Conversation

Notices

1. Nate Cull (natecull@mastodon.social)'s status on Friday, 01-Dec-2017 00:20:28 UTC Nate Cull

   One of the things that most annoys me about computing circa 2017 - which has kinda always been true but I hoped it would get better, and it hasn't - is that we have a serious lack of truly fundamental, correct concepts.

   Information is almost literal thought-stuff. Costs a bit of energy, space, time, yes, but compared to bricks, it's like magic super-Lego. Almost anything we can imagine, we can create.

   But we've created a vast tangled web we can't really understand. Didn't mean to. Just did.

   • Hallå Kitteh likes this.
   • Hallå Kitteh repeated this.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 00:56:53 UTC Hallå Kitteh
     @natecull Unicode will definitely survive. ASCII is still around, and Unicode is more tied into the ecosystem than even ASCII was -- several languages that have already reached COBOL status have Unicode at their core.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 03:11:12 UTC Hallå Kitteh
     @natecull Whatever character encodings and code points we are using in 2200, it will all be called Unicode, or the equivalent of "Unicode" in Sino-Anglo-Slav-Spanish (which is probably going to be "Unicode" anyway!). Such is the nature of Unicode. :-D
     
     Seriously, decoupling the code point tables from the encoding was a great call, and one that means that replacing the code point tables is basically meaningless and a waste of effort. But UTF-8 is probably not the last encoding. But I do think that even in 2200, there will definitely be UTF-8 lying around, and probably even in prominent use still.
     
     We are using safe subsets today. The problems we have seen in certain TLDs have been due to their failure to follow the guidelines explicitly instructing them to use safe subsets.
     
     Full-blown Unicode is great for human display purposes. For identifiers, not so much.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 03:12:06 UTC Hallå Kitteh
     @natecull Scope solves everything when it comes to magic values.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 03:27:49 UTC Hallå Kitteh
     @natecull Nope!
     
     I think that is something that will have to be done per domain. There is no generic safe subset.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 03:31:21 UTC Hallå Kitteh
     @natecull It's not incidental Unicode complexity, it's essential complexity of human language and its expressions.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 03:36:50 UTC Hallå Kitteh
     @natecull ASCII is a safe standardized subset of Unicode. It's just that most people consider it insufficient.
     
     Now I'm imagining CJK unification, but with Latin, Cyrillic and Greek and probably other alphabets I'm not aware of.
     
     That should be fun.
   • D. Joe (deejoe@mstdn.io)'s status on Friday, 01-Dec-2017 03:49:10 UTC D. Joe

     in reply to

     • Hallå Kitteh

     @clacke @natecull

     most language complexity arises out of what you might call folkways.

     to the extent unicode represents existing natural language, I'm down.

     to the extent it grows in disparate directions to sell phones, not so much.

     this is where we get to the gripe about having :poo: before Bengali was adequately included

     Hallå Kitteh likes this.
     Hallå Kitteh repeated this.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 05:53:50 UTC Hallå Kitteh

     • h
     @h @natecull Yes, but ASCII is one single thing. And the unicode code points for those various encodings is one single table, where -- if you normalize -- there is only one "latin letter ï".
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Friday, 01-Dec-2017 05:59:05 UTC Hallå Kitteh

     • p̸h̴o̸⁴c̵e̴ₓa̵
     @pho4cexa @natecull
     
     Also I believe in worse is better. Not as an absolute, but as a true an beneficial force. Better is better is a beneficial and opposite force.
     
     We wouldn't have all these internets and pocket super computers if we cared to get everything right.
     
     If you accept this, there is still an argument to be had about "how much worse?". How much are our pocket miracles and instant access hurting us and how much are they hurting us and our environment and societies? And how can we mitigate that?
     
     I don't believe that waiting for Xanadu would have made a better society. I found the 80s super boring and I'm happy for the shambled mess that Sir Tim and others brought us.
   • ❄️🦊 (icefox@octodon.social)'s status on Monday, 04-Dec-2017 01:19:48 UTC ❄️🦊

     @natecull part of the problem is that mathematics is pure but the machines that implement them are impure

     Part of the problem is humans just suck at building systems

     Part of the problem is we still don't understand the details

     Programming is currently like engineering in the 1850s, when the laws were empirical, there were lots of colliding locomotives and collapsing bridges, and every machine shop made it's own, incompatible nuts and bolts.

     Hallå Kitteh likes this.
     Hallå Kitteh repeated this.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 01:26:57 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @seanl @vertigo @natecull @deejoe And then we just need a way to map different encodings to each other and boom there is Unicode.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 01:44:12 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @vertigo @natecull @deejoe @seanl "Unicode itself" isn't applied, every encoding is an application of it. That's why it's misguided to argue against Unicode.
     
     Use the bits and bobs that make sense. Don't put poop emojis in your font if you don't like them. Don't allow them in your IM application if you think it detracts from its function or presentation.
     
     Unicode doesn't have much of an agenda apart from what people bring to it and "encode all the things". Possibly an agenda against redundancy (CJK), except when it doesn't (Mathematical Alphanumeric Symbols). If the Bengali Unicode block was based on ISCII and ISCII was lacking, that's where to put the blame. There is no anti-Bengali agenda, just nobody pulling for support.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 01:46:28 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @seanl @vertigo @natecull @deejoe Are there character encodings that don't round-trip into Unicode and back?
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:10:05 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @natecull @vertigo @deejoe @seanl I agree that this hurts our heads. But I don't see a solution that isn't limiting ourselves.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:12:17 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @seanl @vertigo @natecull @deejoe English on Twitter and Facebook ...
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:15:44 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @seanl @vertigo @natecull @deejoe That means any software that allows you to get exposed to a service in mass media, and then seek it out, expecting some measure of security.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:20:15 UTC Hallå Kitteh

     in reply to

     • Hallå Kitteh
     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @vertigo @natecull @deejoe @seanl Or, shorter: Any digital model that maps to human society is broken by design.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:44:28 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @vertigo @natecull @deejoe @seanl If you have a glyph that needs a serial number, might as well put it in the universal table of numbers for glyphs.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:51:25 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @vertigo @natecull @deejoe @seanl I don't know at a glance what's stack-based about BERT, but I'm noticing Factor among the dozen of implementations.
     
     When the document came to BERP I just felt that this whole thing is an excuse for a Rick&Morty fan to get to name things. :-D
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:55:53 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @seanl @vertigo @natecull @deejoe That sounds like something that could replace UTF-8 for texts that are largely outside the Latin plane -- like a shift-UTF-8.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 02:58:57 UTC Hallå Kitteh

     • Vertigo
     • h
     • D. Joe
     • Cozy Eugen
     @h @vertigo @gargron @natecull @deejoe The fact that Masto wants to get into Unicode is a sign of the things I do agree are broken. Smiling poop with customizable hair belongs more in Unicode than Adobe product logos do.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 03:00:14 UTC Hallå Kitteh

     • Vertigo
     • Sean R. Lynch ☑️
     • D. Joe
     @vertigo @natecull @deejoe @seanl Indeed Pleroma and Mastodon already solved the emoji entity problem. ;-)
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 03:04:18 UTC Hallå Kitteh

     • Vertigo
     • h
     • D. Joe
     • Cozy Eugen
     @h @vertigo @gargron @natecull @deejoe
     
     Ah yes, sorry. FontAwesome maps to PRIVATE_USE_AREA. Different thing.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 04:12:53 UTC Hallå Kitteh

     • h
     @h
     
     > Katakana
     
     You mean Kanji. There are ~46 katakana.
   • Sean R. Lynch ☑️ (seanl@social.literati.org)'s status on Monday, 04-Dec-2017 04:52:46 UTC Sean R. Lynch ☑️

     • Hallå Kitteh
     • Vertigo
     • h
     • D. Joe

     @natecull @deejoe @clacke @h @vertigo It's all about which specific properties you value. If you value ease of indexed access over compact representation, you use a fixed-width representation like UTF-32. But you're always engaging in some sort of compromise. Unless you're Terry Davis.

     Hallå Kitteh repeated this.
   • Hallå Kitteh (clacke@social.heldscal.la)'s status on Monday, 04-Dec-2017 05:18:57 UTC Hallå Kitteh

     • tekk
     • Sean R. Lynch ☑️
     @tekk @seanl God is the one factor he doesn't compromise with.