Hacker News Reader

Comment by Macha | Hacker News Reader

To add to the mess, the WHATWG now maintains a "URL" spec that covers URIs and declares URIs an outdated term for URLs: https://url.spec.whatwg.org/

So now we have a URL spec from WHATWG that says URI is an old term for URLs, and a URI spec from the IETF that says that URL is just a term for a certain subset of URIs.

View on HN

> the IETF that says that URL is just a term for a certain subset of URIs

This is the more broadly accepted definition, right?

That may be true within their bounded context. WhatWG is the web/HTML working group, and a URL is something you can type in the address bar, and this tells the browser how to find it, although the browser's configuration may be used implicitly in the process. URIs the browser can't use to locate something are useless in the context of the web. For example news:comp.lang.c may be considered a URN, but when the browser is configured with a news server, it can also locate it, so for web purposes it's a URL.

> URIs the browser can't use to locate something are useless in the context of the web.

The obvious counterexample to this is the mailto: URI scheme. It has no location but it is still a useful thing to link to on the web.

It locates and opens your email program. Sometimes it even locates and opens another web page (Gmail/O365).

It doesn’t locate my email program. It contains zero information about that.

> declares URIs an outdated term for URLs

The term I would use is "a distinction without a difference", but they're not entirely wrong.

> URI and IRI are just confusing.

WhatWG wasn't wrong.....

Reply on news.ycombinator.com