Google Wants to Kill the URL | Hacker News Reader

Google Wants to Kill the URL

wired.com

52 pointsLopRabbit24 comments

> But over time, URLs have gotten more and more difficult to read and understand. As web functionality has expanded, URLs have increasingly become unintelligible strings of gibberish combining components from third-parties or being masked by link shorteners and redirect schemes.

There's no reason a well-behaved web sire (or application) needs to have a complicated URLs, that's just bad design. No need to scrap the entire concept of URLs because some people use it improperly.

Yeah, I think Google could accomplish a similar result by publishing a 'nice URL' standard, where your page can be dinged on SEO if it does not conform.

Getting rid of the thousands of ugly URLs used by their own products could also go a long way...

I'd be curious to see what people would consider that to be.

Using HN as an example, what would the nice URL look like for this page?

There's not much room for improvement there. However, the page I am writing this reply on could - it's: <https://news.ycombinator.com/reply?id=17913014&goto=item%3Fi..., which is getting ugly.

The prime candidates for URL deuglification would be sites like amazon.com: https://www.amazon.com/Bushnell-Falcon-133410-Binoculars-Bla...

I mean, what the hell are all those identifiers, and why should anyone besides amazon.com's internal staff even see them?

actually your amazon link is a reflink, the whole amazon link is just:

https://www.amazon.com/Bushnell-Falcon-133410-Binoculars-Bla...

which is basically the product name + model number + asin

ASIN: https://de.wikipedia.org/wiki/Amazon_Standard_Identification...

Link can also be shortened to just the asin. (ASIN = ISBN applied to everything)

Thanks for the info, but the link I pasted was taken by just going to www.amazon.com front page, and copying the link for the "Deal of the day" product I saw there. So it's the link all customers see when they click on the products, and no amount of explaining what it is makes it go away, or justifies it being there in the first place.

Still, Amazon is far from the worst offender. I have seen kilometer long URLs that can span several lines of text.

I would like meaningful subdomain names, avoid putting data (particularly if it needs URL encoding) in query params, try to make urls somewhat parsable by non-technical humans.

Also tracking codes in URLs are kind of crummy because they impose a cost to the user that typically only benefits the publisher. Could be cool to write a tool that automatically strips tracking codes when you copy a URL to clipboard that contains one.

Or randomizes tracking codes!

I hope this will make URLs more transparent and _easier_ to "hack", not the opposite. URLs and deep linking is what makes the web great. If this means making that feature of the web accessible to everyone, that's awesome. If it means trying to get rid of it, that's bad.

For example, allow websites to customize how the URL is displayed. Every website nowadays has breadcrumbs. Let them put them in the URL instead. Let them provide icons and dropdowns for segments, but still allow users to change, copy and paste the raw URL (like the Windows file explorer).

At this point, I am beginning to believe this might be a non-problem for the vast majority being framed as a problem for google to milk it somehow. We haven't connected the dots yet.

So, what, then?

All POST, all the time?

Seen versions of that. No, thanks.

Just hide the address bar?

We already quasi have this, where people just type what they want into a Google search.

And if Google doesn't list it prominently, they are at a loss and/or use some other site, instead.

People used to be totally overwhelmed by security. You know what? When it started to impact them, personally, they learned. I'm continually surprised by the conversations I now hear around me, by "normal people".

The URL isn't "too difficult". It's that people don't care about it. And "tech" keeps trying to hide it further.

Reading the article, I get the impression all they are thinking of is tweaking the way URLs are displayed (“We want to challenge how URLs should be displayed and question it as we’re figuring out the right way to convey identity.")

I think a simple improvement would be to bolden the domain name in a URL.

Browsers could also add some styling to hyperlinks indicating whether they are to the same domain, to another domain that user trusts (for example because a login cookie has been set on its domain), but I expected diminishing returns on that. Even if it can be done reliably from a technical point of view, users won’t learn to discriminate multiple kinds of links.

Also any modifications to the space that belongs to the page is just begging for someone to build an in-page facsimile, especially when that modification is meant to engender trust.

> I think a simple improvement would be to bolden the domain name in a URL.

They already did that. If you open Chrome right now, you'll see that the non-domain part is greyed out. It doesn't help much.

I think URLs will die; it's just a matter of when

Most people don't go to addresses; they go to sites. In other words, people don't care about putting that "https" before they visit Facebook; they want Facebook and know that typing in "facebook.com" gets them there

Additionally, most of today's browsing isn't even done by web browser; it's done through mobile apps. That's why Apple wants to make the iPad the primary computing device; they already have a market-leading position on apps, so once iPads become the way that people compute, the efficacy of web browsers greatly diminishes

Additionally still, Chrome sets itself up to use Google as its primary search engine. Thus, if I type "facebook" into the search bar, the first result is almost always what I want (facebook.com). As well, Google (and other search engines) already do crazy-accurate location-based and context-aware searching so that they know that typing "ford" in Akron, OH likely means "ford.com" and typing "ford" in Toyko likely means "ford.co.jp."

So it makes sense for Google to think "What if people just typed the name of the thing they wanted without having to worry about whether it needs "https" or if it's a .com or a .org? What if objects on the internet can be located via some other means?"

I disagree. URLs still save state in a standard, copyable way. I find that to be very important and experience situations in which I miss them quite regularly.

Reminds me the effort Douglas Crockford was making (still making?) years ago when he was rethinking web infrastructure [1]

[1] https://www.youtube.com/watch?v=7TwrABEAfTk - 7m30s

I wonder what is Google's real reason for this. I have a suspicion it may have something to do with Chrome adding tracking variables to all the URLs you visit, or something along those lines. And they don't want most people to know about it.

Tracking and AMP abstraction were the first to come to my mind. Without a url you could AMP the desktop experience and just manage the entire experience.

Google gave a demo where the browser address bar will no longer show the URL of the web server (e.g. Google AMP), it will show the authenticated origin of the Web Package, https://news.ycombinator.com/item?id=16456025

Shift user behavior from URLs to branded search queries where either the brand pays exorbitant CPMs (low CPC but high CTR, leading to higher earnings per mille than even insurance or law related queries) to protect their branded traffic or risk losing traffic to competitors trademark bidding.

After typing something into the firefox search bar, I was greeted by a google captcha that would not let me pass, since I was using a VPN. I then had to punch in duckduckgo.com into the bar to get what I wanted. It was unfamiliar and awkward to not be able to access google to find what I needed. Imagine the url didn't exist, the user would have no way to connect to the internet without going through google. I think Google wants to receive more traffic, collect more data, all while controlling users access to the internet.

I think making web apps similar to the mobile experience would be awesome.. then we can have less startups with wacky names because they can't get a unique domain.

Reply on news.ycombinator.com