feature completely disconnected from any of their business plans I understand the skepticism about Google's motivations but "making Chrome better and more feature rich so people will want to use it" is completely in line with Google's business plans.

gcb0 8y ago
I believe you. That's what the fine folks at the IE team did with ajax too, for example. And I am not being ironic, but the line is so thin that I have to point it out either way.