LWN: Comments on "Schrödinger's 😻 and outside-the-box naming"

proper German

daenzer — Fri, 17 May 2013 14:01:13 +0000

"Der Dativ ist dem Genitiv sein Tod." :)

Schrödinger's box and outside-the-box naming

n8willis — Tue, 14 May 2013 22:22:50 +0000

On the contrary, I contemplated using U+25A1 in the headline, but this way we get the best of both worlds: cat graphics on supported systems, where the hilarious outside-the-box pun does double-duty, and a robust fallback character that still provides a rich, nonsensical experience.

Nate

Schrödinger's 😻 and outside-the-box naming

chithanh — Sun, 21 Apr 2013 03:11:40 +0000

Next release name should maybe contain some CTL. Perhaps Hindi (Ordering) or Arabic (Bi-directional context-sensitive shaping à la فدورا ١٨).

If they are less daring, they could name it aleph-0 (written as א‏‎₀ with possible fallback to ℵ₀)

Schrödinger's 😻 and outside-the-box naming

rbowen — Fri, 19 Apr 2013 19:09:51 +0000

Absolutely, call this release Schrodingers Cat, and move on. But call the release 20 Gödel's Dog and make that the timeline for fixing the problems.

proper German

panzi — Fri, 19 Apr 2013 14:21:48 +0000

Well, Schrödinger was Austrian, so I'd say: "De Kotz vom Schrödinga"

Schrödinger's 😻 and outside-the-box naming

mirabilos — Thu, 18 Apr 2013 15:16:41 +0000

There’s the issue with X11: BDF fonts may contain only glyphs within a range of roughly 65535, meaning that, if your font includes a glyph for U+0000 or even a space U+0020 you don’t get to add SMP characters to it.

And I *like* fixed-misc!

Schrödinger's 😻 and outside-the-box naming

mirabilos — Thu, 18 Apr 2013 15:15:12 +0000

On the other hand, 「’」 *is* the proper character to use for a Genitive construction in the English language, not 「'」.

Schrödinger's 😻 and outside-the-box naming

mirabilos — Thu, 18 Apr 2013 15:12:40 +0000

The BMP (Basic Multilingual Plane) is mostly decent.

Best thing, it fits into 16 bit (0‥FFFD inclusive) and its UTF-8 form uses only up to three octets.

Schrödinger's 😻 and outside-the-box naming

mirabilos — Thu, 18 Apr 2013 15:11:16 +0000

Also in Konqueror (KDE 4.10, Debian experimental)…

Schrödinger's 😻 and outside-the-box naming

meyert — Sat, 13 Apr 2013 12:00:54 +0000

Omg! What year is it again?

Schrödinger's box and outside-the-box naming

redden0t8 — Thu, 11 Apr 2013 18:01:09 +0000

Somewhat off topic... I just about died laughing when I discovered iOS will speak unicode symbols as their formal description - by asking my phone to read a sentence containing a "PILE OF POO".

Schrödinger's 😻 and outside-the-box naming

jond — Thu, 11 Apr 2013 14:22:30 +0000

This is the first time I've wanted a "report abuse" button on LWN.

Schrödinger's 😻 and outside-the-box naming

mordae — Thu, 11 Apr 2013 11:17:55 +0000

Nice kitty in Firefox on F18. :-)

proper German

Seegras — Wed, 10 Apr 2013 08:34:51 +0000

Yes, but this was about "proper" German ;) I'd also say "Em Schrödinger sini Chatz", but that's rather middle-high-german influenced, and surely not "proper" modern high german.

Not a UTF-8 issue.

jezuch — Tue, 09 Apr 2013 06:59:05 +0000

> Hm. Surely Bilious Badger would be an Ubuntu codename, not Fedora...

Or it could be truly international with Żółty Żółw :) (Yellow Turtle, that is).

Schrödinger's 😻 and outside-the-box naming

Comet — Mon, 08 Apr 2013 18:09:27 +0000

Provide correct metadata that matches their advice.

Schrödinger's 😻 and outside-the-box naming

jwboyer — Mon, 08 Apr 2013 15:13:11 +0000

>Robinson noted that Fedora 19's release name was chosen roughly six months
>ago during the Fedora 18 Alpha period; nevertheless it took six months for
>anyone to encounter a bug related to it precisely because of how deeply
>buried the problem was.

Actually, the problem was discovered really quickly. The name was chosen 6 months ago, but nothing actually switched to using the name until the fedora-release package was updated. That happened when Fedora entered Branch state, and the problem was hit and reported within a day or two.

It wasn't buried deeply, it was just that there was no software using the new release name. The rawhide repo is always just 'rawhide'.

Schrödinger's 😻 and outside-the-box naming

hamjudo — Mon, 08 Apr 2013 15:02:54 +0000

There are a variety of Unicode regular expression libraries. Characters are in classes, alphabetic, currency symbols, numerals,uppercase, lowercase, etc... Companies running search engines will use their own rules (aka. business logic) on what they choose to index for each part of a document, possibly dependent on type of document. Some may recognize source code, and then use programming language specific rules for indexing symbols (C is case sensitive, Fortran is not).

If enough people start using the cat characters in semantically significant ways, there will be a business case for entities to index those characters.

Schrödinger's 😻 and outside-the-box naming

dwmw2 — Mon, 08 Apr 2013 13:49:27 +0000

Which is a duplicate of #404856, filed in 2007.

Remind me why I'm still actually using Firefox?

Schrödinger's 😻 and outside-the-box naming

dvdeug — Mon, 08 Apr 2013 11:13:10 +0000

What's Unicode supposed to do about it? ASCII provides a ' character that does all sorts of different jobs. There's no obvious errors in the Unicode's expansion of that to a large character; even if it's arguable they could have done better in retrospect, even the best would have given them several new characters and a confused role for '. You could use it for apostrophe, at the cost of confusing the ASCII straight line with the curved line expected in typography. No real win here.

Not a UTF-8 issue.

nix — Mon, 08 Apr 2013 08:11:27 +0000

Hm. Surely Bilious Badger would be an Ubuntu codename, not Fedora...

Not a UTF-8 issue.

sfeam — Fri, 05 Apr 2013 20:27:00 +0000

Furthermore, a binary file containing well-formed UTF8 is presumably harmless if printed or otherwise treated as text - at least it's no worse than a binary file that happens to contain only ascii characters.

Not a UTF-8 issue.

smurf — Fri, 05 Apr 2013 20:14:39 +0000

it is very easy to check for well-formed UTF-8 encoding.

The chance that a binary file contains only well-formed UTF-8 is reasonably small even with fairly short files.

So what exactly is the problem?

Schrödinger's 😻 and outside-the-box naming

smurf — Fri, 05 Apr 2013 19:56:24 +0000

Oops, my fault -- I mistakenly thought that the title's cat has a code < U+10000. Apparently that was a copy+paste error on my part.

Schrödinger's 😻 and outside-the-box naming

k8to — Fri, 05 Apr 2013 10:24:17 +0000

Schrodinger's box with 01F63B or something inside.

Right.

Schrödinger's 😻 and outside-the-box naming

epa — Fri, 05 Apr 2013 09:22:13 +0000

Is there some officially defined subset of Unicode which includes characters needed for writing but excludes cats, PILE OF POO and so on?

Not a UTF-8 issue.

jezuch — Fri, 05 Apr 2013 08:26:24 +0000

> Even in ISO8859-1, the heuristic limit of 2% meant that you need a file that's 50 characters long before you're allowed even a single accented character.

Any limit other than 100% would make it impossible to name a Fedora release "żółć" ("bile").

Schrödinger's 😻 and outside-the-box naming

gdt — Fri, 05 Apr 2013 02:15:40 +0000

Also problematical might be

From

Schrödinger's 😻 and outside-the-box naming

Comet — Thu, 04 Apr 2013 23:47:00 +0000

You're right, thank you for the education.

Unicode: consistent, until you want to rely upon the consistency.

So what's the point of marking characters as opening or closing punctuation if a machine can't rely upon well-formed text actually using that, because the distributed files themselves recommend using the characters in violation of their metadata?

*sigh*

💩

Schrödinger's 😻 and outside-the-box naming

Comet — Thu, 04 Apr 2013 23:40:36 +0000

So 0001f408 is out for not being in the basic plane, but SMILING CAT FACE WITH HEART-SHAPED EYES, at 0001f63b, is okay?

(FWIW, in Chrome I get the replacement box in the page and the 0001f63b rendered correctly in the table title, which is ironic, but in Firefox things render fine in the page itself).

Schrödinger's 😻 and outside-the-box naming

blitzkrieg3 — Thu, 04 Apr 2013 20:17:23 +0000

Link for the lazy: https://ubuntuone.com/3u64PJ78P0p0eFrSqo38I5

Schrödinger's 😻 and outside-the-box naming

hamjudo — Thu, 04 Apr 2013 17:45:53 +0000

It is all a matter of fonts for me. The character is formally known as "SMILING CAT FACE WITH HEART-SHAPED EYES (U+1F63B)" Right now I'm using Chrome on my employer provided Windows 7 system, and I see a white box with a black border when using the default font. At home on a personal system running Chrome on Ubuntu, I saw a white cat with a black border. On my Nexus 4, the cat is black. A little googling took me to http://www.fileformat.info/info/unicode/char/1f63b/fontsupport.htm, which contains a link to an evil flash based Local Font List tool. Which will show which fonts (that are visible to flash in your browser (assuming you want to run flash apps at all.)) have a glyph for that character.

But none of that is important. What is important is that most search engines don't index Unicode emoticons. So if I search google for "Schrödinger's 😻" I get the exact same results as just searching for "Schrödinger's", since google ignores the "😻". Unicode searches are supported, as long as the characters are in the one of the indexed classes, so a search for "这" returns billions of hits (google tells me it means "this").

Schrödinger's 😻 and outside-the-box naming

BlueLightning — Thu, 04 Apr 2013 17:29:13 +0000

> u1F408 is not in the basic plane; many systems don't have a character
> set installed which has that glyph. My Debian system apparently doesn't,
> and I'm a font junkie.

Indeed, same here on Kubuntu. The "fancier" one in the title renders just fine though.

Schrödinger's 😻 and outside-the-box naming

smurf — Thu, 04 Apr 2013 17:02:00 +0000

>> 🐈 not good enough

u1F408 is not in the basic plane; many systems don't have a character set installed which has that glyph. My Debian system apparently doesn't, and I'm a font junkie.

Schrödinger's 😻 and outside-the-box naming

mezcalero — Thu, 04 Apr 2013 16:47:17 +0000

There's some mix-up in the article regarding the names of the files.

There's /etc/os-release which is an environment-like file, which contains more than a single line, has a man page, and is somewhat standardized among distros, because we require it in systemd. And there's Fedora's/Red Hat's old /etc/system-release which really just contains a single line, doesn't have a man page (in fact I am not aware of any documentation of it). Both are provided by fedora-release.rpm.

At this point /etc/system-release really should be considered legacy.

Schrödinger's 😻 and outside-the-box naming

jwakely — Thu, 04 Apr 2013 16:47:02 +0000

I think you're agreeing with the parent, which is saying character set non-performance would be "UTF-8 fail" but seeing a box (with or without hex digits inside) is font non-performance, and that's what the grandparent was describing.

Schrödinger's 😻 and outside-the-box naming

smurf — Thu, 04 Apr 2013 16:22:05 +0000

No, that'd be character set nonperformance.

Font nonperformance (i.e. no local font has the glyph) would get you one box with hex digits inside.

Schrödinger's 😻 and outside-the-box naming

jimparis — Thu, 04 Apr 2013 15:19:50 +0000

It's been reported here: https://bugzilla.mozilla.org/show_bug.cgi?id=857913

proper German

hirnbrot — Thu, 04 Apr 2013 15:06:39 +0000

Where I am we'd say "Dem Schrödinger seine Katze".

Schrödinger's box and outside-the-box naming

geuder — Thu, 04 Apr 2013 15:02:35 +0000

Right. http://picpaste.com/pics/box1-VvZq4ok7.1365086770.png

I guess the author might not have understood the double irony in his headline :)

After having "debugged" the first email from an airline with a small airplane http://www.fileformat.info/info/unicode/char/2708/index.htm as the first character of the subject already some time ago, I had a look at some Unicode tables and could easily guess what the box stands for. Really wondered how they accepted
http://www.fileformat.info/info/unicode/char/1f4a9/index.htm ...

Lesson learned: airplane is more likely to appear in the browser's default font than kitten.