tumor (noun) An abnormal growth of tissue resulting from uncontrolled, progressive multiplication of cells and serving no physiological function; a neoplasm
Interesting that onkernel.com intentionally animates and slows down the loading of the web interface, making it harder to scroll and scan the site. Irony or good design?
In my experience the cars are highly regulated and very predictable on the main roads. Marked out lanes, low speed limits, many red lights.
But pedestrians and bikes are everywhere, and they don't always obey the rules. I wonder how well collision detection handles sharing lanes with bikes without suddenly braking in the middle of the road.
I was in a Waymo last week in San Francisco that had a bicyclist cut in front of us, no hard breaking or jerk just a gradual decrease in speed. I would've hit the break relatively fast myself but Waymo handled it without issue in a very smooth way.
I don't know how it'll handle significantly more pedestrians, but I assume that they're confident enough in the models and have run enough simulations to expand to Tokyo.
I bike in SF regularly and I actually cut in front of Waymos intentionally since I know they will slow down gracefully vs regular drivers who might not see me for a number of reasons or trigger road rage. Several months ago a Waymo would over correct but now they're so good at anticipating where I'm heading.
In San Francisco, pedestrians and bikes can appear anywhere around your car, at any speed in any direction. Including wearing all black, no light and taking 2 steps forward then turning around and taking 2 more steps forward (repeat but vary everything randomly). Or doing lasso circles with something heavy at the end of a rope in the middle of the street (last weekend). Just because this is not 100% of the time is not a license to run over ANY of them.
This is a kind of environment that human drivers are NOT made for. All the more not while clicking around on their Uber app or changing the music track or trying to read street signs or understand a Nissan dashboard map. In San Francisco, computers with multiple sensors have a gross advantage over humans.
I live in rural Japan, and in my experience, drivers are also crazy:
* They stop on the side of roads and streets whenever they feel like it without worrying about blocking the traffic.
* They don't turn the lights on no matter how bad the weather is. Super fun to be in the middle of a snow storm and people are driving white cars with the lights off.
* Whenever someone wants to turn right at an intersection, the cars behind will pass it on the left, without worrying if someone is coming the opposite direction, which is really dangerous. I am not sure about the Japanese law, but in my home country (Spain) that is highly illegal.
* Many people watch TV or anime while driving. I even saw one guy reading a book while driving, somehow holding it open over the driving wheel.
Add to this the awful state of most streets and roads, and I can see more accidents here in one year that I saw in 33 years in Spain.
How would passing on the left would risk a collision with a car coming from an opposite direction? The opposite lane is on the right hand side (in Japan)?
I assume they mean: Given turning/waiting car A, passing car B that is going straight and car C coming from right going straight. Car B likely has right of the way as they are going straight and they are on the left from viewpoint of car C, but car C may not see behind car A (like due to it being truck for example). Both B & C might think they have right of the way. B due to fact they should normally only need to avoid people coming from left and he can see there is no one there, C due to fact he entered intersection when he only saw A who is waiting to turn. I don't know Japan law so I don't know actually has right of the way here (in Finland A would actually have right of the way in mirrored situation as you always need to yield those coming from right in equal intersection, even if they are turning left).
Yes, it is a question of visibility. You have two cars coming from opposite directions, both want to turn right (Japan drives on the left), and both think they can do it safely, but all of sudden another car pops from behind.
I have seen many near accidents due to this. That's why such a maneuver is forbidden in many countries.
True. It doesn't help that most streets don't have sidewalks and you are forced to walk on the road. And even when there are sidewalks, people ignore them...
Oh, I mean — that's a thing even in the densest parts of Tokyo, people (and not just tourists!) just have no concept of moving aside to not block the flow, frequently walk like 3/4-abreast and block the entire width of the stairs/sidewalk ;P
Kinda surprising considering how everyone can do it perfectly on the escalator. To the point of ignoring the signs that tell them to stand two abreast.
Does anyone know if there is a tool or platform that makes it easy to make/publish this type of "leave footnotes on an article"? The UX here is nice and simple, and I think there is utility in sharing feedback this way (vs. blockquotes in email, or Google Docs comments, etc).
Hypothesis is a really interesting platform, but in my mind it's kind of too community-focused. Hypothesis's restricted groups (https://web.hypothes.is/help/annotating-with-groups/) are closer to what I think people would want, but to get them you'll have to self-host.
Not trying to throw shade at them, but I feel like their push for annotations is hampered a little bit by how much their specific implementation of accounts and permissions feels like yet another social network, with all of the negatives that entails. There are other issues as well, but that's the big one that kept me using the platform.
That being said, there is also an Open standard for this stuff that I remember at the time it came out being pretty excited about. But I haven't seen much if any adoption of it, so it again makes me wonder if there's something wrong with it or if it's just that nobody has made anything super-attractive yet to take advantage of it.
Which standard is that? ePubs, for example, use canonical fragment identifiers, which could also work on HTML. I think hypothes.is uses them too but adds fuzzy matching fields so that they can match up after document changes. At least from what I remember
I'm working on an annotation project myself, though not really related to what the parent commenter is looking for. A generic solution to the highlight/annotation problem is quite difficult, especially if you allow for document changes.
I read through the docs when they first came out but I don't remember all of the details anymore. Fragment identifiers are part of it though. I've gone back and forth on how much I like them, allowing for document changes is always going to make stuff really complicated, fuzzy/fragment matching is about the best you can do I think without the cooperation of the source you're annotating.
It feels like annotations probably should be tied to specific document versions, but I also get the reasoning why they're not in this proposal. Unless you're self-hosting everything or forcing everything to go through the Internet Archive... it's just kind of difficult.
I've been curious about trying to get annotations to work before using something like Matrix to handle accounts/groups, but not curious enough to actually try to build a working example.
There are people who are adamant Bill Gates is trying to depopulate the Earth through COVID-19 or something. They're wrong, but people do believe that.
Pinboard doesn't have near the traffic or complexity as goodreads. It's basically fetching text out of a database with a user-id lookup. No multiple databases for books, user profile, reviews, etc.
