Hacker News Reader

Comment by efaref | Hacker News Reader

For something that claims to "grok" URLs, using "://" as part of the logo seems to be misunderstanding how URLs (are supposed to) work.

According to RFC3986, URIs (the proper name; URLs are a kind of URI) are correctly of the form:

    <scheme>:<scheme-specific-part>

For some schemes, the <scheme-specific-part> may contain a <hierarchical-part>, the format for which is:

    //<authority>[/<path>] | <path>

So it's only hierarchical URIs with a defined authority that have the (admittedly iconic) "://" in them.

I know most people identity URIs as having that sequence in them, but these are all valid URIs:

    mailto:bob@example.com
    sip:bob@example.com
    tel:+18005550123
    file:/path/to/file

I'm not saying I expect everyone to understand these subtleties, but people who claim to grok URLs ought to.

View on HN

In the article:

"I also suggested we’d try to work with :// (colon slash slash) somehow as a symbol, since after all the :// letter sequence is commonly used in all the URL formats that curl supports. And it is sort of a global symbol for URLs when you start to think about it. Made sense to me."

They seem to grok it perfectly well, but recognize the modern association people have with :// and urls.

This might shed some light - Curl Developer on URLs

https://webcache.googleusercontent.com/search?q=cache:hGPFWP...

It's a logo dude...

But I liked his comment, and explanation! I also learned some. What are we supposed to say, "nice!"? :-)

Does cURL support mailto/sip/tel?

No, it supports mainly schemes with a hierarchical-part - even though it works with 'file:'

For a logo there's absolutely no problem in using an element that's common and pretty identifiable

While on the subject, what is the difference between file:/path/to/file and file:///path/to/file ? Usually I only see the latter being used.

In file:///, the file:// part signifies the protocol (think http://) and the / in /path/to/file signifies the root directory (think / in /home/user/)

I have never seen file:/ used before, so I can't really comment, but it seems invalid to me (and browsers auto-redirect it to ///).

Almost.

The file: signifies the protocol. The /// is actually //<machinename>/ with an empty machinename as a shorthand for 'localhost'.

Interesting.

Why is this interpreted as a relative path?

     cd /etc
     curl file://resolv.conf

Well, it's ill-formed, so what the UA does with it isn't really specified; cURL presumably treats it as a path relative to . because that's what the developer(s) decided that cURL should do. (For comparison, Firefox transforms it into file:///resolv.conf, then tries to find /resolv.conf on localhost and fails.)

Not to mention that Tim Berners-Lee himself says that the // was superfluous in retrospect. Regardless, I don't think that paying homage to a widely-known idiosyncrasy of our technological evolution is the worst way to go for a project like curl (a historic project in its own right).

Every time I see the word grok I grok in my mouth a little.

Why? What's wrong with it? It's a real word (i.e. it's in Merriam-Webster).

Reply on news.ycombinator.com