Hacker News Reader

Comment by elcritch | Hacker News Reader

If you're really adventurous, you could just use Braille which has 255 Unicode symbols. Ahem

ip6emoji("fe8000000000000003ceecdfffe30c27",Char(0x2800)) => "⣾⢀⠀⠀⠀⠀⠀⠀⠃⣎⣬⣟⣿⣣⠌⠧"

Then:

   deadbeef000000000000000000000001 
   2607f2f8a36800000000000000000002
   fe8000000000000003ceecdfffe30c27
   fe800000000000000000000000000001
   2607f8b040078090000000000000200e

Becomes:

   ⣞⢭⢾⣯⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠁
   ⠦⠇⣲⣸⢣⡨⠀⠀⠀⠀⠀⠀⠀⠀⠀⠂
   ⣾⢀⠀⠀⠀⠀⠀⠀⠃⣎⣬⣟⣿⣣⠌⠧
   ⣾⢀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠁
   ⠦⠇⣸⢰⡀⠇⢀⢐⠀⠀⠀⠀⠀⠀⠠⠎

At first I was just playing around, but after a bit it begins to resemble one of those binary clocks. It even becomes somewhat natural to read. Might actually use this for myself... something nice about the 2x4 bit block patterns. 64bit pointer addresses?

View on HN

That results in a different usability problem. Installed fonts are very nonstandard.

At the moment I happen to be on an up-to-date version of Windows 7 and here's what it looks like:

http://imgur.com/U5vZFRQ

*edit: though it appears to render properly on Debian

"up-to-date version of windows7" reads like an oxymoron.

Why? Windows 7 gets regular updates. Windows 8/8.1/10 are not updates to Windows 7, it is a different operating system.

Edit: Not sure why all the downvotes, I honestly don't see how "up to date Windows 7" is an oxymoron.

Upvoted because downvotes without reasoning is pointless.

In regards to why I think you may being downvoted:

> reads like an oxymoron

The parent was being a bit facetious. We all know they are different operating systems. But it's far less likely that an operating system that no longer receives "feature" updates would be up-to-date on its fonts.

In any case, an oxymoron is (per Google):

> a figure of speech in which apparently contradictory terms appear in conjunction (e.g., faith unfaithful kept him falsely true ).

Which I think the parent accurately described. (It's apparently contradictory, but has meaning.)

> Why? Windows 7 gets regular updates. Windows 8/8.1/10 are not updates to Windows 7, it is a different operating system

I'll address this point, the sibling comment already addressed the other point (it wasn't meant as a criticism, by the way).

It's true that Windows 7 still get updates. However, these are mostly security updates, or bugfixes. We don't expect it to get new features. In this case, unicode font rendering. It is likely (but not guaranteed) that a user running Windows 10 would see a much better result.

But then that user would be running Windows 10, which means they get to pay Microsoft for the privilege of having their data harvested and having an advertising ID assigned to their device.

Why any sane individual willingly chooses windows 10 is beyond me.

edit: Can't reply, but turning off the "phone home features" doesn't actually turn them off. It still phones home and you still get assigned an advertising ID. Sure, I could block it at the firewall potentially, but this is about taking an ideological stance.

It looks just like yours on Win 10 under Chrome. I was curious and tried in Edge and it renders fine.

Also, the settings in Win 10 allow you to turn off a lot of the phone home type stuff. Other than that, my opinion is that 10 mostly feels like a more polished 7.

Windows 10 on Chrome here, I see the same as the Windows 7 user above.

Actually, it renders perfectly on Edge; so the problem is probably with Google Chrome.

Doesnt work on my main up-to-date Win8 computer either.

(though interestingly when I open this page up with links2 on my vps, accessed through a terminal on the iphone, which is using a monospace font... it renders fine.. it doesn't render fine if I see the same thing with Putty on the Win8 computer. I guess fonts support in general on Win is bad)

>That results in a different usability problem. Installed fonts are very nonstandard.

And some users change their default font and don't even use installed fonts and override preferences because they're in charge of how things display on their computer.

I use Sofia-Pro and this is what I see: http://i.imgur.com/8PoNfdJ.png

Your browser/OS should be using font substitution¹ though when it encounters a character that the current font doesn’t have a glyph for. Font substitution is pretty much necessary since it’s currently impossible for one font to hold Unicode’s repertoire of 110,000+ characters; even the OpenType format is limited to a maximum of 65,536 glyphs.

Assuming that it is using font substitution (as it should be whether or not you’ve chosen Sofia Pro as your browser’s default sans-serif font… unless you’ve also changed some other settings too), then the reason those characters show up as question marks is because you don’t have any fonts installed containing glyphs for those characters. (I’d recommend Everson Mono² or Symbola³).

――――――

¹ — https://en.wikipedia.org/wiki/Font_substitution

² — https://en.wikipedia.org/wiki/Everson_Mono

³ — https://web.archive.org/web/20150625020428/http://users.teil...

>Assuming that it is using font substitution (as it should be whether or not you’ve chosen Sofia Pro as your browser’s default sans-serif font… unless you’ve also changed some other settings too),

Yes, actually. The only two fonts my browser is permitted to use are Sofia Pro and Meiryo. My point being nothing is wrong with alphanumeric representations of hexadecimal. Short of glyph fonts like Webdings, every font supports alphanumerics - even CJK fonts.

Some users change settings - heavily so. The "safest default" should be the assumption. Changing something that works for most people to work only for "people with a proper supporting font installed" is breaking the web as far as any devs should be concerned.

I parsed that as Conway's Life. Which is odd, because I've not played with, nor seen it, since about 1993! I'll be plumbing that into Life to see what happens.

hehe, now I'm going to envision a giant Conway's game of life being played every time I see a room full of IoT devices on IPv6.

Wonder if there would be an interesting way to visualize a cascading failure of the kind that brought down AWS in past years. Which of course makes one wonder if there would be any kind of useful calculus for a series of such glyphs.

PS: according to wikipedia there is a calculus of communicating systems of sorts: https://en.wikipedia.org/wiki/Calculus_of_communicating_syst...

I like where you're going with this. By using a base 256 representation you shrink hex's 32 charachters down to 16, almost the same as IPv4's decimal dot representation of 15 characters: 255.255.255.255. But obviously you don't address the typeability issue.

Without using the shift key it's easy to type base 32 numbers. That comes out to an average of 26 characters per IPV6 address.

So Hex

  deadbeef000000000000000000000001 
  2607f2f8a36800000000000000000002
  fe8000000000000003ceecdfffe30c27 
  fe800000000000000000000000000001 
  2607f8b040078090000000000000200e

Becomes base 32:

  6ULMVEU0000000000000000001
  160VPFH8R80000000000000002
  7UG000000000007JNCRVVU6317
  7UG00000000000000000000001
  160VSB0G07G28000000000080E

If we are willing to use the shift key we could move up to Base 64 and get it down to an average of 21 characters. (Anyone know why the standard base 64 alphabet starts at A instead of 0?):

  3q2+7wAAAAAAAAAAAAAAAQ
  Jgfy+KNoAAAAAAAAAAAAAg
  /oAAAAAAAAADzuzf/+MMJw
  /oAAAAAAAAAAAAAAAAAAAQ
  Jgf4sEAHgJAAAAAAAAAgDg

Now if we are willing to use unicode characters, why stop at base 256? Unicode has 95,000 characters so we could use a base 65,536 number and cut it down to 8 characters, now we're talking! Unfortunately to get it down to 4 characters would require 4,294,967,296 different characters. Even the extended unicode set won't get us there.

But maybe an alphabet with emoji could be practical. You could have smiley or sad faces in your IP address. All kinds of possibilities with that. Though software such as this HN website would have to be fixed to be able to display it. But you know, a long term project.

I like the base32 one. If you put two colons in front then you can even copy/paste and it distinguishes it from other text as an ipv6.

  ::6ULMVEU0000000000000000001
  ::160VPFH8R80000000000000002
  ::7UG000000000007JNCRVVU6317
  ::7UG00000000000000000000001
  ::160VSB0G07G28000000000080E

There's an RFC for IPv6 addresses in Base85: https://tools.ietf.org/html/rfc1924

> 7. Implementation Issues

> Many current processors do not find 128 bit integer arithmetic, as required for this technique, a trivial operation. This is not considered a serious drawback in the representation, but a flaw of the processor designs.

Oh, IETF, please never change.

Why, thanks! The original inspiration is probably related to http://what3words.com/. Mainly I thought it'd be "mad scientist" thing but it actually is fun and possibly useful in some contexts. As for usability, you could use text substitution [1] and have something like `ip174` change into `⢮`. Setting up 256 mappings isn't too terrible, especially since there is a system to Braille dot ordering.

[1]: http://lifehacker.com/162484/save-time-with-text-substitutio...

Base 32 seems like a good half-way solution of length reduction vs glyph complexity, as Base 64 or Base 89 (or whichever the RFC suggest) include too many distracting characters (IMHO).

Of course, Chinese/Japanese speakers have a natural advantage here (Katana and Hirigana both fall short of a contiguous 64 character mapping):

ip6emoji("fe8000000000000003ceecdfffe30c27",Char(0x3300),stride=8) => "㏰㌀㌏㏷"

( though hopefully that's not some form of insult in Chinese! ;) )

For reference, it looks like this on Safari v9.0.2 on OS X v10.11.2 'El Capitan':

http://f.cl.ly/items/0E182U1p3r430M073w2i/braille.png

Seems like the font that OS X’s text rendering system substitutes for the braille characters (U+28E3, etc.) is Apple Braille Regular over the other fonts installed that also have glyphs for those codepoints (Apple Symbols, Everson Mono (font I installed myself), and Symbola (font I installed myself)).

As an aside (and yes, I’m copying part of a post I made more than a year ago¹; I’m still interested in knowing the answer to this!) I’m pretty interested on how OS X and Windows decide on which font to use when there are multiple fonts installed containing the required glyph. For example, I have two other fonts on my OS X system that have a glyph for U+2705 (Everson Mono and Symbola), but OS X always seems to consistently pick Apple Color Emoji’s glyph. Maybe OS X’s text rendering system goes through the fonts in alphabetical order and uses the first one it finds containing the required glyph? It would be great if end users could have a bit more control over the font substitution process. I know it’s possible to do in some text editors like Emacs², but I believe that programs like that use their own text rendering systems instead of that supplied by the OS (could be wrong though).

――――――

¹ — https://news.ycombinator.com/item?id=8865067

² — http://stackoverflow.com/questions/6491202/overriding-emacs-...

I guess if you make sure to always use a mono-space version of Braille (I'm not sure that makes sense, it could be a requirement it be mono-spaced) and have really good whitespace awareness. That output gives me a very good quick general feel of an address, but very poor fidelity.

Reply on news.ycombinator.com