Better discussions with Zulip (https://zulip.com). Recurse Center '14. Previously: Dropbox, Ksplice, MIT.
Communicating through and about code. Secretly still a mathematician. He/him. 🏳️🌈
Across countries and even across US states, more isolated capitals are associated with worse government, less transparency, and more corruption:
https://www.aeaweb.org/articles?id=10.1257/app.20170111
https://www.aeaweb.org/articles?id=10.1257/aer.104.8.2456
It's a choice for Egypt to decamp from Cairo and give contracts to military affiliates rather than to build up Cairo's infrastructure with competitive bidding, and for Iran to move the capital to where dispossessable minorities live rather than to force its agriculture sector to be more water-efficient.
Capital Cities, Conflict, and Misgovernance by Filipe R. Campante, Quoc-Anh Do and Bernardo Guimaraes. Published in volume 11, issue 3, pages 298-337 of American Economic Journal: Applied Economics, July 2019, Abstract: We investigate the links between capital cities, conflict, and the quality of go...
#Thanksgiving is coming! For many, that means politico-talks w/ #MAGA fam
Sharing a recent convo, to show how they can be less-awkward/more-effective
Keys
* Ask questions - engage left brain
* Establish trust - take down defenses
* Focus - ignore, call out distractions
* NO debate, compromise, or appease
* Bonus lesson: manipulation in news/socials
IRL, they ack'ed they were wrong about the protests. Small win, but progress!
Would love to hear others' experiences in replies
🙏Please boost 🤙
@janxdevil @mhoye @irvingreid I think this fundamentally goes back to my first reply: by that logic, anything is possible.
You're describing a scheme that would be completely incompatible with all the software we have today that processes text. Could somebody in the future spec out such a scheme? Sure maybe, who knows. But it doesn't have anything to do with Unicode as it actually exists today, and an incompatible break like that could do anything. There's no basis for specific predictions like having 6 bytes be the new longest encoding.
The last incompatible change anything like that was the introduction of UTF-16 itself (i.e., of surrogates), and that was 1996. At that time Unicode had barely begun to be implemented. A comparable change today would be vastly costlier.
"Multiple cisgender women have turned down spots on a list of the top female cyclists in the U.K. because the organization compiling the list banned trans women.
'If they don’t want to ride with all women, then it’s not the kind of ride I want to be part of,' Claire Sharpe told the BBC of Cycling U.K."
Raw numbers are perhaps not enough to get across just how bad the centralisation of the Fediverse by mastodon.social is. I've done a chart to show this visually, hopefully this will make the situation clearer?
Given the circumstances, getting people to join other servers except mastodon.social would greatly help decentralise the Fediverse. It's no longer big vs medium vs small, it's more like supergiant deathstar mastodon.social vs everyone else 😞
@mdiluz @jplebreton @lexfeathers Yeah, that's the behavior I seem to observe.
Related behavior, to demonstrate there is a difference: if you go to https://mastodon.social/@jplebreton on the web, the default "Posts" tab includes the message with which JP started this thread, but not his reply above. If you go to "Posts and replies", both appear. And that's true even though the reply https://mastodon.social/@jplebreton/115588707331529098 shows the "globe" icon indicating it's public.
@janxdevil @mhoye @irvingreid In the Unicode and UTF-8 we actually have, the longest encodings are 4 bytes. That's no temporary accident; it's a fundamental consequence of the design.
Specifically, the space of code points ends at U+10FFFF. That's because UTF-16 cannot represent anything bigger: only 10 bits' worth of space in the first 2^16 numbers was reserved for UTF-16 to use in each surrogate-half, so it can represent only 2^20 = 0x100000 values with surrogate pairs, plus the code points up to U+FFFF which get represented directly.
UTF-16 isn't going anywhere — it's how Java and JavaScript represent strings, baked deep into those languages' APIs. (Not to mention all the existing software now that knows how UTF-8 works and knows that the largest code point is U+10FFFF.) So code point space isn't getting any bigger.
And under the basic scheme of UTF-8, 4 bytes expresses 21 bits, so conveniently just slightly more than enough to cover that space.
@janxdevil @mhoye @irvingreid In the Unicode and UTF-8 we actually have, the longest encodings are 4 bytes. That's no temporary accident; it's a fundamental consequence of the design.
Specifically, the space of code points ends at U+10FFFF. That's because UTF-16 cannot represent anything bigger: only 10 bits' worth of space in the first 2^16 numbers was reserved for UTF-16 to use in each surrogate-half, so it can represent only 2^20 = 0x100000 values with surrogate pairs, plus the code points up to U+FFFF which get represented directly.
UTF-16 isn't going anywhere — it's how Java and JavaScript represent strings, baked deep into those languages' APIs. (Not to mention all the existing software now that knows how UTF-8 works and knows that the largest code point is U+10FFFF.) So code point space isn't getting any bigger.
And under the basic scheme of UTF-8, 4 bytes expresses 21 bits, so conveniently just slightly more than enough to cover that space.
Alon
Jasper Corydon
Irving Reid
Race Bannon
Fedi.Garden