Looking at the second screenshot in this article (the one titled "An early CERN Web browser, circa 1990"), I feel this is what I'd like the modern web to be: crisp, clean, high density information, using colors only where it is relevant.
Modern web pages have become a seizure-inducing cacophony of colors, fonts, animations, ads, videos and intrusive modals ("please sign up", "subscribe to our newsletter", "do you accept cookies?", "please disable ad-block", "people in <your town> are excited about this new weight loss product!") etc.
> I feel this is what I'd like the modern web to be: crisp, clean, high density information, using colors only where it is relevant.
Take a trip only 10 years back. Browse with a Wii, or a series 40 Nokia phone. Or even Lynx using a color terminal. The experience bears considerable resemblance to the one you describe.
Everything pulling the experience away from this is a cultural problem. On the technical side there's the increasing tendency to see the browser as a weird VM/runtime first (or, worse, not to think about the browser much at all, and instead see the web primarily through the lens of several popular technological hammers to which every problem looks like a nail). On the business side there's various legal, economic, and marketing incentives. Not all of these are trivially dismissed concerns, but it's still possible to do what you want in simple enough circumstances or with some careful thinking/work in more complex ones.
I'm going to get flamed for this but, if all you're interested in is the text, you could do worse than Lynx. I mean, it's not pretty, but it does get to the point. For sites that absolutely require JavaScript to load content you're screwed, but it does a surprisingly decent job on many others. Can be helpful if you need to avoid becoming distracted.
I actually used text based browsers as my main thing for a few weeks, when I had really bad internet. Ran it thru ssh on a VPS, much faster that way. I found the w3m browser gave the best formatting and mouse support.
When I have a slow internet connection I use the Dillo browser with CSS and image loading disabled. Dillo is also really snappy even on an old computer.
I'd use ELinks or Links2 over Links, AFAIK the originall Links is no longer maintained. There's also partial JS support in ELinks which fixes some sites that require it.
One of the university labs I had to use did not allow access to Netscape for whatever reason, so I actually used Lynx a lot. For reading text on webpages it actually did a quite sturdy job.
In college I had one of those eeePC netbooks. It came with Windows XP on it, but it was so underpowered it couldn't really run a browser, so I threw headless Ubuntu onto it and got briefly good with emacs. I tried a few different console browsers; as I recall elinks was slightly better than lynx, but you could get most of the content you wanted back then, at least. I suspect that a large portion of the web is completely useless if you are still trying to use those browsers now.
I still have one of the original models that came with its own version of Linux. It worked really well with Fluxbox but unfortunately that's no longer being supported, it's currently running Puppy Linux.
What's true is that images & video are effective distractors, and most content "servers" want impressions and ad clicks. I don't think this means that most visitors/readers are necessarily looking for image-/video-heavy content.
The likes of Youtube, etc. may be an exception (where it's clear people have come to the site looking specifically for a video), but even there, the suggested videos sidebar, etc. are not things that visitors necessarily ask for/want, but rather things that benefit advertisers in terms of user attention.
This idea that techies are a different species to "normal people" is an overused red herring. Of course every user/demographic has their different preferences, but generally speaking the only difference with "techies" in this particular instance is awareness: due to familiarity with the medium, they are more deliberate in what they want—and more conscious of image-/video-heavy content being a distraction from what they're looking for—than the average user. This doesn't necessarily mean they're actually looking for different things.
I'd suggest that people go where people are. They probably do want life and would likely be happier if they were interacting with other real live people. But our society is increasingly fragmented with people with good jobs being forced to work all the time and people with bad jobs being forced to work irregular hours and the people with stable reasonable jobs shrinking. This is probably just one factor that leads people to try to socialize online and thus seek a simulacrum of authentic experience which leaves them lacking often in comparison to what a real life experience would be. People who are seeking utility or information through the Internet are less inclined to desire distractions, but people who are seeking distractions or entertainment are probably more inclined towards shiny. The challenge we face is that all human activity is being pushed into the domain of a surveillance panopticon house of mirrors where even people seeking to interact with other people face to face are required to participate lest they be disinvited. This is specific to things like Facebook groups and events to coordinate activities. But yeah advertisers do follow people and they also help corral people by feeding social media companies into places so they have a captive audience.
I remember the first time I saw Mosaic - a friend of mine was reading a digital graphic novel. I was absolutely gobsmacked at how incredible that was: not only did someone create something in a country far from mine, I was able to look at it in my home—perhaps minutes after its creation. So quickly, we become accustomed to the amazing.
I remember first seeing Mosaic when I started taking classes at Brown University. The university CS department people had very wisely set it up so that any CS students could not only very easily create their own Web pages from files in their home directory, but also easily do CGI scripts that were publicly accessible. The easy CGI wasn't so great for security, but was fabulous for lowering the barrier to experimenting.
Using that CGI, I wrote several services, mostly in Perl, and one in C. (The C one was so that I could use the GD library to draw some GIF parts of a hypermedia image of the main workstation lab, with faces of people there, which machines were free, etc., since it wasn't doable as well in HTML at the time.)
After seeing it at the university, I also went and and installed Mosaic and a the NCSA or Apache HTTP server at my workplace (where I was in an R&D group), and wrote some demonstrations of it (e.g., doing our engineering documents in it instead of Interleaf/FrameMaker, clickable hypermedia floorplan).
I'd actually already been on the Internet as a kid (gateways, email, Usenet, ftp, etc.), and had already worked with a few offline hypertext systems before I saw the Web (as well as being aware of earlier grand visions), but the Web was obviously going to let us do Internet things we couldn't before, and also accelerate bringing the goodness of the Internet to everyone. (Well, we dropped the ball on some of the goodness, but it still happens, just not as much nor nearly as universally as we'd hoped, and we can still improve that.)
Likewise. I remember vividly seeing Mosaic for the first time, running on a SPARCstation workstation at university, circa 93. I had been using gopher and lynx for a few months (and BBSes for years before that), but seeing a browser loading a page - with graphics and colors! - was a pivotal experience for me.
(I know, "graphics" is an overstatement, given it meant gimmicky static images at the time, at best. I don't think even tables were available yet)
I don't remember my initial impression. I just remember how 10 minutes later I was trying the hardest pretending I had just stumbled onto pictures of naked women by accident LOL.
I first used SPRY Mosaic as provided by CompuServe (my first ISP). I had written an article for a late 1995 special edition of Doctor Dobbs Journal which still has my screen image captures of the SPRY Mosaic browser:
I mentioned this in a different comment, but I have a jar from back then with that name. Early "if they'd only done research" brand accident. Or maybe they meant to?
Anyone here ever used the web-browser that comes with GeoWorks/New Deal? Pretty amazing software. In the late 90's I was studying in Poland. In the computer lab of my dorm there were two Windows PCs and half a dozen text terminals. People were always waiting to use the more modern machines. The terminals were only good for e-mail (via telnet). They were XT-class machines with Hercules graphics. One day I decided to put a copy of GeoWorks on one of them. From then on I could always surf the web whenever I want. People were quite amazed to see a full GUI on a 12" monochrome screen.
Back then I used lynx on the text terminals at university and shortly after upgraded my PC at home (running Linux) from 4 MB to 8 MB of RAM (which did cost about an arm and a leg) so that I could run Mosaic with its statically linked Motif libs on it... When I think back the thing that amazes me most is the fact that I ran a Unix system along with X11 and a Motif application in all of 8 MB of RAM.
I still use lynx now and then even today. Haven't seen any Motif app for a long while, though.
somehow, seeing Mosaic in 1993 or so was not a big deal.. Gopher was in use and the various FTP clients to transfer files, and Mosaic seemed like a small step from those, just unified into a page. Composable HTML was not a thing yet. The GUI was so primitive compared to daily desktop software, that it wasnt very compelling. Public access to a unified Internet was also not a thing yet, networks were restricted and siloed by governance and protocols. Usenet news already had some binaries if you really wanted to distribute or look for something, but the content was limited. basically, not a fan on first looks
If anyone is wondering what was Netscape’s first big feature differentiator was, it was inline JPEGs. Mosaic could just display GIFs inline. JPEGs would require an external handler, typically LView on Windows, or xv on Linux.
Netscape's huge differentiator was running on 16-bit Windows. All the other browsers needed either NT or a Unix workstation. This was prior to Windows95 so regular folks were still running Windows 3.1.
I remember running Mosaic on Windows 3.11 with WinSock in the 93/94 time frame. Maybe this was a 3.11 thing though... I never tried it with standard 3.1.
I don’t know. I was running it before 95 and never had NT. There may have been some random libraries like WinG and WinSock to install, but the lack of install libraries isn’t really an end user wowing feature.
That reminds me of the story of Trumpet Winsock and it's author Peter Tattam. Before Windows 95 you need 3rd party software to make a PPP connection to the Internet.
It was such a widely used and distributed piece of software - magazines would ship it on a CD attached to the cover which would usually contain Netscape as well.
Andreesen was hated at UIUC in the mid 90s for essentially taking the NCSA work and and then getting rich off of it. Then to add insult to injury he called his company Mosaic Communications. The university went after him for the trademark infringement (or whatever), but I don’t think there was anything legally wrong with him taking the code. Not only was Netscape, not the only company spun off of the Mosaic work, but pretty much every browser had some NCSA code in it for years and years.
Andreessen was loved by the entire core Mosaic technical team. Management did not like him, because he was a leader they could not control: opinionated, outspoken. The core Mosaic tech team all left Mosaic project on the same day to join Marc at Netscape, and pursue the pure vision of what browser should be. Mosaic project never recovered.
Source: I was Mr. MacMosaic.
I don't remember that, and can't find a source for it. The source code was released, but I don't think it ever had a license that allowed commercial use, and was it licensed commercially to many others.
Edit: including to Spyglass, for their browser, which was later licensed by Microsoft for Explorer.
I think Netscape was a clean rewrite of the source code of NCSA Mosaic, but with the benefit of past experience and being able to look at the code. Not illegal, but still somewhat dodgy, and understandable that people who worked on the NCSA version might get mad. I remember hearing that Andreesen was none too popular at UIUC after founding Netscape, and the lawsuit over the trademark; there's a source for that here:
I once took a programming class from a later (post-Andreessen)NCSA Mosaic team who would joke about Netscape coincidentally having the same bugs as NCSA Mosaic.
I remember trying it in the mid 90s and then not using it much. I remember it being quite slow. The lack of even rudimentary multi-tasking caused it to hang easily. I could see where it might be useful in a lab with 8086 and 286 machines. But I think even at the time, this software had what seemed like a high cost, like $99. If there had ever been a DOS-only port of Lynx (without requiring Win32), that would have been a better option instead.
Since the 486 I used could run Windows 3.1, Trumpet Winsock and Netscape and later Opera were good enough for the times. It was claimed that Opera was written in assembler (and it did feel like it was faster than other browsers) and it introduced MDI (tabbed) browsing.
Arachne's still (freely) available though it hasn't been updated since 2013 (and 2010 for the last Linux (curses) version). There is now a version for 386 and better which should be a bit quicker:
http://www.glennmcc.org/
I used MacWeb [1] on my PowerBook 145b in the early 90s. Lovely solid machine, 25 MHz Motorola CPU, 4 MB RAM, and a massive 80 MB hard disk. And yes, it was possible to surf the web with that, with information dense content, mostly text (the "T" in HTML and HTTP!).
rolls out 300bps acoustic coupler to slowly watch lynx update on a remote system....
Ok, so I really used 2400bps and later 14k4 all the way up to 56k (speed!) for IP based traffic. It didn't take long using 300bps to demonstrate the wisdom of upgrading.
Interesting to see Viola mentioned so prominantly here, and to see its impact on Andreesen, Mosaic, and Netscape. Whilst I'd never used it myself, I ran across it a year or so back when looking at how the n=modrn browser mechanics were estabished: navigation, history, bookmarks, scropting. Viola extended many of these concepts, and explored ground that's since been lost; Applets, scripting (Tcl), stylesheets, frames, forms, equations editor, columnar layouts, sidebar as integral element, and more.
I used the Atari browser CAB [1]. It had auto reload capabilities, you could have the browser open in a window in the background and it would reload the page while I was editing the html in the foreground window. Very handy for developing. Not sure about timeframe though, 98?
Cello was the first graphical browser I used after Lynx for a short while. Though I appreciated the graphics in Cello, I did initially find the Web (graphics or just text) to be messy compared to the much more orderly Gopher. Once I started dabbling with HTML and made a crude site for my employer, I quickly went from office clerk to webmaster. At that point Gopher was pretty much forgotten.
I also used Cello, on an early Windows NT beta. It was multi threaded and async, much nicer and faster compared to Mosaic. Problem was it was also very buggy and crashed all the time. After some time I returned to Mosaic that would run without crashing.
Lynx was widely used at my university when I started there in 1993. I remember vaguely not liking Mosaic and other early inline graphic browsers, not because they weren't revolutionary, but because they were too slow on our relatively slow networks and phone connections.
> So in his spare time, [Berners-Lee] wrote up some software to address this shortfall: a little program he named Enquire. It allowed users to create "nodes"—information-packed index card-style pages that linked to other pages. Unfortunately, the PASCAL application ran on CERN's proprietary operating system. "The few people who saw it thought it was a nice idea, but no one used it. Eventually, the disk was lost, and with it, the original Enquire."
Sounds like this was pre-NeXT, or at least on another in-house OS. It would be interesting to see what it was all about.
The Wikipedia article on ENQUIRE says it ran under SINTRAN III (the standard OS for many Norsk Data mainframes in the 1970s-1980s). So not a CERN-specific OS, just one not known to many people outside European users of mainframes at the time.
Having been involved in the BBS scene from the very early 80s and being involved with all things computer related since then, I have experience with all the browsers listed, and could name a few others.
I clearly remember using the web browser bundled with early NYC ISP The Pipeline in like 94. Netscape eventually came and blew it away, but it wasn’t bad for what (I think) was the work of basically one guy.
I had Mosaic as well as Spry. I always thought it funny that my wife had an old antique jar in the kitchen (still have it) for flour that was for a brand of lard called "Spry"
My first graphical web browser was Slipknot. From what I remember you would telnet into a shell account and it essentially acted as a graphical front end for Lynx.
(Realizing I miswrote that yesterday — obviously the point of Slipknot was that you would dial up to your shell account, because you didn’t have telnet access on your machine.)
Modern web pages have become a seizure-inducing cacophony of colors, fonts, animations, ads, videos and intrusive modals ("please sign up", "subscribe to our newsletter", "do you accept cookies?", "please disable ad-block", "people in <your town> are excited about this new weight loss product!") etc.