In 1992ish I worked at RNEC Manadon (UK, Devon). I was asked by my boss to investigate this new www thing.
I telnetted to the nearest VAX from my Win 3.1 PC. I then telnetted to the X.25 PAD and used that to go via the US to Switzerland and CERN. It looked just like gopher and WAIS to me and that's how I reported back - "it looks the same as gopher".
When Tim BL invented www, html and that, browsers were telnet and graphics was a nonsense.
The experience was very different on a NeXT computer.
WAIS was modeled after the built in DigitalLibrarian software. You would select a site in the upper pane, and enter a search term in the box in the middle, and a list of documents would come back in the bottom pane that you could double click and open. Very search engine like.
Gopher was structured and I think Gemini today still sticks with the format. You load a site and the hierarchy of links appeared in a column browser up top and selected documents appeared in the bottom pane.
WWW didn't seem like much in comparison because they were freeform documents without app level navigation support and there wasn't support for images or much formatting and people had not learned to make web pages so it was really hard to see the future of what it would grow to become.
My early career was defined by showing up ten minutes late to several revolutions in a row.
I had a friend who was the most junior developer on the Mosaic team and one day he took me to his office to show me a text document with an image in the middle of it. In theory I met Marc Andreesen and Eric Bina that day but I just wanted to go do something with my friend. I did not get it. At all. A year later my girlfriend had to re-explain it to me and then another few months later I applied to work there in a support role. I don't think she knew what to do with the level of enthusiasm I wasn't bringing to this opportunity.
A year after that I'm sitting in a bar after a tech convention in Chicago, wearing my Mosaic t-shirt, and someone said, 'where did you get that shirt?' When I told them we were on the team, you'd have thought I'd said we were Madonna's backup band.
I never entirely understood that "I'd rather be lucky than good" sentiment until my luck ran out, and now I know.
There was a website back when I was in college where you'd click through some presentation or tutorial or something about bitcoin and at the end they'd just give you 0.5 bitcoin for free.
I have decided that the “start investing early for compound interest” advice is actually a very clever white lie told to young adults everywhere.
The point of starting early is not compound interest. It’s to experience loss when you still have a pittance in the market. The older you get the bigger the chunk of cash you can put in, and if you don’t understand Let it Ride and rebalancing before 20% is a loss of thousands instead of hundreds of dollars, you’re gonna have a bad time.
The only compound interest that really matters is what you get when you have a substantial stake that you also haven’t blown up chasing fads or snake oil. So the advice is technically true but also technically beside the point.
We used to tests servers before deploying them to customers and for that we ran intensive CPU software for days.
I told my direct manager to mine bitcoin for fun. But he being a nerd for UFOs proposed to use Seti@Home.
This was 2009, months after the official launch.
We had extremely expensive servers with multi-cpu setups continuously running. We could have become easily one of the top miners nodes in the world back then. But instead we helped to proof the lack of alien communication towards the earth.
I worked at an EDI company in the mid 90s. X.25 was the wild west. We had a router set up on it that would happily stand up a ppp session to anyone that knew the node name. No password, right on the core network lol.
It certainly was! I remember connecting to Tymnet and Sprintnet/Telenet as a teenager, probably around 1990 or 91. Someone on a local BBS gave me a username that let me connect to QSD and another European chat system. Someone on there had taken over the "system" account on a VAX and was giving out accounts that let you use it as PAD. This went on for weeks. The company must've freaked when they got their x.25 bill. Zero security in those days. The early Internet was just as bad.
Everything connected to the internet was really bad until automatic updates that are enabled by default (or enforced by sysadmins) became a thing. Wordpress, Mysql, Active Directory... all those things had unpatched exploits that you could trivially tap in to until the 2010s if you knew how to use nmap and metasploit. Add insecure wifi standards like wep and basically every other network was fair game for people who had some basic skills. Heck, facebook only made https mandatory in 2013 after someone made a browser plugin that let literally everyone steal cookies on public networks and log in to other people's accounts. Gen Zers never saw this, but the modern web as a secure place where you can comfortably buy stuff or do banking without worries is a relatively recent invention.
I got on the 'Net in 1993. The Web was very "meh". A lot of tutorials on how to write HTML, very little useful content yet. IRC and Usenet were where the action was.
Every time I'm downloading something and "only" getting a megabyte/sec download I take a deep breath and remember the days when I was getting 0.5KB/sec.
Fun fact: Erwise[0] was the first _graphical_ browser developed by a group of students in Helsinki University of Technology with Sir Berners Lee. Sadly there was no funding in Finland available at the time and they had to abandon the project and most of the group ended up working at Tekla, contributing to a bunch of cool AEC CAD technology (Tekla is now a Trimble subsidiary).
Not really. In Graz we had our better Hyper-G graphical browser before CERN, with a completely integrated system to ensure link consistency. Every browser was also the editor. In 1989. https://arxiv.org/pdf/2105.08057
At CERN they wanted to enrich gopher with multimedia data to share building plans and images of their complicated plans, in Graz we wanted to provide a rich teaching and information platform for students. Sadly we went commercial and not open source, so worse got better. Well, a session-less server as httpd was actually better.
It's a real shame both Job's movies skip right over his NeXT and Pixar days..
In 1983 he predicted 10-15 years until home network connectivity is "solved". 10 years later the world wide web released to the public, originally developed on his company's NeXT platform in 1989..
Indeed, to the exclusion of the iPod or iPhone launches.
And it gets the core idea right, too, that NeXT was a commercial failure but built the core OS technology launchpad for the mobile revolution—after saving Apple, of course.
It’s told in a wildly ahistorical framing, but I find the stage play-like “your life passes before your eyes” structure to be much truer to Jobs’ story—and more entertaining—than the Isaacson book.
Do we know that they didn't have some backend code handing the editing?
I don't think a web where every page is globally editable by default would be a good idea, but I can't imagine at all how it would work without a backend, unless all of the changes are just local. But that seems pointless.
Being able to change stylesheets, disable or enhance various JavaScript scripts, add notes and annotations, and other things, is exactly the idea of a user agent.
The user makes a request, and then does whatever they like with the answer. Not just whatever is sensible, but whatever they want to do.
If that concept somehow became accepted again... I think the accessible web might well become a solved problem, rather than an endless slog.
I'm aware of plugins. But these used to be builtin features. Developers needed to work with them, rather than making it harder and harder to use them to make the users life easiee.
You'll need to do a bit of work to make it the way it used to be. Editing any text on a page, or having your changes save persistently, needs a bit of a... Framework, to keep things together, rather than being the expected mode of interaction.
Sure, I can add a p to the tree. But if I refresh, its gone. I'll probably need plugins to keep my own stylesheets and JS changes around.
I love what the CERN team did here visually with the NeXT UI. Rebuilding a historical browser inside a modern one is a fun rabbit hole, but man, it is the same technical wall to hit every time: iframes.
You build this beautiful retro UI, you wire up the address bar, and then you try to load a modern site and just hit a wall of CORS, X-Frame-Options, and CSP blocks. Which, tho is probably precisely things should work. Otherwise people arbitrarily iframe the open web opening up a massive clickjacking-pocalypse. It makes total sense for security....sigh.
But I sitll wanted a way to get around it to capture that 90s nostalgia (tho NeXT and this browser were actually from the late 80s), the real open web inside a retro recreation not just a crippled, iframe-blocked imitation. Or "everything links to archive org" stuff.
To make that work, I had to make a custom embedder API. It basically pipes a fully isolated remote Chromium instance right into the retro shell through an iframe in a custom element. The engine is real, and it respects the native security boundaries because the browser is physically isolated, but it wears that heavy 90s UI so you get the 90s feel.
If you want to mess around with a different flavor of 90s nostalgia that can actually surf the modern web, I put up a live version here: https://win9-5.com/demo. Sound on for the retro modem dial-up elevator music. The non-graybeards may never have experienced the modem's mating call in the wild.
When watching this I'm shocked how bad the UX Was these days. The scrollbar left, the triple steped menu...
What was improved sometimes is only visible when we see how it was back in the past.
> When watching this I'm shocked how bad the UX Was these days. The scrollbar left, the triple steped menu...
Perhaps the only thing "bad" about it is that you're simply not used to it. I can certainly think of someone used to that UI thinking the same thing about today's interfaces, with disappearing scrollbars, flat design and confusing icons.
> The deeply nested menu for entering the url, that’s bad, I agree.
I'm not saying it is perfect, but it was not that bad, really. It's only one level down. And then you could also use a keyboard shortcut for it, which is always faster than anything mouse-driven if your hands are on the keyboard, which they would be, if you wanted to type a URL.
And even if you had to use the mouse, there is an interface feature we have lost: tear-off menus. If you found that you needed something in a nested menu often, you could simply tear-off that submenu and pin it on your desktop so you can always have direct access.
It's a javascript-based imitation, much like all of those js-based imitations of various Windows versions.
The original source code isn't really involved, which is a shame, since it is actually available.
IMHO this should have been (something along the lines of) GNUstep + TimBL's original code (mirror: https://github.com/cynthia/WorldWideWeb) + Emscripten + getting Emscripten to work with ObjC. Now, that would have been cool.
This is the most commented HN posting on this from that time (2019):
TimBL's original NeXT is still on display at CERN, I've seen it.
I've even stood in the office that was his when he wrote it (it was empty when I was there, but had recently been in use by some incredibly high-end physicist).
That makes me think about the whatng cartel apocalypse.
People lost themselves, forgetting how important noscript/basic (x)html (aka basic HTML forms, nowdays which could be augmented with <audio> and <video>)) has been for web technical independence.
Part of the original requirements was the decentralized nature, which I always found extra interesting:
> CERN Requirements - Non-Centralisation - Information systems start small and grow. They also start isolated and then merge. A new system must allow existing systems to be linked together without requiring any central control or coordination.
Doesn't directly answer your question I suppose, but gives at least one perspective on how at least one person saw it at that point :)
It wasn't a rocket science amazing idea in 1989. It was a pretty obvious thing anyone in the space could see would be interesting to try -- hypertext already existed. The internet was just taking off. The idea that you could host pages that hyperlinked to each other on the internet was totally obvious and if you'd explained it to anyone active in the internet at the time (obviously not that many of those people), they'd have nodded. I should add that almost nobody had a computer with a graphics display in 1989 (I did, at work) so that further constrains the set of folk to whom the idea would make sense. The fact that everyone now has many 64-bit computers with extremely high resolution displays is probably the more surprising thing to 1989-dude.
network users at that time already had software for ftp and other common tools. Gopher sort of linked logically to an ftp idea. Mosaic was often introduced in the same sentence as "uses a format called HTML" .. Mosaic seemed interesting but also it was obvious that pages in that format would have to become popular, to make more of them. There wasn't a big reason to switch your daily software to Mosaic since stable apps were better for their existing uses. It was a very rare thing to have access to a NeXT machine (maybe not on YNews).
From my point of view it was Netscape that made a big splash, a year+ later, with a lot of publicity and good graphic design. Mosaic itself was an awkward demo with an interesting nerdy story.
WorldWideWeb didn't originally support inline images, and while using a graphical toolkit rendered pages more like Lynx, albeit with the ability to vary fonts. Lynx wasn't the first WWW browser, but came along shortly after, a year or so after WorldWideWeb, and is the oldest browser still maintained. See https://en.wikipedia.org/wiki/History_of_the_web_browser#Ear...
I'm having trouble pinning down when WorldWideWeb got inline image support, but based on https://www.w3.org/History/1991-WWW-NeXT/Implementation/Feat... I'm guessing sometime between 1992 and 1994, when there are screenshots with inline images, so maybe after Lynx was published.
WorldWideWeb could display images, but originally only in a separate window when you clicked on them, similar to the way audio, PDFs, and other multimedia worked (and sometimes still work). The wording of one of the people involved seems to confirm this:
> How was I to know that I was passing an historical milestone, as the one above was the first picture of a band ever to be clicked on in a web browser!"
It's been a very long time, but my recollection was the Mosaic did images first, and it was non-standard. (The beginning of the end.) I might be thinking of some other feature though.
I was also disappointed that the editing went away after the first browser. (There was "Amaya" which had editing, but it was a research thing and not a commonly used browser.)
There was also OmniWeb on the Next machine, but there weren't a lot of NeXT machines around.
Mosaic was the first browser to support images because HTML didn't support images and Marc Andreesen and Eric Bina sat in a coffee shop on campus while Marc talked himself into going rogue and making his own tag while Eric didn't talk him out of it (source, Eric Bina, ACM lecture at UIUC ca 1995)
This lecture Alan aimed at this particular audience, the computer science (programming) students at University of Illinois, where they programmed the second browser, the second broken wheel 20 years after Alan and Dan had showed them how do do it better.
Dan Ingalls implemented most of Alan Kay's invention of the personal computer, in the following demo's he shows how to fix the webbrowser's broken wheel a bit.
The Lively Kernel would be another way to fix html but retain the web. Two demos says it all:
> The Lively Kernel would be another way to fix html but retain the web.
The Web is not HTML (and it's not JavaScript). It's URLs. It's a machine-readable graph of clickable references on cross-linked Works Cited pages. It's certainly not Smalltalk-over-the-Internet, and it's not trying to be (at least it wasn't when TBL created it).
The biggest problem facing the Web in the 90s and still today is that everyone who saw it then hallucinated TBL describing an SRI-/PARC-style application platform because that's what they wanted it to be—including people like Alan Kay—who then perversely go on to criticize it for being so unaligned with that vision.
It is both surprising and unsurprising (given this reaction) that the industry managed to make it all the way through the 90s without Wikipedia showing up until after the crash.
88 comments:
In 1992ish I worked at RNEC Manadon (UK, Devon). I was asked by my boss to investigate this new www thing.
I telnetted to the nearest VAX from my Win 3.1 PC. I then telnetted to the X.25 PAD and used that to go via the US to Switzerland and CERN. It looked just like gopher and WAIS to me and that's how I reported back - "it looks the same as gopher".
When Tim BL invented www, html and that, browsers were telnet and graphics was a nonsense.
The experience was very different on a NeXT computer.
WAIS was modeled after the built in DigitalLibrarian software. You would select a site in the upper pane, and enter a search term in the box in the middle, and a list of documents would come back in the bottom pane that you could double click and open. Very search engine like.
Gopher was structured and I think Gemini today still sticks with the format. You load a site and the hierarchy of links appeared in a column browser up top and selected documents appeared in the bottom pane.
WWW didn't seem like much in comparison because they were freeform documents without app level navigation support and there wasn't support for images or much formatting and people had not learned to make web pages so it was really hard to see the future of what it would grow to become.
I'm not known for picking winners :-(
My early career was defined by showing up ten minutes late to several revolutions in a row.
I had a friend who was the most junior developer on the Mosaic team and one day he took me to his office to show me a text document with an image in the middle of it. In theory I met Marc Andreesen and Eric Bina that day but I just wanted to go do something with my friend. I did not get it. At all. A year later my girlfriend had to re-explain it to me and then another few months later I applied to work there in a support role. I don't think she knew what to do with the level of enthusiasm I wasn't bringing to this opportunity.
A year after that I'm sitting in a bar after a tech convention in Chicago, wearing my Mosaic t-shirt, and someone said, 'where did you get that shirt?' When I told them we were on the team, you'd have thought I'd said we were Madonna's backup band.
I never entirely understood that "I'd rather be lucky than good" sentiment until my luck ran out, and now I know.
Love it. This is all soo very Nelson Bighetti.
Nelson did quite fine in the end.
>"My early career was defined by showing up ten minutes late to several revolutions in a row."
Ha, I missed so many great things. The most obvious was not to buy $10K worth of bitcoin when it just started.
Luckily (or not) I am an easy going person and do not dwell on things.
It could be worse: you could have bought bitcoin when it started and then have sold it for a profit of $40. ;)
There was a website back when I was in college where you'd click through some presentation or tutorial or something about bitcoin and at the end they'd just give you 0.5 bitcoin for free.
I could have retired making early iPhone apps but I was already so burned out on how shitty mobile carriers were behaving that I just sat it out.
Hey at least you didn't aggressively "day trade" it all away with your idiot friends who moved in and tried to start a "fund". Good times.
I have decided that the “start investing early for compound interest” advice is actually a very clever white lie told to young adults everywhere.
The point of starting early is not compound interest. It’s to experience loss when you still have a pittance in the market. The older you get the bigger the chunk of cash you can put in, and if you don’t understand Let it Ride and rebalancing before 20% is a loss of thousands instead of hundreds of dollars, you’re gonna have a bad time.
The only compound interest that really matters is what you get when you have a substantial stake that you also haven’t blown up chasing fads or snake oil. So the advice is technically true but also technically beside the point.
In the beginning you didn't really buy BTC. You could mine a few off of your nvidia card in less than a week.
I was focused on doing useless things like cracking md5 hashed passwords and didn't really believe you could pay for things with it.
Regret on a different level.
We used to tests servers before deploying them to customers and for that we ran intensive CPU software for days.
I told my direct manager to mine bitcoin for fun. But he being a nerd for UFOs proposed to use Seti@Home.
This was 2009, months after the official launch.
We had extremely expensive servers with multi-cpu setups continuously running. We could have become easily one of the top miners nodes in the world back then. But instead we helped to proof the lack of alien communication towards the earth.
> The most obvious was not to buy $10K worth of bitcoin when it just started.
"Forty quid for a string of hex digits? Nah, I don't think so..."
- me, some time in 2010.
- me, also in 2010, when a junior colleague ("what would he know?") spent his bonus on bitcoin.
But you were lucky. You were in the right places at the right time, just didn't realize it.
This is lack of vision, not lack of luck.
> WAIS was[…] Gopher was[…]
> You load a site and the hierarchy of links appeared in a column browser up top and selected documents appeared in the bottom pane.
You're mentioning formats and protocols but describing application UI designs.
Something that just occurred to me: RAGs are almost Gopher for AIs.
> people had not learned to make web pages
Because there wasn't a widespread usable browser until Mosaic came along, 2 1/2 years after WWW.
I worked at an EDI company in the mid 90s. X.25 was the wild west. We had a router set up on it that would happily stand up a ppp session to anyone that knew the node name. No password, right on the core network lol.
It certainly was! I remember connecting to Tymnet and Sprintnet/Telenet as a teenager, probably around 1990 or 91. Someone on a local BBS gave me a username that let me connect to QSD and another European chat system. Someone on there had taken over the "system" account on a VAX and was giving out accounts that let you use it as PAD. This went on for weeks. The company must've freaked when they got their x.25 bill. Zero security in those days. The early Internet was just as bad.
Everything connected to the internet was really bad until automatic updates that are enabled by default (or enforced by sysadmins) became a thing. Wordpress, Mysql, Active Directory... all those things had unpatched exploits that you could trivially tap in to until the 2010s if you knew how to use nmap and metasploit. Add insecure wifi standards like wep and basically every other network was fair game for people who had some basic skills. Heck, facebook only made https mandatory in 2013 after someone made a browser plugin that let literally everyone steal cookies on public networks and log in to other people's accounts. Gen Zers never saw this, but the modern web as a secure place where you can comfortably buy stuff or do banking without worries is a relatively recent invention.
I got on the 'Net in 1993. The Web was very "meh". A lot of tutorials on how to write HTML, very little useful content yet. IRC and Usenet were where the action was.
Wired Magazine famously agreed with you. Usenet was where it was at then.
Internet commercialization wasn't really on until 1994. Then anyone could get dial-up IP, they could put ads on their webpages, and etc.
I remember that. I had almost zero interest in www until geocities came along and then …it was something else to compose and publish a “website”
The whole thing was atrocious but at least introduced me to the concept.
In fact, I had to spend like three days downloading Netscape to try it out because I didn’t even have a graphical browser yet.
Every time I'm downloading something and "only" getting a megabyte/sec download I take a deep breath and remember the days when I was getting 0.5KB/sec.
Fun fact: Erwise[0] was the first _graphical_ browser developed by a group of students in Helsinki University of Technology with Sir Berners Lee. Sadly there was no funding in Finland available at the time and they had to abandon the project and most of the group ended up working at Tekla, contributing to a bunch of cool AEC CAD technology (Tekla is now a Trimble subsidiary).
[0]https://en.wikipedia.org/wiki/Erwise
> Erwise[0] was the first _graphical_ browser
No, as indicated in the submission the original WorldWideWeb.app (developed on a NeXTCube) is a graphical Web browser.
Not really. In Graz we had our better Hyper-G graphical browser before CERN, with a completely integrated system to ensure link consistency. Every browser was also the editor. In 1989. https://arxiv.org/pdf/2105.08057
At CERN they wanted to enrich gopher with multimedia data to share building plans and images of their complicated plans, in Graz we wanted to provide a rich teaching and information platform for students. Sadly we went commercial and not open source, so worse got better. Well, a session-less server as httpd was actually better.
It's a shame that Hyper-G has not been better preserved. It had some remarkable features (disclosure, my article: https://oldvcr.blogspot.com/2025/05/prior-art-dept-hierarchi... ).
Fixed link: https://oldvcr.blogspot.com/2025/05/prior-art-dept-hierarchi...
Ah - I stand corrected - thanks!
Related. Others?
WorldWideWeb – the first web browser - https://news.ycombinator.com/item?id=34218591 - Jan 2023 (18 comments)
The Browser – WorldWideWeb Next Application (2019) - https://news.ycombinator.com/item?id=26680839 - April 2021 (21 comments)
The Browser – WorldWideWeb Next Application - https://news.ycombinator.com/item?id=25013103 - Nov 2020 (8 comments)
CERN 2019 WorldWideWeb Rebuild - https://news.ycombinator.com/item?id=24939929 - Oct 2020 (7 comments)
CERN 2019 WorldWideWeb Rebuild - https://news.ycombinator.com/item?id=19249373 - Feb 2019 (45 comments)
CERN 2019 WorldWideWeb Rebuild: 2019 rebuilding of the original NeXT web browser - https://news.ycombinator.com/item?id=19183316 - Feb 2019 (1 comment)
It's a real shame both Job's movies skip right over his NeXT and Pixar days..
In 1983 he predicted 10-15 years until home network connectivity is "solved". 10 years later the world wide web released to the public, originally developed on his company's NeXT platform in 1989..
The 2015 movie has the entire second act dedicated to the Next launch no?
Indeed, to the exclusion of the iPod or iPhone launches.
And it gets the core idea right, too, that NeXT was a commercial failure but built the core OS technology launchpad for the mobile revolution—after saving Apple, of course.
It’s told in a wildly ahistorical framing, but I find the stage play-like “your life passes before your eyes” structure to be much truer to Jobs’ story—and more entertaining—than the Isaacson book.
One of my favorite films of the last 20 years.
Direct link to the browser: https://worldwideweb.cern.ch/browser/
Did you notice you can click anywhere in the text and edit it?
Something was lost along the way.
(Nowadays you need a separate wiki engine on a site to be able to do that)
F12, Console, type
(it is slightly different though, as links cannot be followed)
F12, Console, type document.designMode = 'off'
I had no idea. You just blew my mind
> (Nowadays you need a separate wiki engine on a site to be able to do that)
No you don’t. These browser simply PUTs the request and your web server simply edits the document. Versioning is optional, of course.
The original read/write web
Do we know that they didn't have some backend code handing the editing?
I don't think a web where every page is globally editable by default would be a good idea, but I can't imagine at all how it would work without a backend, unless all of the changes are just local. But that seems pointless.
Being able to change stylesheets, disable or enhance various JavaScript scripts, add notes and annotations, and other things, is exactly the idea of a user agent.
The user makes a request, and then does whatever they like with the answer. Not just whatever is sensible, but whatever they want to do.
If that concept somehow became accepted again... I think the accessible web might well become a solved problem, rather than an endless slog.
>Being able to change stylesheets
https://addons.mozilla.org/en-US/firefox/addon/styl-us/
>disable or enhance various JavaScript scripts
https://addons.mozilla.org/en-US/firefox/addon/ublock-origin...
https://addons.mozilla.org/en-US/firefox/addon/tampermonkey/...
Yeah you can't directly alter scripts being ran (as far as I know?) but you can usually override/extend behavior and can definitely add your own
>add notes and annotations
https://cwmonkey.github.io/greasemonkey/make-note/
(I haven't actually used this one, just first result)
I'm aware of plugins. But these used to be builtin features. Developers needed to work with them, rather than making it harder and harder to use them to make the users life easiee.
In what way is that not currently possible? All browsers I know of you can edit whatever you want in any page you download
You'll need to do a bit of work to make it the way it used to be. Editing any text on a page, or having your changes save persistently, needs a bit of a... Framework, to keep things together, rather than being the expected mode of interaction.
Sure, I can add a p to the tree. But if I refresh, its gone. I'll probably need plugins to keep my own stylesheets and JS changes around.
HTTP has PUT and DELETE for a reason ;-)
> But that seems pointless.
Making notes for your own consumption?
Upload the file when you are done, perhaps?
It has been about 16 years since I fired up my old NeXTStation Color where I had a copy of 1.0 or a late beta.
The last time I tried about the only site that worked was useit.com, former home of Nielsen Norman UX experts ;-)
I love what the CERN team did here visually with the NeXT UI. Rebuilding a historical browser inside a modern one is a fun rabbit hole, but man, it is the same technical wall to hit every time: iframes.
You build this beautiful retro UI, you wire up the address bar, and then you try to load a modern site and just hit a wall of CORS, X-Frame-Options, and CSP blocks. Which, tho is probably precisely things should work. Otherwise people arbitrarily iframe the open web opening up a massive clickjacking-pocalypse. It makes total sense for security....sigh.
But I sitll wanted a way to get around it to capture that 90s nostalgia (tho NeXT and this browser were actually from the late 80s), the real open web inside a retro recreation not just a crippled, iframe-blocked imitation. Or "everything links to archive org" stuff.
To make that work, I had to make a custom embedder API. It basically pipes a fully isolated remote Chromium instance right into the retro shell through an iframe in a custom element. The engine is real, and it respects the native security boundaries because the browser is physically isolated, but it wears that heavy 90s UI so you get the 90s feel.
If you want to mess around with a different flavor of 90s nostalgia that can actually surf the modern web, I put up a live version here: https://win9-5.com/demo. Sound on for the retro modem dial-up elevator music. The non-graybeards may never have experienced the modem's mating call in the wild.
I wish someone would write a reference implementation in a functional language.
At least that would formalize the specification.
When watching this I'm shocked how bad the UX Was these days. The scrollbar left, the triple steped menu... What was improved sometimes is only visible when we see how it was back in the past.
> When watching this I'm shocked how bad the UX Was these days. The scrollbar left, the triple steped menu...
Perhaps the only thing "bad" about it is that you're simply not used to it. I can certainly think of someone used to that UI thinking the same thing about today's interfaces, with disappearing scrollbars, flat design and confusing icons.
The deeply nested menu for entering the url, that’s bad, I agree. But why is a scrollbar on the other side better or worse?
I have the minimap configured on the left in vs code and use it as scrollbar. It’s quite nice actually.
> The deeply nested menu for entering the url, that’s bad, I agree.
I'm not saying it is perfect, but it was not that bad, really. It's only one level down. And then you could also use a keyboard shortcut for it, which is always faster than anything mouse-driven if your hands are on the keyboard, which they would be, if you wanted to type a URL.
And even if you had to use the mouse, there is an interface feature we have lost: tear-off menus. If you found that you needed something in a nested menu often, you could simply tear-off that submenu and pin it on your desktop so you can always have direct access.
It has a certain charm, like everything engineered without immediate commercial considerations
Menus can be torn off and positioned as desired.
I miss that.
It's a javascript-based imitation, much like all of those js-based imitations of various Windows versions.
The original source code isn't really involved, which is a shame, since it is actually available.
IMHO this should have been (something along the lines of) GNUstep + TimBL's original code (mirror: https://github.com/cynthia/WorldWideWeb) + Emscripten + getting Emscripten to work with ObjC. Now, that would have been cool.
This is the most commented HN posting on this from that time (2019):
https://news.ycombinator.com/item?id=19249373
A WASM emulator of 68040 and NeXT, the original OS and compiler, then run WWW on top of that.
The performance would likely be comparable %)
The performance would likely be much better. :-)
An example from a year later; 1990:
https://copy.sh/v86/?profile=windows30
Every version of NeXT is actually catalogued and can be run in the browser at: https://infinitemac.org/
I'm pretty sure the CERN WorldWideWeb application is also included in the "bonus software" HDD, but I'm on my phone right now and can't confirm. :-)
You can already run nextstep in browser, see https://aresluna.org/frame-of-preference/ (section "Back to 1992")
TimBL's original NeXT is still on display at CERN, I've seen it.
I've even stood in the office that was his when he wrote it (it was empty when I was there, but had recently been in use by some incredibly high-end physicist).
That makes me think about the whatng cartel apocalypse.
People lost themselves, forgetting how important noscript/basic (x)html (aka basic HTML forms, nowdays which could be augmented with <audio> and <video>)) has been for web technical independence.
All that is very sad, and toxic.
i want to meet those people who did this originally in 1989 and to ask them, did they ever think it would be this today?
The original WWW proposal is quite easy and interesting to read through: https://www.w3.org/History/1989/proposal.html
Part of the original requirements was the decentralized nature, which I always found extra interesting:
> CERN Requirements - Non-Centralisation - Information systems start small and grow. They also start isolated and then merge. A new system must allow existing systems to be linked together without requiring any central control or coordination.
Doesn't directly answer your question I suppose, but gives at least one perspective on how at least one person saw it at that point :)
It wasn't a rocket science amazing idea in 1989. It was a pretty obvious thing anyone in the space could see would be interesting to try -- hypertext already existed. The internet was just taking off. The idea that you could host pages that hyperlinked to each other on the internet was totally obvious and if you'd explained it to anyone active in the internet at the time (obviously not that many of those people), they'd have nodded. I should add that almost nobody had a computer with a graphics display in 1989 (I did, at work) so that further constrains the set of folk to whom the idea would make sense. The fact that everyone now has many 64-bit computers with extremely high resolution displays is probably the more surprising thing to 1989-dude.
Better than chrome!
(2019)
Some previous discussions:
2023 https://news.ycombinator.com/item?id=34218591
2021 https://news.ycombinator.com/item?id=26680839
2020 https://news.ycombinator.com/item?id=25013103
2019 https://news.ycombinator.com/item?id=19249373
> The WWW project does not take responsability
I guess that let them off the hook for incorrect spelling. :-)
Links where called Pointers back then apparently.
network users at that time already had software for ftp and other common tools. Gopher sort of linked logically to an ftp idea. Mosaic was often introduced in the same sentence as "uses a format called HTML" .. Mosaic seemed interesting but also it was obvious that pages in that format would have to become popular, to make more of them. There wasn't a big reason to switch your daily software to Mosaic since stable apps were better for their existing uses. It was a very rare thing to have access to a NeXT machine (maybe not on YNews).
From my point of view it was Netscape that made a big splash, a year+ later, with a lot of publicity and good graphic design. Mosaic itself was an awkward demo with an interesting nerdy story.
No link to EDH.
All the links should point to the 1989 internet instead of "Not Found"
:-)
Interesting, for some reason I thought lynx was the first browser. I thought I read that a while ago.
But it makes sense it is a GUI browser since it was developed on a NeXT
WorldWideWeb didn't originally support inline images, and while using a graphical toolkit rendered pages more like Lynx, albeit with the ability to vary fonts. Lynx wasn't the first WWW browser, but came along shortly after, a year or so after WorldWideWeb, and is the oldest browser still maintained. See https://en.wikipedia.org/wiki/History_of_the_web_browser#Ear...
I'm having trouble pinning down when WorldWideWeb got inline image support, but based on https://www.w3.org/History/1991-WWW-NeXT/Implementation/Feat... I'm guessing sometime between 1992 and 1994, when there are screenshots with inline images, so maybe after Lynx was published.
Well there was this image:
https://www.themarginalian.org/2014/07/18/les-horribles-cern...
WorldWideWeb could display images, but originally only in a separate window when you clicked on them, similar to the way audio, PDFs, and other multimedia worked (and sometimes still work). The wording of one of the people involved seems to confirm this:
> How was I to know that I was passing an historical milestone, as the one above was the first picture of a band ever to be clicked on in a web browser!"
Source: https://musiclub.web.cern.ch/bands/cernettes/firstband.html
It's been a very long time, but my recollection was the Mosaic did images first, and it was non-standard. (The beginning of the end.) I might be thinking of some other feature though.
I was also disappointed that the editing went away after the first browser. (There was "Amaya" which had editing, but it was a research thing and not a commonly used browser.)
IIRC, Viola also got scooped by Mosaic, which was the first browser most people used, before you could buy one shrinkwrapped at a store.
https://en.wikipedia.org/wiki/ViolaWWW
There was also OmniWeb on the Next machine, but there weren't a lot of NeXT machines around.
Mosaic was the first browser to support images because HTML didn't support images and Marc Andreesen and Eric Bina sat in a coffee shop on campus while Marc talked himself into going rogue and making his own tag while Eric didn't talk him out of it (source, Eric Bina, ACM lecture at UIUC ca 1995)
You can't fix a broken wheel. Let the downvotes illustrating the ignorance of HN pop culture start....
This web link post, the original NEXT webbrowser as a web page, tries to celebrate and revive the reinvention of the broken wheel.
The World Wide Web, browser and html standards are a very broken wheel. Alan Kay, the inventor of personal computing, explains why:
https://youtu.be/FvmTSpJU-Xc?t=961
Some of the comments of youtube are fun too.
This lecture Alan aimed at this particular audience, the computer science (programming) students at University of Illinois, where they programmed the second browser, the second broken wheel 20 years after Alan and Dan had showed them how do do it better.
Dan Ingalls implemented most of Alan Kay's invention of the personal computer, in the following demo's he shows how to fix the webbrowser's broken wheel a bit.
The Lively Kernel would be another way to fix html but retain the web. Two demos says it all:
https://youtu.be/gGw09RZjQf8?t=147
https://youtu.be/QTJRwKOFddc?t=234
Their Squeak, Etoys and Croquet fixed it completely:
Early Croquet demo (there are several others): https://www.youtube.com/watch?v=XZO7av2ZFB8
Croquet in webbrowser: https://codefrau.github.io/jasmine/
Demo of webbrowser replacement: https://www.youtube.com/watch?v=1s9ldlqhVkM
Squeak and all its predecessors: https://smalltalkzoo.computerhistory.org
Etoys: https://squeak.js.org/etoys/
> The Lively Kernel would be another way to fix html but retain the web.
The Web is not HTML (and it's not JavaScript). It's URLs. It's a machine-readable graph of clickable references on cross-linked Works Cited pages. It's certainly not Smalltalk-over-the-Internet, and it's not trying to be (at least it wasn't when TBL created it).
The biggest problem facing the Web in the 90s and still today is that everyone who saw it then hallucinated TBL describing an SRI-/PARC-style application platform because that's what they wanted it to be—including people like Alan Kay—who then perversely go on to criticize it for being so unaligned with that vision.
It is both surprising and unsurprising (given this reaction) that the industry managed to make it all the way through the 90s without Wikipedia showing up until after the crash.