# URLs with circumflex are not handled correctly

I tried linking to Wolfram Alpha today. The request contains ^ and it does not seem to work:

1. Vanilla C&P
http://www.wolframalpha.com/input/?i=limit+2^%28Sqrt%28n%29%29%2F1.0000000001^n
It is not even recognised as URL (it was in the comments, though).

2. Escaping ^
http://www.wolframalpha.com/input/?i=limit+2\^%28Sqrt%28n%29%29%2F1.0000000001\^n
Thought Markdown might interfere; no change.

3. Using HTML character code
http://www.wolframalpha.com/input/?i=limit+2&#94;%28Sqrt%28n%29%29%2F1.0000000001&#94;n
Works here, but did not in the comments.

4. Using ASCII code
http://www.wolframalpha.com/input/?i=limit+2%5E%28Sqrt%28n%29%29%2F1.0000000001%5En
Works both here and in the comments.

5. http://www.wolframalpha.com/input/?i=limit+2^%28Sqrt%28n%29%29%2F1.0000000001^n
<http://www.wolframalpha.com/input/?i=limit+2^%28Sqrt%28n%29%29%2F1.0000000001^n>
Nope. And ugly.

I think we should be able to copy&paste URLs and the engine should take care of the rest.

Why are comments even handled differently than posts?

## 1 Answer

I'm pretty sure this is : you're supposed to use the standard URL encoding %5e. Likewise, use %2a when an URL contains a *.

Note that ^ is not a valid character in URLs or URIs, per RFC 1738 (§2.2) and RFC 3986 (§2.2, §2.3).

• If this is by design, the design is broken. I should be able to use any URL my browser can handle.
– Raphael Mod
Apr 9 '12 at 23:53
• @Raphael No. Just because your browser (like some web servers) is more permissive than the standards doesn't mean SE need to accept invalid URLs. Apr 10 '12 at 0:00
• I see. Does not mean it has to not to either, though. The user experience for those who don't know the precise URL standards and what to replace special characters with is not good.
– Raphael Mod
Apr 10 '12 at 5:19