Hacker News Reader

Comment by sp332 | Hacker News Reader

As a user, that would really annoy me. I like to know where I'm going before I put a URL in my browser.

Any URL can send you anywhere. If you trust the origin of the QR-code, there's no reason to not trust going to /qr.

For the record, I agree with reviewing the URL before hitting it, but if the domain matches what I expect, I'd be strongly inclined to trust the rest.

I don't think the majority of users check the URL of the QR code. Why wouldn't you trust the advertiser? I think it's really rare to see an advert that looks innocent, but actually isn't, and at the same time contain a QR code that has a "malicious-looking" URL.

I guess I'd trust ads more than random QR codes on blogs. My point is, if the point of a redirect is that you can send people to a different page later in time, then you really can't depend on the QR code taking you where you expect, or even to the same page that it used to go to.

If this is such a concern, you can always have your browser ask you for permission before it follows a redirect. Any URL anywhere could potentially redirect.

I guess the point is that there is a certain amount of comfort to be had if the url is http://www.domainname.com/ because at least then you know that any redirects are intended for the entire world: no one is shuffling you though ad trackers making you feel like a pawn.

And if you care that much about the URL you should be able to infer from website.tld/qr that it's a QR code specific URL to track hits.

But if it's a redirect, it could send me anywhere. That's what I don't like about it.

You just scanned an unreadable symbol. It could already send you anywhere. Vetting URLs by whether they look like they point at a redirect script seems like overkill.

Oh, actually I've never used a QR code reader that didn't show me the URL before letting me decide whether to go there.

Even that's not an absolute assurance, as millions of people who have failed to renew expired domains can attest. And the HTTP status codes have a whole block of redirect codes while just about every web server permits backend proxies. When you visit http://viral.example/ it can already send you anywhere; seeing the URL http://viral.example/qr only tells you that you're getting QR-specific content, and still does not tell you what that content is.

I guess the only problem is people can't recognize familiar URLS if you use a redirect page. If you link to a Facebook event for a concert, for example, people could see the trusted domain before visiting.

Consider what can be done with a punycode'd URL here, too.

Mix a QR with a little punycode and you easily can end up anywhere, even if you do (quickly) review the URL before the jump.

Would you really be worried if AirBNB had a QR code which took you to http://airbnb.com/qr because it might be a redirect which could take you anywhere?

Currently, no. But I would be mad if they sold a redirect to someone else and it took me where I didn't expect. And if that happened often, then I would get more suspicious of redirects over time.

How about the more likely version: http://qrserv.clearcast.com/?adid=239af932e which redirects to http://airbnb.com/?

I don't much use QR codes, but when I want to go to a URL I got in email, and it looks like that, even if the email otherwise seems legit, I just type in the URL to the real site directly. Clicking on such a link feels like deliberately falling for a phish.

It's very common to use shortened URLs to decrease the complexity and size of the resulting code. I don't think most users scanning codes are any more concerned than they are clicking on a shortened twitter link.

Reply on news.ycombinator.com