K‮ly‬e @khird@qoto.org

@design_RG

You'll notice that the odd post does still come through - there was one today, in fact, which is what prompted me to investigate since I had previously thought the bot was deactivated when I stopped seeing its posts regularly. I also retain my status as a follower of these accounts, which is normally not possible if a block is in place. This makes me think we're not subject to such a restriction.

@dragfyre

Hey @dragfyre, for some reason, I only see a small subset of @bahai's and @bahai_wiki's posts from my account, but many more (I presume all of them) when I visit the page directly in a browser. The problem seems to have started around 4 March 2020. Do you have any idea what might be happening here?

@coldwave Kind of. It reads like someone asked, "Why now?" and the author spent most of her time considering the "now" instead of the "why", going on about George Floyd and recent events. There's also a lot on the lack of diversity among media employees, which would be important if they were changing their hiring policy but is kind of tangential to the style change.

In particular, she mentions the "proper noun" principle that underpinned their old style rule, but the thought is left unfinished. There's no indication what principle of English now supersedes it to justify the new rule.

It's important to note that this is the CBC, not a private broadcaster which is supposed to change according to shifting preferences and market pressure. One of its functions is to serve as a language model for people learning, improving, or maintaining proficiency in Canadian English, so it's supposed to observe the principles of the language pretty zealously.

Here's a kind of non-explanation ("we saw a lot of other people doing it and the activists said it's good"): https://cbc.ca/1.5626669

@freemo Especially in the first example, the upper diagonal looks like a continuation of the L the same way the horizontal is a continuation of the K. So I read those two letters as LH or TH more easily than as LK.

@realcaseyrollins I'm not on the west coast, but it worked great here. The plywood is off the windows, the people protesting racism and the people protesting the mask order stand out of the way and shout at passersby, and life goes on.

The first few nights just saw the police outnumbered. Every arrest took two or more cops away from crowd control at a time when they needed every man they could get. Then policy changed, and they decided to just march down the street with pepper spray and tear gas if things got rowdy. Over the course of a few days, the story went from the crowd re-forming after being cleared, to the crowd dispersing after being cleared, to the crowd behaving and no longer getting cleared at all.

@realcaseyrollins ah, but that's a different proposition. "Stopping riots" != "Feds arresting rioters".

Alice, for instance, opposes *arrests* because she sees them as an ineffective use of resources. In her view, the police can more effectively stop riots if they forget about making arrests and just chase rioters from the streets, so the feds are allowing the riots to continue longer than necessary by pursuing bad strategy.

In compsci we talk about separating mechanism from policy. The same idea applies here - the question of mechanism (should the feds be making arrests) should not constrain the choice of policy (are we fine with rioting).

https://en.m.wikipedia.org/wiki/Separation_of_mechanism_and_policy

@realcaseyrollins I don't think that follows. Consider:

Alice isn't "fine with rioting". But a given number of police can disperse many more people than they can detain, so she wants the feds to stop making arrests and just break the riots up instead.

Bob isn't "fine with rioting". But he feels strongly that the feds need to mind their own business and leave enforcement to the municipal or state police.

Carol is "fine with rioting". She has a jury summons and hopes to make a political statement by refusing to convict a rioter (jury nullification). She wants the police to arrest as many as possible to maximise her chances of getting on a rioter's trial.

@realcaseyrollins

Well the RF part of TERF stands for "radical feminist" and I'd be mildly surprised if anyone identified you as such. Not all transkeptics are TERFS.

@realcaseyrollins if QOTO had single sign-on, I think that'd really be all that's needed.

I've recently noticed a number of media outlets capitalising "Black" when referring to a person's race. Here's an example, from a CBC story:

"Officer Garrett Rolfe fatally shot Brooks in the back after Brooks fired a Taser in his direction while running away after a struggle with officers outside a Wendy's restaurant on June 12. Rolfe, 27, is white. Brooks, 27, was Black."

Here's another, from FiveThirtyEight:

"Some political science research shows that Black people vote at higher rates when a Black candidate is on the ballot, although that finding is somewhat contested, and that research is about voting for a Black candidate, not a white candidate with a Black running mate."

This is an error in English because "black" doesn't derive from a proper noun (unlike "African" which does). However, these are both fairly respected news organisations - the CBC in particular is supposed to serve as something of a language model - and the pattern is consistent enough it's clearly intentional.

Would anyone care to offer any insight as to why the writers or editors are deviating from conventional English practice here, and selectively so (note that in both examples, "white" is not capitalised)? I'm specifically interested in the reasoning from their perspective; if all you have to contribute is "the media is full of liberals" then you're not going to advance the conversation much.

@HighFunctioningSunflower H is just the letter I chose because its Persian equivalent ه has four very distinct shapes, so it's easy to see how it changes based on context. Since Persian is written right-to-left, the initial letter of a word appears on the right, while the initial letter of a Latin-alphabet word appears on the left. But in both cases you type the word first-to-last and the computer arranges the letters left-to-right or right-to-left as appropriate for the script you're using.

@HighFunctioningSunflower I assume when you ask "how it works" you mean "how the different character shapes are selected"? The keyboard just enters letters as in any other language, but it's up to the font rendering engine to decide the appropriate glyph to represent each character based on context. For example, here's the Persian letter H, and a small line (kas͟hída) that represents any letter to which the H might be connected.

The initial form (هـ) is used when there is a letter to the left, but not to the right.
The medial form (ـهـ) is used when there are letters to the left and right.
The final form (ـه) is used when there is a letter to the right but not the left.
The isolated form (ه) is used when there are no adjacent letters.

In all of those lines, I typed the same character on the keyboard for H, and the context (i.e. position of the adjoining line) tells your device which shape to use. The Latin alphabet equivalent would be typing (H_), (_H_), (_H), & (H) and seeing the letter H assume a different shape in each instance, even though you typed it in with the same key all four times.

@realcaseyrollins@mstdn.social
Additional discussion of this feature is at https://github.com/tootsuite/mastodon/issues/8649

@zleap @design_RG @globcoco @freemo

QOTO differs from vanilla Mastodon in its post length limit. While we enjoy 64kchar of space, users on the flagship instance get only five hundred. So to post a longer message, a vanilla user will break it into chunks ≤500 characters each and post each chunk as a reply to the previous one. That "show thread" link pops up to make it clear the post is possibly dependent on another for context.

If you reply to a message from someone else, your reply isn't published to the local timeline (although it may show up in your followers' home timelines). Self-replies, however, do show up in the local timeline unless prevented by privacy settings. So the need to warn about hidden context is much less for cross-replies than for self-replies).

https://github.com/tootsuite/mastodon/pull/9228

@freemo I'd suggest ditching the phone and envelope glyphs. If your recipient might run OCR on your correspondence (many do, so as to be able to grep it later on), those sorts of things could result in your contact info getting garbled.

@vibhi Odd. Well, turning incognito mode off fixes it - although why that should be the case doesn't make a whole lot of sense. Does your CSS delivery scheme rely on being able to save cookies?

@vibhi still the same behaviour - here's what I see.

@vibhi @freemo

"Something went wrong while trying to load the full version of this site. Try hard-refreshing this page to fix the error."

Not a great first impression.