Log in

No account? Create an account
Lindsey Kuper [entries|archive|friends|userinfo]
Lindsey Kuper

[ website | composition.al ]
[ userinfo | livejournal userinfo ]
[ archive | journal archive ]

Hackable URLs: speed, usefulness, fun, respect [May. 20th, 2008|10:58 pm]
Lindsey Kuper

Audrey on Dave on Geohash:

It turns your lat/long into a single alphanumeric string. The cool thing here is that as you lose characters from the right side, the sequence remains valid at a lower accuracy, describing a larger and larger bounding box. I could see this being really useful for a site that had location urls matching some collection of data. it’s human-editable enough that people could expand the search area just by editing the url.

Aside from that just being cool, I have a soft spot for services that not only tolerate, but actually encourage URL hacking! del.icio.us is probably the canonical example -- they put it right in front of you -- but Adam Darowski points out1 some other folks that have nicely sliceable URLs, too. And docreader lets you do it along multiple axes at once, causing people's heads to explode.

Hackable URLs appeal to me because they have the potential of making it faster and easier to find stuff and link to stuff. Hackable URLs appeal to me because it's fun to tinker. But hackable URLs appeal to me most of all because they imply trust and respect for users (as opposed to URLs that imply "don't tinker or we'll sue you").

It's also interesting to think about how URL was never supposed to be an exposed part of browser UI, considering how well entrenched it is now.2

  1. Darowski didn't coin the "URL as UI" phrase he uses. Jakob Nielsen has been writing about it since roughly the beginning of recorded history.
  2. Not necessarily everywhere, though. In Japan, we saw a lot of ads that showed search terms (or phone-scannable 2D bar codes which encoded URLs!) much more prominently than they showed domain names. Our patient teachers told us that it was because a typical domain name is hard to chunk if you're not a native reader of the English alphabet. And, indeed, when we visited Roy and Electra's Japanese teacher, I noticed that she didn't have the URL field showing by default in her browser. (From what I can tell, the idea of domain names that contain non-ASCII characters is still largely just an idea.)

[User Picture]From: poodleface
2008-05-21 09:03 am (UTC)
Not really the same but I've been having a field day typing in random strings of text by just haphazardly mashing a few keys to search for hastily uploaded videos on YouTube:


(Reply) (Thread)
[User Picture]From: lindseykuper
2008-05-21 06:26 pm (UTC)
Where, exactly, was the skater planning to go if they hadn't crashed into the wall?

Also, it seems weird that people have time to upload videos but they don't have a few seconds to give 'em meaningful titles.
(Reply) (Parent) (Thread)
[User Picture]From: jes5199
2008-05-21 03:33 pm (UTC)
firefox is totally willing to accept Unicode URLs. I've never seen anything useful at the destination, but for example, there exists http://☻.com
(Reply) (Thread)
[User Picture]From: lindseykuper
2008-05-21 06:23 pm (UTC)
That's true; there's browser support. But there's browser support for a lot of stuff that almost no one uses.

Also, the top-level domain extensions are ASCII-only, so you still have to know how to read those, at least to the extent of being able to remember them and tell them apart.

For better or worse, I guess it's always been, "You want to play? Learn to read English, the more the better."
(Reply) (Parent) (Thread)
[User Picture]From: jes5199
2008-05-21 06:32 pm (UTC)
there's an argument that goes something like:
computers are in english
music is in italian
poetry is in french
church is in latin
and this is normal.

fair? ridiculous?
but I guess nobody had to learn a new language to talk on the telephone.

well, there's always volapuk.
but it's probably harder to do that with CJK characters
(Reply) (Parent) (Thread)
[User Picture]From: phthoggos
2008-05-21 11:10 pm (UTC)
For me this issue comes up especially with Firefox's quick-search keyword-based bookmarks.

"w foo" gives http://en.wikipedia.org/wiki/Foo
"l gloria" gives http://www.perseus.tufts.edu/hopper/morph.jsp?l=harmonia&la=la
"g a)reth/" gives http://www.perseus.tufts.edu/hopper/morph.jsp?l=a%29reth/&la=greek (unfortunately the Perseus project still doesn't accept Unicode input; you have to type in beta code)
"w umeå" gives "http://en.wikipedia.org/wiki/Ume%C3%A5"
"w sv:umeå" gives "http://en.wikipedia.org/wiki/sv%3AUme%C3%A5" which gives "http://sv.wikipedia.org/wiki/Ume%C3%A5"

Using Safari, I think I can also get the following to work (though keyword address bar searching requires a plugin):
"w ἀρετή" gives http://en.wikipedia.org/wiki/Ἀρετή
"w zh:社会主义荣辱观" gives http://en.wikipedia.org/wiki/zh:社会主义荣辱观" which gives "http://zh.wikipedia.org/wiki/社会主义荣辱观"

Whereas they give me problems in Firefox -- FF's address bar appears to recognize ISO-8859-1 but not anything beyond: "w 社会主义荣辱观" yields "http://en.wikipedia.org/wiki/%3F%3F%3F%3F%3F%3F%3F" (which is "http://en.wikipedia.org/wiki/???????")
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-05-22 12:43 am (UTC)
I wonder if we can rig up a YubNub command that will solve that problem.
(Reply) (Parent) (Thread)
[User Picture]From: idealisms
2008-05-26 05:40 pm (UTC)
Firefox's address bar handles different encodings for URLs just fine. It may be the case that the default keyword for w is set up wrong or maybe it was a keyword you added? E.g., if you go to wikipedia, right click on the search field and select "add a keyword", it will properly handle "w 社会主义荣辱观". When the keyword is created, it grabs the encoding out of the form or from the document charset and saves it with the bookmark.

Re: IDN- it's not that common and there are only a few domain name registrars with permission to assign them. Due to phishing concerns, Firefox only allows IDN if they come from certain registrars (a white list). Part of the reason it's not that common is that typing with more complex character sets like Chinese is more difficult. This is part of the reason why many Chinese sites use just numbers for their domain name (e.g., 163.com or 263.net).
(Reply) (Parent) (Thread)
[User Picture]From: lindseykuper
2008-05-26 09:39 pm (UTC)
Yeah. In Japan, we saw all kinds of interesting things being done to ameliorate the "typing is a pain" problem. It makes me wish I had more time there to learn about it.
(Reply) (Parent) (Thread)