Graphemica - for lovers of characters and symbols | Hacker News Reader

Graphemica - for lovers of characters and symbols

graphemica.com

66 pointsbleakgadfly14 comments

Thanks for the HN love. This is actually one of my side projects. I'm a character and type geek and use this as a place to experiment and play with the data. Really glad to see other people find it interesting.

Let me know about anything that would make it more useful. I've kept it pretty minimal so it's easy to change things.

There are lots of little features that aren't immediately obvious, like user profiles: http://graphemica.com/people/alphabetum Also, cool stuff on the cjk characters, including definitions: http://graphemica.com/戰

Nice work. I had the idea to do something like this a few months back and never followed through on it. Yours is 100 times better than mine would ever have been, so now I'm glad I never got to it.

I would love to see something like detexify: http://detexify.kirelabs.org/classify.html but that worked for unicode chars as well.

I think the guy who made it has some posts about how it works on his blog at http://detexify.posterous.com/ as well.

Try Shapecatcher (http://shapecatcher.com/). Could be improved, but hey, it works.

> cool stuff on the cjk characters, including definitions

The Firefox add-on "Character Identifier" does some of this too.

https://addons.mozilla.org/en-US/firefox/addon/character-ide...

This is simply excellent—beautiful and clean and just works. But if you really want points for improvement, I’ll do my best. :P

• For your CJK example, it shows “ZHAN4” for the Pinyin, which would be better rendered as “zhàn”, using tone marks instead of tone numbers.

• When a character name is too long, you omit the “Unicode code point: U+XXXX” altogether (e.g., for U+2918); it would perhaps be better to shorten it first to just “U+XXXX”, or make the layout fluid to allow that column to remain visible.

• It would also be cool to search characters’ descriptive text, such as that shown (and optionally searchable) in the Ubuntu character map. But that might be getting a bit on the extreme side.

• The /search?q=x URL could be prettier—/search/x for instance. Just my obscure personal preference though.

I would like if the results were grouped by whether the word was connected to the meaning or the appearance of the character. For example "star" yields "★" and "星", I was probably looking for "★" and if I wanted "星" I'd reach for a dictionary instead.

Nice. OS X has this built-in though. Other things I've noticed:

- The "no results found" bar overlaps the search box.

- Can't search by HTML entity

- Seems to be missing a few of the more exotic ones, e.g.:

💩 PILE OF POO Unicode: U+1F4A9 (U+D83D U+DCA9), UTF-8: F0 9F 92 A9

Feedback noted. The reason some characters are missing is because I haven't had time to update it to include the new Unicode 6 characters. I'll do it ASAP.

This is rad: http://graphemica.com/search?q=apl

This is a bit off topic, but I copy-pasted some of the weird characters from the site to name.com, and I found that probably all .com combinations are taken.

I put cloud icon dot com, it's taken. I put 10 consecutive cloud icons dot com, also taken. I put many random characters dot com, also taken. What does it mean?

a bug in the code checking for them?

Nice site. I would use this more for Windows ALT key codes—I have a few like the em dash or ALT+0151 memorized—but maybe this isn't something you want to add.

could add some semantic markup or rdfa or something, could be cool

Reply on news.ycombinator.com