The engagement metrics have got to be better with Home. It makes sense, since the whole idea of an algorithmic timeline is to take the stuff the company knows is getting a reaction and feed it to more people.
Personally I think this is a great example of how engagement metrics mislead companies. The numbers go up, but it fundamentally changes the nature of the experience.
When you're on a reverse-chron timeline you have a sense of people's lives in real time and can have conversations. You only see people you have chosen to follow, so you recognize people and get a sense of their personality. A lot of the content is fairly banal, which gives an (accurate) sense that life for most people is kind of calm and mildly interesting.
When you're on an algorithmic timeline, you're seeing content that is really likely to get a reaction--that usually means it has a lot of emotion tied up in it: outrage, anger, sadness, humor, etc.
But it's also content that is old and choked with replies. So you can't have a normal conversation around it; there are too many people acting too emotionally. And you're seeing content from more people you don't follow--Twitter's Home timeline brings in a higher percentage of tweets that your follows have liked or replied to.
So if your timeline is filled with this kind of content, it creates the (inaccurate) impression that life is full of really extreme moments and amped-up strangers!
I can almost always tell when Twitter has switched me back to home. Suddenly I realize I'm feeling more angry, and the numbers on most of the tweets are bigger. I also see more tweets from people I don't recognize. It's like opening my kitchen door and suddenly finding a loud restaurant with spicy food in my house.
Seems to me like they are going out of their way to force their "top tweets" algorithm on users.
I would bet "top tweets" improves engagement with the passive consumption audience while saving money by abandoning consistency and improving caching.
I'm sure if Twitter was a product which came out of Google's culture, it would've worked that way.
They should have ignored the A/B test results, but investors give them money not to!
To my great rage, some Twitter design poobah was energetically explaining how they could "prove" that was better. Actual people saying no, it was actually worse for them were of course ignored.
As much as I love Twitter, the extent to which users are the product, not the customer is obvious. At best, you're a dairy cow, producing valuable content. Maybe you're a beef cow, carved up and sold to the real customers, the advertisers. And maybe you just end up eaten the rats that are always in the shadows and frequently come out in hungry swarms.
OK. Won’t use the product then.
With lists, if I swipe out, it loses my place.
Small issue but it puts a hamper on me using lists.
really? One of my major gripes with twitter is I'll be about 20 (ajax loaded) pages down the timeline, click on some retweet or name, and lose my place. At best it takes me to the bottom of the first page. What browser/os are you using?
That's kind of against the point of CSS, but it's also a useful dev tool to make your code more readable. I agree that it can be annoying on the user side, but it makes sense that they made this change.
Also, most likely, these change every deploy (when the depedency manager builds the app).
I've run a Twitter ad for the app so at least someone there knows about it. I don't know if the app was a factor, but certainly hiding what they think of their users' output is in line with all the other sneaky things they do to their users.
[1] https://github.com/necolas/react-native-web/blob/master/docs...
Sounds just like Facebook now.
[1] I’m kinda glad Twitter doesn’t force me to have an account, no account makes it so much easier to not get sucked into the reply-rants.
Bookmark this URL: https://facebook.com/?sk=h_chr
Your newsfeed is now in "Most recent" mode.
There may be A/B testing going on though.
Yeah I'm not a fan of the "Home" view.
twitter.com##div[aria-label="Timeline: Trending now"]
twitter.com##aside[aria-label="Who to follow"]
twitter.com##a[aria-label="Search and explore"]
twitter.com##a[aria-label="Bookmarks"]
ublock origin also seems to block the ads just fine at least.
I would be a little surprised if this wasn't just caused by using a library like styled-components for managing CSS. I'm sure it could also be partly anti-adblock or anti-scraping, though.
Never forget: you are the product. On whatever is the ad industry equivalent of HN, there is probably someone gloating about how much they love that it does that.
I wonder also if it’s cheaper. Computationally I mean. You don’t need to worry about consistency if you don’t care about ordering. You can also just show the same few popular tweets to everyone without needing to compute an individual timeline for each person.
Is it to make the best tool for conversation, or is it to make money? If it's both, how important is each?
The previous design was very good about keeping whatever tweet you were scrolled to in focus when new tweets come in.
I'm not sure how widespread this issue it, but I have seen other people in my personal feed express frustration with this issue.
Non-deterministic class names have been popular for a while, especially in React; open up Chrome inspector on a Gmail tab. Sometimes they are used for minification. This might not be a (strictly) hostile change.
It annoys me too, I write custom CSS for any site I use a lot.
HTML class names would not affect accessibility.
Non-deterministic class names are quickly becoming the norm for a lot of good reasons.
It's not always wrong to do things that are self serving at the expense of our users. But we should never confuse ourselves into thinking of that as a feature. It's always important to keep in mind who we serve.
Is there some advantage other than removing these two features that dynamic CSS classes provide?
The best they could do to make it less obfuscated is to include the name with the hash. So the class name may look like 'fkljdkldjf9303-button'. However, the hash may change if the component changes for example, so still not super reliable from your point of view. Also, most users won't care, so they'd just be including extra page bloat to add the human readable names along-side the hash in production builds.
I'd argue that CSS modules and similar reduce many many bugs I've seen people make over the years and that it's a huge win in terms of developer speed and reduction of bugs. I highly doubt anyone made the decision explicitly to obfuscate the CSS...
https://github.com/css-modules/css-modules or similar does this because the CSS global scope is more and more painful as applications get bigger.
Unfortunately they've now ported one of the most offensive feature from mobile to desktop. The "Home" timeline, with tweets out of order. And the real kicker; you can still select "latest Tweets first" but then the app literally undoes your preference every week or two, forcing you back to their "Home" view. It's offensive.
Also a small thing, but the new desktop Twitter now has obfuscated CSS classes for everything. The names change frequently too, maybe at every deploy? Anyway it makes it a lot harder to modify the desktop HTML presentation with an extension or set of ad blocker rules.