Hacker News new | past | comments | ask | show | jobs | submit login
Unicode Snowman for You (xn--n3h.net)
115 points by jamesjyu on Dec 23, 2010 | hide | past | favorite | 48 comments



http://railssnowman.info , for those of you who haven't seen it.

I believe it ended up being a ✓ in the final build, though.


Looks like it was originally committed as a snowman, changed, then committed again, then finally removed. Quite the flurry of activity over a tiny character that sits in the background quietly increasing the consistency of browser behavior and saving the world from cross-encoded text.


It was quite a good case-study in bike-shedding.

For those new to the term, it's effectively "The intensity of discussion is inversely proportional to the complexity of the problem". http://ftp.sunet.se/jargon/html/B/bikeshedding.html http://en.wikipedia.org/wiki/Parkinsons_Law_of_Triviality


I don't know if I'd entirely call it bike-shedding; if they left it as a snowman, people would start to make comments like "Oh those unprofessional Rubyists, putting silly snowmans willy-nilly into code!"

Then again, they might not. But a check seems more 'professional.'


I agree with you on the end result, but I remember following the comments and discussion on GitHub when the changes were being made. It sure seemed like a big discussion over not much across several commits, and more thank 50 comments on GitHub alone.


Yeah, I saw nothing of that, so... you may certainly well be right.



The sad story of the unicode snowman:

http://news.ycombinator.com/item?id=2035926


I'm not clear how how the snowman translates to xn--n3h...?




Does anyone know how to get the URL to be displayed as http://☃.net rather than http://xn--n3h.net in Firefox?


it's displayed like that on purpose, so folks don't fake banking site or other important domains with unicode characters that are similiar to the standard latin characters but aren't actually them.

look in about:config for

"network.IDN.whitelist.tld" and put .ws in there


Ah didn't think about that. I guess I should leave it as it is after all ;/

But won't this have to be enabled in other countries (like China), where they actually want to see the Unicode characters? Isn't that going to increase their risks of being phished?


Different top-level domains have different rules about what characters they'll allow in domain names. If a registrar has policies that prevent homograph registration, Firefox (and other browser vendors) will add them to the default whitelist.

If the .cn registrar has a policy that says "only ASCII and CJK characters will be allowed in domain names", then it's pretty unlikely that a homoglyph attack will work, and browsers will show the properly-decoded string as the domain name.

(and sure enough, my copy of Firefox sets "network.IDN.whitelist.cn" to true)


I only see a black bordered box. I'm on WinXP and latest dev channel chrome. Am I missing something?


http://dl.dropbox.com/u/4073000/sonwman.png (Screenshot for all those who can’t see it.)


Funny it looks different on my computer: http://dl.dropbox.com/u/2601554/snowman.png


Awww, yours looks much nicer. The page doesn’t specify a font so I guess it’s just using the default font of the browser (or, failing that, any font that contains the Unicode character; I would be surprised if many do). My default font is Helvetica Neue.

– edit: I did some digging around. There are a few fonts on my computer which contain a snowman. My browser uses (for whatever reason) Hiragino which is a family of Japanese fonts that comes standard with every Mac. Your nice snowman is the one in Arial Unicode. Here are all the snowmen on my system: http://dl.dropbox.com/u/4073000/snowmen.png


I wonder if Hiragino isn’t the choice due in part to the prevalence of emoji in Japan.


I'm seeing the MS Gothic Snowman from your screenshot with Opera 11 on Win XP. The Kozuka Snowman is cute. Thank you all for the fun with unicode and fonts :)


Another snowman. This one on Chromium under Arch Linux, no idea what font its ended up using, but I think it looks much happier.

http://dl.dropbox.com/u/6266437/Snowman.png


Of all the screenshots, mine looks the ugliest. ( http://i.imgur.com/zX3tG.png ). Chrome Beta on OS X Snow Leopard.


Best one imo. Our snowman has a fez!


Here's mine, using Firefox 4 Nightly on Arch Linux. http://imgur.com/yGXVS.png


Firefox 3.6.13 on Arch Linux: http://i.imgur.com/G8AVb.png

My default font is called 'serif'.


Firefox 3.6.6 on Ubuntu 10.04 - He looks pretty happy! http://i.imgur.com/v8zXx.png



This claims to be a mirror, but it's sourcing the font file from the original website. That means if unicodesnowmanforyou.com goes down, so does the mirror. Not a mirror.


So why does Unicode Snowman exist? What language uses it?


It's in the Symbols & Dingbats subrange, along with the comet (☄), the cloud (☁), the umbrella (☂), the hot springs (♨), all the card suits, all the chess pieces (both white and black), a few musical notes, and various and sundry others. Open Character Map in Arial Unicode font and you can see for yourself.

Not all Unicode symbols are meant for languages.


I love Unicode (my license plate? 0x2764 ❤) and that these symbols exist, but get upset that they still don’t see a reason to add a padlock (or two, one open and one closed) or play/pause/ffwd/rev symbols.


I have nothing but sympathy for the Unicode Consortium which finds itself in a difficult situation. The main purpose of Unicode is to contain all characters of all the world’s writing systems, not to be a directory of the world’s symbols. Yet there clearly are some symbols which absolutely have to be part of Unicode. Punctuation, for one.

Where to draw the line? That’s not easy.


I don't know, but it would sure make it easier to add a bare-bones media player widget or a weather report widget to a spartan desktop interface such as xmonad. Plus your widgets would be automatically restyled to be aesthetically consistent with the rest of the interface!


Unicode 6.0 has: 1F512 and 1F513 closed and open locks. Also 1F511 is a key, 1F510 is a key and lock, and 1F50F is a lock with ink pen(?).

Unicode 6.0 also adds ffwd and rewind symbols but frustratingly leaves out pause and play. Plenty of triangles for play (which may or may not match the new additions). But two vertical lines or boxes are going to be very inconsistent from font to font in looking like a pause symbol.

My biggest peeve? The standard power button symbol on every computer and almost everything else these days, the IEC 5009 standy icon, also recognized as the power symbol in IEEE 1621. You can approximate the IEC 5010 on/off button with some other things, but it won't reliably look the same, and won't look at all like the IEC 5009 icon.

Pause, play, and power buttons are WAY WAY WAY more useful (i.e. necessary) in text than a nail polish icon.


Thanks for the info, and I cannot agree enough.


I'm holding out for Mr Potato Head combining characters.


> ...or play/pause/ffwd/rev symbols.

I'm not sure about "pause", but Unicode has had U+25B6 BLACK RIGHT-POINTING TRIANGLE for "play" for some time, and Unicode 6.0 introduces U+23E9 BLACK RIGHT-POINTING DOUBLE TRIANGLE for "fast-forward" and U+23EA BLACK LEFT-POINTING DOUBLE TRIANGLE for "rewind".


Unicode isn't strictly for language characters. It includes many other common symbols, borders, decorations, and other misc.

See here for the full block that it resides in: http://en.wikibooks.org/wiki/Unicode/Character_reference/200...


Why is it wearing a fez?


Because it can.


http://www.gооgle.com <-- attention to details...



One of my friend had a similar domain but he got it refunded because it didn't worked in all browsers etc and would not in future either according to him because of ICANN's specs which go against symbols in domain names.


> go against symbols in domain names

Can you clarify which symbol it was? The only related problem I know is that homoglyphs in domain names are problematic due to their potential abuse for phishing. So for example a smallcaps letter O or C that looks like a lowercase o or c would be problematic.

Edit in response: The Apple symbol  (U+F8FF) is not cross-platform. It’s explicitly defined as a code point that a system (OS X in this case) can define for its own “private use.” Given this, it’s not surprising that it would go against ICANN policy and cause problems in various browsers.


The symbol was  [Edit: Image link http://i.imgur.com/Pf0Br.png ]

The domain name was .ws


On my machine, that looks like the Euro symbol.


How to monetize this? In general, how to monetize being funny/entertaining (on the web)?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: