Somehow around the 2010s we all decided that everything in the web had to become "reactive" and "asynchronous" - which is a fancy way of saying things can theoretically happen at any time but realistically if you try to make it happen in ways that don't resemble the previous serial approach you get weird race conditions - and instead of making sure this was implemented in HTML now we have to write another web browser on top of our web browser in javascript, using a thing called shadow DOM.
Also somehow now you have to understand how the internet works at protocol level unless you want way worse performance than that page written in Dreamweaver 20 years ago.
But this is fine because this is the way big companies run things, which we all know they always make the correct decisions like giving their AI full access to their login and password recovery process.
What a brilliant piece of writing. I remember almost every single step—safe for actually getting angry emails. Maybe I ended up being the one writing them.
What a glorious time period.
Interestingly, it would’ve been impossible to share this writing with as many people as the author did by publishing it on mastodon and then it ended up on HN in 1998. The network effects are real.
I enjoyed this. But reading the claim that the iPhone was bad compared to other phones of the day makes me question it all. That's so incredibly backwards. It _was_ a much better internet in your pocket. If you couldn't see that, it says something about you, not phones.
Very little of this rings true for me, but that part worst of all.
The mobile web pre-iPhone was terrible. Nobody used it, nobody wanted to use it, and nobody wanted to build it. At best there was a shitty cut back version on the `m` subdomain. WAP/WML were terrible and didn’t give you anything close to the real web, and XHTML Basic was still-born.
The iPhone came along with its “desktop class web browser” and it genuinely worked. Steve Jobs got on stage and told everybody if they wanted to build apps for the iPhone, they should be web apps. Then he told everybody Flash was terrible – which it was – and that we should all use open standards instead.
Practically overnight, everybody commissioning websites wanted them to be “iPhone-compatible”. They did not ask for mobile sites – they specifically asked for them to be iPhone-compatible.
And because WebKit was open-source (thanks to it being based upon KHTML), all the other phone vendors took the code and ran with it, including Android.
This is why I say there is no single organisation that has done more to push the mobile web forward than Apple. The difference in attitudes and capability towards the mobile web changed practically overnight, and it’s directly attributable to Apple’s intentional actions to develop and promote the mobile web.
The original iPhone had a 3.5” screen, Wi-Fi, and EDGE. The iPhone 3G came a year later, with, shockingly, 3G. So yeah, that web browser that actually worked did actually work. Sure, it was slow compared with today’s speeds, but giving people what they want slowly matters far more than giving people what they don’t want marginally faster. Remember most people were on dial-up at home back in those days so the web was hardly fast on desktop either.
WAP might’ve been able to convey information, but so could Lynx in an 80×24 terminal – people want more than that. WAP sites were never popular (aside from some Japanese platform, IIRC) and I don’t think the average web developer had even heard of WAP or WML at the height of their popularity.
Symbian phones were also based on WebKit, however it wasn't nearly as usable due to many factors, mostly because it emulated mouse pointer (and you controlled it via the joystick) and didn't provide the pinch-to-zoom fluidity of the iPhone Safari. Also iPhone supported most of the major fonts used on the web, had font smoothing, etc. I've used both Symbian WebKit and Safari on first iPhone, and they aren't even close even though they're based on the same underlying browser tech.
I’m not saying the iPhone was the first phone with a web browser, I’m saying that it was the first phone with a web browser that was good enough for average people to want to use it.
Clients were lining up en masse for iPhone-compatible websites, but none of them ever cared about S60 compatibility.
The original iPhone was genuinely terrible. A 2006 Nokia could surf the web on the go and tell you where you are. The iPhone could do neither, since Apple did not include a 3G modem or GPS. It also did not have any apps, and one of the key features highlighted by Steve Jobs was voicemail. The 3G one year later was the first truly usable iPhone.
I seem to recall I had a windows phone at the time, with a full keyboard. I could use OS maps on the thing, and although it didn't have GPS, it could get my rough location by tower.
My mate had an iPhone, and it had an app where you could pretend to drink a pint of beer.
That might prove the novelty of the App Store back then. Also, let’s not forget that its’ screen was better than most other smartphones of its time. It was a rather limited phone, but not bad exactly.
My Nokia 5800 express could connect to a VPN and run a bridged internet connection over Bluetooth in 2008. I clearly remember using it in Switzerland around 2010 to pay for a tramway using a QR code.
Sure, UI was way worse compared to UIKit but in term of features Symbian phones were light-years beyond the iPhone at least until the iPhone 4.
Another way of looking at it is that it made the "real" web shit.
The mobile-web world before the iPhone was one where mobile devices were second class citizens; desktop "real" websites first, and scaled down versions of those sites second (for mobile devices). The information itself was readily available to both, even if the presentation was lacking on the latter.
Jobs knew the only way to win was to not play the same game, because the open web is an even playing field. If you control a new platform on the other hand, you can't lose. So here we are, with locked down, dumbed down toys determining the standard.
Fuck, you missed it being bad at being a phone. It launched on the network that had such bad coverage they needed Apple's hottest product ever to get people to switch. They had to put a cell repeater (or was it a miniature tower?) in the presentation center so the call would go smoothly, and still had to rig the cellular icon to always show full bars.
Judged purely as a "Let's give Ted a call"-phone, it was fucking bad compared to the competiton. They killed it in other areas, don't get me wrong, but not at being a phone.
I read the beginning of the post and immediately tabbed back to start writing a rant about standing up a website being fine, but the real loss of web functionality was Flash, glad I kept reading. I'm quite good with CSS and doing tricks with SVGs, but constantly run into things I want to do in 2D web I find to be complex enough and time consuming enough I don't even bother, while I would have been able to do it in Flash as a 13 year old in a few minutes. The modern web is a prison built out of <div>s, the tricks you see for "amazing" websites with obnoxious scrolljacking/parallax don't hide it.
> Is it easier to build or repair a radio now than it was when they were first sold?
"Build"[1], yes. "Repair", not really. A lot of things (like radio) can be an isolated module you just drop in.
Right now a bright HS student can produce a clock radio by plugging together an arduino, an I2C FM receiver, I2C display, an I2C audio amp and a speaker. Pop it all into an enclosure and you have a real clock radio.
Compare to building a clock radio back in 1990 - you'll have to understand how to wire up an op-amp, the principles behind op-amps, how to drive the readout from a clock, etc.
-------------------
[1] By "build" I assume you mean "assemble from parts".
But this is about building radios, which is easier (probably?) now. You can drag and drop an FM radio in pure software (well, with a USB dongle). Or if you insist on hardware, use free tools and cheap PCB services to prototype your radio, grab the parts from Digikey and follow a Youtube tutorial to solder it all together.
3D printers are mostly still repairable and far more reliable and usable than a few years ago when the majority of the hobby wasn’t making stuff, it was tuning the printer to work reliably. Bambu and California may be signalling that the enshittification inflection point is near.
Not quite the same level, but home/hobby electronics with tiny microcontrollers is more accessible than ever before thanks to the availability of cheap ESP8266/32 clones.
And there are some obvious individual counter-examples - Framework computers, or repairable blenders[0]… but you’ve got to pay a premium for the privilege.
But broadly you’re totally right - in the modern world, by the time something becomes a mainstream product aimed at general consumers, there’s a profit to be made and it’s likely on a downwards path.
My experience was almost identical. Don’t forget all the time spent balancing image quality with load speed to get that perfect blend of shit quality and slow download.
> Making a website should be the easiest fucking thing in the world by now. There should be a program that you can use to spit out a website as easy as Word spits out words on paper
squarespace?
> and we - us fucking foss nerds or whoever - should've made it.
SSI is still the perfect balance of just enough power to do templating with .html fragments with a minimal attack surface and no maintainence from version churn. It's stable, it's tested, it's left alone. It's in most major webservers as a core module. It's as good now as it was in 1998.
the prediction wasn't even wrong, it just split in two. deploying a static site in 2026 is genuinely easier than 1998. you drag a folder into netlify and you're done, no ISP instructions, no FTP client, no guy with a tarp. what exploded is everything before the deploy. and the funny thing is the actual complaint buried in this thread, "i don't want to paste my nav into every page and update it by hand," is the exact problem we've spent 25 years re-solving. frames, then SSI, then php includes, then templating engines, then the whole frontend framework industry. react is, underneath everything, a really elaborate way to not repeat your navbar.
Those angry emails from guys (it's always guys) felt so contrived and wedged in just to attack other guys. I reminded me of the tweet about how people online invent someone doing a hypothetical situation and then get mad at them
He's missing the last step:
Somehow around the 2010s we all decided that everything in the web had to become "reactive" and "asynchronous" - which is a fancy way of saying things can theoretically happen at any time but realistically if you try to make it happen in ways that don't resemble the previous serial approach you get weird race conditions - and instead of making sure this was implemented in HTML now we have to write another web browser on top of our web browser in javascript, using a thing called shadow DOM.
Also somehow now you have to understand how the internet works at protocol level unless you want way worse performance than that page written in Dreamweaver 20 years ago.
But this is fine because this is the way big companies run things, which we all know they always make the correct decisions like giving their AI full access to their login and password recovery process.
And there were a bunch of WYSIWYG editors in the mid-late 90s. It seems like everyone had one, including Netscape: https://en.wikipedia.org/wiki/Netscape_Composer
The very first web browser that ever existed – WorldWideWeb – had a WYSIWYG editor.
https://en.wikipedia.org/wiki/WorldWideWeb
What a brilliant piece of writing. I remember almost every single step—safe for actually getting angry emails. Maybe I ended up being the one writing them.
What a glorious time period.
Interestingly, it would’ve been impossible to share this writing with as many people as the author did by publishing it on mastodon and then it ended up on HN in 1998. The network effects are real.
I enjoyed this. But reading the claim that the iPhone was bad compared to other phones of the day makes me question it all. That's so incredibly backwards. It _was_ a much better internet in your pocket. If you couldn't see that, it says something about you, not phones.
Very little of this rings true for me, but that part worst of all.
The mobile web pre-iPhone was terrible. Nobody used it, nobody wanted to use it, and nobody wanted to build it. At best there was a shitty cut back version on the `m` subdomain. WAP/WML were terrible and didn’t give you anything close to the real web, and XHTML Basic was still-born.
The iPhone came along with its “desktop class web browser” and it genuinely worked. Steve Jobs got on stage and told everybody if they wanted to build apps for the iPhone, they should be web apps. Then he told everybody Flash was terrible – which it was – and that we should all use open standards instead.
Practically overnight, everybody commissioning websites wanted them to be “iPhone-compatible”. They did not ask for mobile sites – they specifically asked for them to be iPhone-compatible.
And because WebKit was open-source (thanks to it being based upon KHTML), all the other phone vendors took the code and ran with it, including Android.
This is why I say there is no single organisation that has done more to push the mobile web forward than Apple. The difference in attitudes and capability towards the mobile web changed practically overnight, and it’s directly attributable to Apple’s intentional actions to develop and promote the mobile web.
> Very little of this rings true
How old are you? I'm betting mid-twenties.
What 'real' web are you going to get on a 1.5in phone screen?
What real web are you going to get when there's no 3g, much less a reliable signal on a lot of places.
Wap worked well for actually relaying information.
The original iPhone had a 3.5” screen, Wi-Fi, and EDGE. The iPhone 3G came a year later, with, shockingly, 3G. So yeah, that web browser that actually worked did actually work. Sure, it was slow compared with today’s speeds, but giving people what they want slowly matters far more than giving people what they don’t want marginally faster. Remember most people were on dial-up at home back in those days so the web was hardly fast on desktop either.
WAP might’ve been able to convey information, but so could Lynx in an 80×24 terminal – people want more than that. WAP sites were never popular (aside from some Japanese platform, IIRC) and I don’t think the average web developer had even heard of WAP or WML at the height of their popularity.
There were loads of Symbian/S60 phones with browsers.
2.5g/GPRS was usable enough at the time.
WAP was more 1999, a lot happened between that and the iPhone.
Symbian phones were also based on WebKit, however it wasn't nearly as usable due to many factors, mostly because it emulated mouse pointer (and you controlled it via the joystick) and didn't provide the pinch-to-zoom fluidity of the iPhone Safari. Also iPhone supported most of the major fonts used on the web, had font smoothing, etc. I've used both Symbian WebKit and Safari on first iPhone, and they aren't even close even though they're based on the same underlying browser tech.
I’m not saying the iPhone was the first phone with a web browser, I’m saying that it was the first phone with a web browser that was good enough for average people to want to use it.
Clients were lining up en masse for iPhone-compatible websites, but none of them ever cared about S60 compatibility.
The original iPhone was genuinely terrible. A 2006 Nokia could surf the web on the go and tell you where you are. The iPhone could do neither, since Apple did not include a 3G modem or GPS. It also did not have any apps, and one of the key features highlighted by Steve Jobs was voicemail. The 3G one year later was the first truly usable iPhone.
It was a bad phone, poor battery life, fragile, and relatively poor reception.
That was more than offset by the unmetered internet connection + decent browser, but that’s a feature not everything.
> It was a bad phone
It got a bit of traction, I think they did ok.
Out of interest I checked. I live in New Zealand and iPhone revenue is about 60% more than my countries GDP.
it's 3 products: a widescreen iPod, a phone, and an internet communicator. Are you getting it?
I seem to recall I had a windows phone at the time, with a full keyboard. I could use OS maps on the thing, and although it didn't have GPS, it could get my rough location by tower.
My mate had an iPhone, and it had an app where you could pretend to drink a pint of beer.
That might prove the novelty of the App Store back then. Also, let’s not forget that its’ screen was better than most other smartphones of its time. It was a rather limited phone, but not bad exactly.
The App Store wasn’t planned on release. Apple was pushing for html5 mobile apps at the time. If people had native apps it was via jailbreak
My Nokia 5800 express could connect to a VPN and run a bridged internet connection over Bluetooth in 2008. I clearly remember using it in Switzerland around 2010 to pay for a tramway using a QR code.
Sure, UI was way worse compared to UIKit but in term of features Symbian phones were light-years beyond the iPhone at least until the iPhone 4.
Another way of looking at it is that it made the "real" web shit.
The mobile-web world before the iPhone was one where mobile devices were second class citizens; desktop "real" websites first, and scaled down versions of those sites second (for mobile devices). The information itself was readily available to both, even if the presentation was lacking on the latter.
Jobs knew the only way to win was to not play the same game, because the open web is an even playing field. If you control a new platform on the other hand, you can't lose. So here we are, with locked down, dumbed down toys determining the standard.
It was bad if you had a Nokia smartphone, or a Blackberry. Lots of people felt they weren't ready to give up their keyboards.
It's hard to make objective judgment when you're in one tribe's trench.
Better internet may be, but that doesn't make an entire phone. There was the terrible battery life, limited multitasking, poor camera, etc.
Fuck, you missed it being bad at being a phone. It launched on the network that had such bad coverage they needed Apple's hottest product ever to get people to switch. They had to put a cell repeater (or was it a miniature tower?) in the presentation center so the call would go smoothly, and still had to rig the cellular icon to always show full bars.
Judged purely as a "Let's give Ted a call"-phone, it was fucking bad compared to the competiton. They killed it in other areas, don't get me wrong, but not at being a phone.
It. Was. Bad. But Apple managed to strongarm telcos and it came with free unlimited 3G.
Strike that, not 3G, 2G and Edge !
I read the beginning of the post and immediately tabbed back to start writing a rant about standing up a website being fine, but the real loss of web functionality was Flash, glad I kept reading. I'm quite good with CSS and doing tricks with SVGs, but constantly run into things I want to do in 2D web I find to be complex enough and time consuming enough I don't even bother, while I would have been able to do it in Flash as a 13 year old in a few minutes. The modern web is a prison built out of <div>s, the tricks you see for "amazing" websites with obnoxious scrolljacking/parallax don't hide it.
Actually, do things ever get easier to tinker with as technology matures?
Is it easier to build or repair a radio now than it was when they were first sold? A computer? A car? A washing machine? A vacuum cleaner?
> Is it easier to build or repair a radio now than it was when they were first sold?
"Build"[1], yes. "Repair", not really. A lot of things (like radio) can be an isolated module you just drop in.
Right now a bright HS student can produce a clock radio by plugging together an arduino, an I2C FM receiver, I2C display, an I2C audio amp and a speaker. Pop it all into an enclosure and you have a real clock radio.
Compare to building a clock radio back in 1990 - you'll have to understand how to wire up an op-amp, the principles behind op-amps, how to drive the readout from a clock, etc.
-------------------
[1] By "build" I assume you mean "assemble from parts".
But this is about building radios, which is easier (probably?) now. You can drag and drop an FM radio in pure software (well, with a USB dongle). Or if you insist on hardware, use free tools and cheap PCB services to prototype your radio, grab the parts from Digikey and follow a Youtube tutorial to solder it all together.
Yes, before the enshittification kicks in.
3D printers are mostly still repairable and far more reliable and usable than a few years ago when the majority of the hobby wasn’t making stuff, it was tuning the printer to work reliably. Bambu and California may be signalling that the enshittification inflection point is near.
Not quite the same level, but home/hobby electronics with tiny microcontrollers is more accessible than ever before thanks to the availability of cheap ESP8266/32 clones.
And there are some obvious individual counter-examples - Framework computers, or repairable blenders[0]… but you’ve got to pay a premium for the privilege.
But broadly you’re totally right - in the modern world, by the time something becomes a mainstream product aimed at general consumers, there’s a profit to be made and it’s likely on a downwards path.
[0] https://www.openfunk.co/pages/re-mix
My experience was almost identical. Don’t forget all the time spent balancing image quality with load speed to get that perfect blend of shit quality and slow download.
> Making a website should be the easiest fucking thing in the world by now. There should be a program that you can use to spit out a website as easy as Word spits out words on paper
squarespace?
> and we - us fucking foss nerds or whoever - should've made it.
wordpress?
but but WordPress is written in PHP screeeee <explode type="nerd-apoplexy/>
SSI is still the perfect balance of just enough power to do templating with .html fragments with a minimal attack surface and no maintainence from version churn. It's stable, it's tested, it's left alone. It's in most major webservers as a core module. It's as good now as it was in 1998.
Not seeing anything
Looks like its getting a ton of traffic
Spinner and load bar at the top
Missed Twitter bootstrap which was huge for web design.
Also missed the HP IPAQ.
Completely ignoring accessibility
My actual thoughts were "Wow, doing front end is a shitty experience in 1998, I'll bet it'll be even shittier in 2026".
the prediction wasn't even wrong, it just split in two. deploying a static site in 2026 is genuinely easier than 1998. you drag a folder into netlify and you're done, no ISP instructions, no FTP client, no guy with a tarp. what exploded is everything before the deploy. and the funny thing is the actual complaint buried in this thread, "i don't want to paste my nav into every page and update it by hand," is the exact problem we've spent 25 years re-solving. frames, then SSI, then php includes, then templating engines, then the whole frontend framework industry. react is, underneath everything, a really elaborate way to not repeat your navbar.
> react is, underneath everything, a really elaborate way to not repeat your navbar.
That specific use-case is now replaced by having a single, small webcomponent for client-side includes.
Is a much better dev-XP than configuring the server, then tying your sources to that specific server.Those angry emails from guys (it's always guys) felt so contrived and wedged in just to attack other guys. I reminded me of the tweet about how people online invent someone doing a hypothetical situation and then get mad at them