I switched to using PWAs for social media apps for similar reasons the author outlines. A pleasant, but somewhat unintended consequence is that I just use them a lot less because the experience is pretty bad. It makes me a little sad because I’ve always believed in the PWA dream, but the reality is that they’re bad because companies certainly don’t want to make an experience that rivals the app they really want you to download.
Expected, but just leads to reinforcing the idea that PWAs won’t ever be as good when every one people try from someone with a popular app is so awful.
What's funny is that desktop versions of websites in a lot of cases are responsive, and work fine on small screen. BUT at the same time the mobile version is crappy and lacks some features (or just shows "download our app").
Recently I've set up Firefox on Android so that it always run in desktop mode. I needed to also change screen width in about:config, because otherwise everything is too small. But after this websites seem to work better.
quite likely that the site has a mobile "mode" and a small-screen mode (for desktop), each made by different teams. some mobile mode website is fine, but others suck. Where as the small-screen mode for desktop tend to be made by the same team/person as the main site (it's a css media query after all) - so it's likely to be more coherent.
Alternatively, you can download Firefox Nightly instead of regular.
"about:config" just works in Nightly. No fuss.
You can sideload extensions in Nightly, too, after you activate the developer options. I don't think they've added that to regular, as yet? At least not with as much flexibility.
Anyway, I'm gonna try this mobile desktop mode thing and see how it goes. Thank you to everyone!
Installing extensions from file is available on the release build as well, after enabling dev options.
I think the only difference is nightly allows installing unsigned extensions, which I don't personally have a need for (as getting a personal/non-published extension signed is very easy).
And you don’t realize that social media apps put cookies on other websites so they know you have been to another website and then start showing you ads based on your interests?
Apps can’t tell what you do in other unaffiliated apps nearly as easily at least now on iOS that there is no globally unique identifier that apps can use to track you.
Apps require you to sign in so they've got you immediately. They can share all your activity with whoever they want. Websites (many) do not require you to login (youtube, reddit, hacker news, etc....)
Apps also try to open all links into their own webview, a webview in which they can track all activity.
All privacy-respecting browsers block 3rd party cookies by default now, which prevents that kind of tracking. There's still other forms of fingerprinting they can use, but those can be used in apps as well.
But I guess apps can run web views that have access to all the same fingerprinting as a standalone browser, minus any ad-blocking plugins (on iOS at least)
With a browser, you have the ability to block cookies, block whole hosts/domains, alter DOM content, alter tracking URL's, and (often) disable low level features you don't like. With apps, not so much.
And still waiting for examples of how apps can track you better. If the server wants to track you by your originating IP, all of the client side blocking will do nothing
What is your definition of "track you" in this context?
If it's to pinpoint a unique device accessing a website even through VPNs and/or other IP changes, there are an untold number of ways that apps can track you better than a website.
Apps have access to many device-specific APIs in addition to all the web ones, and every additional bit of information used can be added to the mix to create an even more unique fingerprint of the specific device accessing a website.
For example with phones, an app (even if it's mostly just a webview) may now also have access to your phone model, phone number, maybe your contacts or GPS location, and many other things.
A website Can easily deduce your phone model based on the browser agent attribute which tells the operating system and the screen resolution with a fair degree of certainty, an app can’t get your phone number, it can get your GPS with your permission. But so can a web page with your permission. There is a standard JavaScript API for it. Contacts are also gated by permissions.
Some apps can/will detect that an OS-level VPN has been activated though, and refuse to work at all. Spectrum TV does this for example, as well as some banking and other types of apps.
Yes and sites can also fairly reliably detect when a user is coming from a well known IP address block belonging to a VPN or VPS provider. It’s a built in feature of I know at least AWS
Assuming you are using a "well known IP address block belonging to a VPN or VPS provider", yes, but it is also possible to setup VPNs/proxies outside of well-known IP blocks.
PWAs can be good, but for a lot of social media, they're only as good as their website experience. Many (most) companies seem to make their website intentionally slow and buggy, probably with the idea that users only need to use their web UI for a short while because they lost access to their apps or something.
For instance, I've installed Mastodon as a PWA and it performs great. Photoprism also works so well I haven't even bothered to look for an app.
The absolutely batshit insane part is that the 'native apps' are almost certainly created using web technologies which call the exact same APIs as the web app.
There's zero reason the web apps should be so slow.
You can't use Facebook Messenger on the web at all, unless you go to Facebook and switch to the desktop version. Then it's a simple matter of zooming in without accidentally clicking anything, using their fiddly interface to load up the conversation you're interested in, and get bounced around the screen as the input focus changes around.
Puzzled because I use Messenger web version all the time on the laptop. Works pretty normally. I don't use Facebook usually. Maybe if they detect a phone they refuse?
Usually they only break the mobile websites. They know they won't get people to install programs on their computers for "online things". Not least lots of people don't have admin rights in their computers, but also people are used to accessing "online things" via browsers on computers rather than an app per website.
But, they've managed to make "phones use apps, not browsers" a social norm that enough people accept tacitly, perhaps because near enough everyone has the ability to install apps on their own phones.
On iOS devices you can turn off the ability to allow apps to wake up on a one by one basis “background refresh”.
And if you are concerned with your privacy, it’s nonsensical to buy a phone run by an adtech company that only made the operating system in the first place to sell ads and collect your data
Pwa with permissions granted gives access to:
Location, create notification, phone state, phone #, IMEI, motion data
Mobile app with permissions gives access to
EVERYTHING a pwa gets PLUS, Contacts, sms, notification content, biometrics data, web browsing data, phone activity history, location history, camera access, microphone access, NFC access, near device history, nearby wifi listing, saved wifi networks, Bluetooth device ID, Bluetooth beacons nearby, some device settings, personal data access(photos/music)
Maybe on Android. Literally half of that shit isn’t accessible by iOS apps even with full permissions. This feels like you’re just throwing shit against the wall.
I was wondering if it's just me. I am using Brave on iOS with all the possible blockers enabled, so I'm not surprised when some website doesn't work well. Instagram literally freezes solid after 5-15s of being on the website, so I usually only quickly scan the top 2-3 posts in the feed. I only follow people I know personally, so this is usually enough to do once or twice a day and stay up to date. If I see a close friend posted a story I kinda want to see then it usually takes two or three hard closes of the browser to actually see it. Sucks, but sucks less than being mental gamed into doomscrolling every time I get an app notification.
By the stopwatch it takes 3x longer for me to upload a photo to the Instagram web app than it does to Mastodon. Facebook's blue website works pretty well but the Instagram site comes across like something that was vibe coded in a weekend or maybe a straw man that was made to prove SPAs are bad. Contrast that to the Mastodon application produced by a basically unfunded application that's fast and reliable.
Just hours ago I couldn't even copy-paste a description of a post I drafted in another app. Literally nothing happened when I tried to paste. No console errors, no feedback, nothing.
It was a bit of a longer one, but still far below Instagram's supposed character limit. The fact that they somehow broke copy-paste functionality really baffles me.
Surely at some point some team that writes this has to demo it and someone checks it. After however many years of it not working, surely that's strategic, not accidental.
It's such a pervasive pattern and somehow always in the direction: the app works better than the website. If there even is a website.
Sometimes it goes the other way, in fact enough it's a running gag that the banner that says "Download our app for a better experience" at sites like Reddit ought to have one of these
I don't know if big companies even know how to make web apps. Honestly. Which is extra insane to me because there's so much investment in web technologies. On my team at $BigTech there's like 1 or 2 people out of 30 people on our team that knows web, the rest are mobile. I'm a web guy but I refuse to touch our web-app because they butchered the tech stack and I don't have the energy to deal with that BS. We still have an mobile-web version distinct from the 'desktop' version because.... I don't know why, whoever wrote it never learned about responsive web design and we never bothered to move out of the stone ages because if people want to use the app on their phone, they should download the native app of course! And by "native" I mean we built our own half-baked framework so that we could cross-compile for Android and iOS.
Also I don't think these people know how capable PWAs are. There's very little you can't do in a web-app that you can do with a native app.
I have had a FOSS web app for learning arithmetic for quite a few years. I occasionally review it, and make changes. Each year Chrome and Safari both nip at the edges of what allows a PWA to be OK. No one really cares until one has to write documentation helping folks install the PWA and avoid issues that did not affect the PWA a few years ago. I mean really, are Tim and Sundar really that afraid ?? I guess so. They have dozens of millions on the line. Capitalism... gotta luv it.
Hmm, I'm making a site and I planned on using a PWA for the app experience instead of a native app. Am I setting up for a bad time? I'm not too worried about the installation hurdle, my potential early adopters are motivated and smart.
If you're using React, I'd recommend using Silk (silkhq.com) to create native-like bottom sheets, pages, sidebar, etc.
Most animations, including the swipe, are hardware-accelerated, and it deals with a lot of common issues you encounter on the mobile web (body scrolling, on-screen keyboard, etc).
Nearly every time we try to fix this problem with money it fails. The problem is not money. All else being equal there is little to no correlation between spend and outcome. Money get's touted by schools and politicatians as a way of pretending to care but not actually do any of the work to improve outcomes.
What does tend to correlate with money and also correlates with outcomes is parental involvement. Solving that problem requires societal and economic change in a district though not giving the school more money.
I would argue spending less money would actually improve things.
Ultimately it’s a culture problem. America’s attitude toward problems is nothing if not “throw more money at the problem and hope it gets better.” See also, healthcare, military spending, college sports, etc.
cost per student is higher for high school students. So if you take an average across all grades for public schools and then compare that to specific cost per grade at private schools, of course private schools are going to look relatively cheaper for younger students.
I think your reasoning is flawed, but fine...if the goal is to try and have the cheapest possible one room school house. That $200k gets eaten up pretty quick by things like security, janitorial, building maintenance, support staff like principals, librarians, guidance counselors etc etc. If you’re meaning to include total cost for the full time employees (the teachers) in the list, then the salaries are a lot less attractive once you’re done covering benefits, etc.
I've got multiple kids, so I'll admit I think about schools here a lot. The absolute cheapest private schools I've seen in San Francisco are subsidized by religious institutions. The tuition for those schools per child is roughly $28k. Non religious private schools usually start in the $40k range and can easily get into the $50s and well beyond.
My point is that it's hard to point at some issue of inefficient public bureaucracy, because clearly private institutions aren't able to do it any cheaper. I would also argue they wouldn't try, because their goal is a good education, or at least better than the public alternative (that only spends $28k per kid).
"I think your reasoning is flawed, but fine...if the goal is to try and have the cheapest possible one room school house."
I was generous in my estimate for each of the line items. I chose a one room school house as an example because it's easy to grok, and anything larger would be cheaper due to economies of scale.
"I've got multiple kids, so I'll admit I think about schools here a lot."
Although I have only one child (in 4th grade), I think about schools a lot, too.
"The absolute cheapest private schools I've seen in San Francisco are subsidized by religious institutions. The tuition for those schools per child is roughly $28k."
This $28k number is false. Most parochial schools charge about $12k. Here is a breakdown by grade level of the number of parochial schools in SF that serve that grade level, and the median tuition among those schools for that grade:
# Median sticker price
Pre-K 7 $16,610
K 29 $11,530
1 29 $11,530
2 29 $11,175
3 29 $11,175
4 29 $11,175
5 29 $11,175
6 30 $11,519
7 30 $11,519
8 30 $11,519
9 4 $31,725
10 4 $31,725
11 4 $31,725
12 4 $31,725
"Non religious private schools usually start in the $40k range and can easily get into the $50s and well beyond."
This 'usually start in the $40k range' is also false. For each of the grades K-5, 33-39% of non-parochial schools in SF charge less than $40k. For each of the grades 6-8, 30% of non-parochial schools in SF charge less than $40k.
"because clearly private institutions aren't able to do it any cheaper"
Non-parochial private schools don't typically price based on cost. The schools that have high demand (due to parents and student population) can charge more. So they don't need to manage their costs tightly. And they can spend lots of money on marketing.
Moreover, not all students pay sticker price. So looking at the sticker prices (which I've listed above) may give an inflated view of total income.
"because their goal is a good education"
Their goal is happy customers (parents). Different schools achieve this in different ways. Some parents choose a school not based on the expected quality of education but based on the expected networking opportunities for themselves and for their child.
I was generous in my estimate for each of the line items. I chose a one room school house as an example because it's easy to grok, and anything larger would be cheaper due to economies of scale.
I would argue that economies of scale don't apply to education in the same way they apply to other businesses at large. Sure, you theoretically get the benefits of scale with central organization, buildings, centralized services, etc, but once you get to the classrooms themselves most of the cost simply scales linearly with the number of students.
This $28k number is false. Most parochial schools charge about $12k.
I'm not sure what we're talking about here anymore. You're using K-8 as the dominating factor for this gotcha a few times in this thread. There are more K-8 parochial schools, yes. "Most parochial schools charge about $12k" is true, unless you're talking about high school. Exactly 1 parochial school is less than $30k (SF Christian, at $16k). From there (limited to religious schools):
- Sacred Heart ($31k)
- Archbishop Riordan ($32k)
- Saint Ignatius ($34.6)
- Sacred Heart ($60k)
- Jewish Community School ($65k)
I might have missed some in here since I'm going by names, but given that SF Christian is the cheapest private high school on SF Chronicle's list[1] I don't think that matters for my point.
You started this thread with average cost per student across all SF public school students, which includes special needs, high school, etc, but move to median prices for debate, and structure most of your argument around the cheapest schools (K-8). Mea culpa on my end, though: you are correct that when I was saying "cheapest I've seen," there was an unfair modifier of "cheapest schools on my personal spreadsheet" which is limited to schools within a reasonable commute and that we'd be willing to send our kids to. You're absolutely correct that there are cheaper parochial schools available as long as you only need K-8.
Using averages for private schools, which feels more applicable to your starting premise, private schools in SF average $27k, $28k, and $52k, for elementary, middle, and high school (again, referencing SF Chronicle's data). I still feel comfortable with my original premise that averaging $28k per student across all of SFUSD students is not an absurd number.
So looking at the sticker prices (which I've listed above) may give an inflated view of total income.
Sure, that's fair! But we're not talking about income, we're talking about average cost per kid. We can't actually know the details under the hood, but again, those schools specifically in your list are usually subsidized by a larger religious organization, so the sticker price doesn't truly reflect that cost anyway.
You started this thread with average cost per student across all SF public school students, which includes special needs, high school, etc, but move to median prices for debate
The reason for this is simple and not nefarious:
- I don't have access to data that would allow me to apportion total SFUSD costs to individual school types
- When considering schools with vastly different prices (and different scales), the median is a much more informative measure than the mean (which could be skewed by an unusually expensive or inexpensive school with a tiny student population).
Another reason for using median is that I was responding to your comments which talked about general price levels ('tuition for those schools is roughly', 'usually start in the $40k range'). You were not talking about averages, but typical prices or minimum (starting) prices. The mean prices have no bearing on the truth or falsity of those claims.
Using averages for private schools, which feels more applicable to your starting premise, private high schools in SF average $27k, $28k, and $52k, for elementary, middle, and high school
If we look only at non-parochial schools, the means are even higher (e.g. $39k for 5th grade, $41k for 8th grade, $59k for 12th grade).
those schools specifically in your list are subsidized by a larger religious organization, so the sticker price doesn't truly reflect that cost anyway
Sorry, I should have been clearer. I meant when we look at the sticker price for non-parochial schools, we should assume their average revenue per student is less than the sticker price, and the average cost per student is less than or equal to the average revenue per student.
I still feel comfortable with my original premise that averaging $28k per student across all of SFUSD students is not an absurd number.
My original point in this subthread was that SFUSD is NOT underfunded. Do you believe it IS underfunded?
which is limited to schools within a reasonable commute and that we'd be willing to send our kids to
If we limit the discussion to only those schools we'd be willing to send our kids to, then that would rule out almost all SFUSD schools, which kind of defeats the point of the discussion!
BTW In case you want to see the SF Chronicle data in a form that's more personalized (showing the schools nearest to you first, filterable by grade levels and price and type), I made a tool to do that: https://tools.encona.com/schoolfinder
My original point in this subthread was that SFUSD is NOT underfunded. Do you believe it IS underfunded?
Your original point was not that it's "not underfunded," it was that it's overfunded (and substantially so, based on other comments). Your top(ish) comment on this thread to the $28k per student average:
I'm saying it's a lot.
My only argument here is that I don't think $28k is unreasonable, particularly when viewed against the cost of private alternatives.
then that would rule out almost all SFUSD schools, which kind of defeats the point of the discussion!
We go to our attendance area SFUSD school and love it. There are plenty of SFUSD and private schools that would not be on our list, be it for academic reasons or logistical.
I made a tool to do that
Cool, I dig it! Annoying, unsolicited feature request would be to allow addresses or zip codes rather than requiring geolocation :)
My only argument here is that I don't think $28k is unreasonable, particularly when viewed against the cost of private alternatives.
OK, so we agree SFUSD is not underfunded?
We go to our attendance area SFUSD school and love it.
That's great! At my attendance area school, two thirds of students are behind grade level in math, and there's no opportunity to be grouped with kids in other grades.
Annoying, unsolicited feature request would be to allow addresses or zip codes rather than requiring geolocation :)
If this is for privacy, don't worry, it's all front end code and your location isn't sent to the server. (You can check the network tab or just look at the code.)
Yes, it was extremely clear that your position is that it’s overfunded.
OK, so we agree SFUSD is not underfunded?
No, we don’t, I just wasn’t trying to make that point. There’s absolutely debates to be had about SFUSD, including how they spend their budget (personally I would make big cuts at the central office and redistribute to the schools), but the thought that $28k/kid is too much in SF just isn’t grounded in reality.
two thirds of students are behind grade level in math, and there's no opportunity to be grouped with kids in other grades.
This type of broad statement is true, but also obscures the realities of the student population SFUSD is mandated to serve. That high level number includes special education, non-English speakers, etc. Generally speaking, the data for kids with a similar socioeconomic background to the one I suspect your kids have are doing fine in SFUSD, particularly in K-8.
For example, a low key popular school with “bad scores,” Flynn[1]. ~30% of the student population met or exceeded the standard for math. That number jumps to 65.4% for kids with college-educated parents, and 81.3% for grad school. Race is an unfortunate proxy here, but it’s 70.8% for white students.
Not trying to convince you to send your kid to public school, of course, just calling out that there’s nuance required when comparing outcomes and what can work for families.
personally I would make big cuts at the central office and redistribute to the schools
Amen.
That number jumps to 65.4% for kids with college-educated parents, and 81.3% for grad school.
Do you think those kids are learning primarily due to their experience at school, or because their parents teach them? Anecdotally, whenever I've walked past Kumon centers during the weekend they seem busy, and when I was driving the other day I noticed Russian School of Math has added a new location.
I think this is overall fair/accurate in terms of crash fatigue, but crypto has crashed by 85+% multiple times in the last decade. I'm not sure that's the best proof point in this list of things that don't actually happen.
I am not fan of crypto and dont hold any, not by reference (etfs etc) either, however, all the crypto fans that told me to buy and hodl did win the jackpot so far. But I agree, not the best example for not crashing: I was just, in the particular case of crypto, referring to that many believe(d) it would drop to 0 and dissappear: in case of bitcoin: did not come close in the past 10 years.
> Gleam takes inspiration from Elm, so if that's your thing and something you'd consider using, probably go with that rather than Phoenix. Again, if your team can handle it.
I’m not sure I’m following. Gleam is inspired by Elm in a lot of ways, and the major draw over Elixir would be the Elm-inspired type system, but Gleam is still a language not a web framework (which Elm is not either, since it only targets clients afaik).
I agree that Oracle has been a perfectly fine trademark holder in all of these regards in that they are entirely irrelevant to JavaScript and have been for as long as I can remember.
The point here is that them not doing those things would be codified. Deno's not trying to take the trademark from them for themselves, they're trying to get the USPTO to agree that JavaScript is a generic term at this point and unable to be trademarked or owned by any one entity.
I'm not sure how that changes any of the bullet points you've got above. It's nice that points 4 and 5 would become completely impossible and not just improbable because the trademark owner currently doesn't care enough to do it.
Not a lot of info on the page about the process, etc, but this is also called "datamoshing." If you're curious, there's a great talk from Demuxed '21 on some of the details: https://www.youtube.com/watch?v=Qtia43DGSrY
It's a product whose largest cohort is designers or design-minded people. Them focusing on that as part of the product itself feels like a perfectly good use of their time.
No one said they were, I don't even think this font is available for use outside of are.na's product. This is about craft.
I think they said it pretty well themselves:
With Areal, Dinamo designed an updated version of Arial especially suited for Are.na, but which still honors the original. Stem thicknesses were streamlined, more characters added (), a monospace version drawn, dark mode functionality optimized. You probably wouldn’t have noticed these changes if you hadn’t read this statement. It’s possible you still won’t. But to us (Are.na and Dinamo) Areal’s existence is satisfying in the way that rewriting an entire front-end is satisfying. As stated in this text block from 5 years ago, “the reason you would create something is because you love it enough to see it exist.”
Just curious, if you’re already in Elixir and using Postgres, why not use Oban[1]? It’s my absolute favorite background job library, and the thing I often miss most when working in other ecosystems.
Oban is awesome, but I really like the ideas around pgmq and most of the logic living in the database. And the idea of "flows" design from pgflow for multi-step processes (which Elixir is naturally a great match for)
Oban is so good! My startup has an extensive graph of background jobs all managed by Oban, and it's just rock solid, simple to use and gets out of the way.
I recently joined a local independent video rental store and it's so, so good. My partner looked at me like I was crazy when I told her, but she was a convert after one trip to pick out a movie in person.
Something about browsing in person is just so much more enjoyable than flipping between 9 services. Having a cinephile right there behind the desk that wants to nerd about movies and help pick something out is awesome. It's not a big store, but they've got thousands of movies in their catalog, which is (apparently) way bigger than any of the streaming services.
This doesn't solve your problem, but for the folks that are near the few remaining physical rental stores: consider supporting them, because they're great.
Most of the collection is DVD and/or Blu-ray, but he's got some VHS tapes and video games across a few platforms. When he was giving me the spiel about joining he was explicit that he doesn't have any Laserdisc or Betamax, though.
Expected, but just leads to reinforcing the idea that PWAs won’t ever be as good when every one people try from someone with a popular app is so awful.