[vovanidze]

the nostalgia for 1999 isnt really about missing dial up or basic html, its about mourning the loss of user agency. back then the browser actually worked for you. today the client is basically a hostile enviroment running megabytes of third party js just to track telemetry.

going back to gopher or text-only browsers feels like admitting defeat tbh. we can still build incredibly fast modern apps if we just stop treating the users hardware like an infinite resource for adtech. you dont need massive frameworks and client--side bloat to make something good.

[jazzypants]

In 1999, hostile websites would pop up endless new windows full of advertisements that you were powerless to stop unless you simply held down "ALT+F4" or "CTRL+ALT+DEL". Part of Mozilla Firefox's appeal is that it came with a pop-up blocker. [0]

Do you know anything about the Browser Wars? People literally had to put up images telling you which browser to use if you wanted to actually experience their website the way it was intended. Otherwise, it was just broken. [1]

[0] https://www.nytimes.com/2004/01/19/business/as-consumers-rev... (sorry for the tracking code, but this is a "gift" article and it was the best source I could find on popup ads)

[1] https://en.wikipedia.org/wiki/Browser_wars

[franga2000]

Hostile pages did that. Today, nearly every page has a dozen tracking scripts, starts off with a cookie popup, probably pops up a "please log in" or "please give me money" after you scroll half way down, still has ads that even more effectively mimic the site topic and design to trick you into clicking them, pops up a newsletter or cupon code popup if your cursor leaves the viewport, might be secretly running experiments on you by A/B testing titles, images or testimonials...

The assault on your attention is way worse these days, it's just (mostly) contained to the viewport.

[acdha]

In 1999, most websites were not hostile because they weren’t chasing diminishing returns from ad-tech companies. Most of the companies I worked with were trying direct revenue models getting people to buy things or subscribe directly, and the ad market paid a lot better for less obtrusive ads - the whole real-time bidding process to run arbitrary JavaScript was yet to come.

Now almost everyone has pressure to find new revenue streams and maximize income while Google and Facebook have sucked up most of the revenue, so you see more and sleazier ads everywhere and sites which rely on you reading or watching are on a much more aggressive treadmill trying to constantly give you new things to see ads on so the experience is more frenetic.

It feels not unlike how in the mid-20th century people left work at work when they went home and only extraordinary circumstances would result in phone calls home, etc. whereas in this century it’s just expected that everyone carries a smartphone and checks email/Slack. More efficient in some ways, for sure, but a lot of stress ground out of people for no extra compensation.

[keepamovin]

You nailed it. Retreating to text-only kinda misses the whole point. The browser back then wasn't just a document viewer. It was a portal.

People forget that the internet used to be a place you went. It was entirely separate from our analog lives. You sat down, you fired up the machine, and your screen became this portal cut right into the universe. The juxtaposition between that visually-stunted, industrial-grade gray interface and the shocking immediate global access we suddenly had... it was the everything. The UI wasn't 'boring'. It was the clunky machine whose buttons you pressed (literally) to touch the world.

Today it's all hyper-lubricated feeds, and scammy-shiny UI trying to hijack your dopamine. But back then, the machine worked for you. It was a tool for discovery. A fucking frontier.

I've been trying to build a shrine to that precise feeling, to see if I can grab the modern web and force it to face it's beautiful glorious past - to that specific gateway-to-the-world, electronic frontier feeling. Just a small set of experiments. Incomplete as a monument to the totality of it. Merely a partial body of work, trying to express what it felt like to be there. I built a Win98/1999 environment: https://win9-5.com/desktop.html to browse the web from a (abominatively) multi-tab Netscape re-imagining. It runs a live, remote modern browser session inside a pixel-perfect 98 shell. Forcing the modern web through that dial-up era lens... it’s evocative + modem sounds. The aim is to remind you what it felt like when the web was a boundless horizon, not a walled garden of weirdo nimby's and microstates and regulatory capture etc etc etc. Sometimes I catch a flash of that fire again using it. Sometimes...

[1vuio0pswjnm7]

"going back to gopher or text-only browsers feels like admitting defeat tbh."

Interesting perspective

I have been using text-only browsers continually since the mid 90s (no lynx after 1999)

As such, I never "went back" to using the text-only browser as I have always used one, but as graphical browsers became worse I used them less

The customised text-only browsers I use today are 1.3 and 2.0 MB. I can compile them in seconds on underpowered computers

The so-called "modern" browser is [rapidly-expanding size] MB, not was easy to customise via editing the source code and takes substantial resources plus time to compile

Today, most www use for me is text-only. I am consuming information not graphics. I prefer textmode to X11 or the like

I avoid making HTTP requests to remote servers with graphical, Javascript-enabled browsers

I prefer using TCP clients and TLS forward proxies for making HTTP requests (at least one forward proxy now even has its own TCP client)

I use the text-only browser to read HTML files

For example, yc.htm is a file I create each day that contains all HN submissions where discussion is still open

Today yc.htm is 12 MB and there are 7268 stories

Using tiny command line utilities I wrote, the HTML in yc.htm is processed to CSV and added to an SQLite database. The unique domain names, today about 3704 of them, are extracted from all item URLs in yc.htm and DNS data is obtained via 1-3 pipelined lookups, each over a single TCP connection. The DNS data is then processed and inserted into an SQLite database

The TLS forward proxy stores the DNS data in memory so that I do not have to make any remote DNS when accessing the www

The yc.htm file can be opened in a text-only browser but I would not attempt it in a graphical one. The text-only browser feels more robust, less likely to stall or crash. I prefer the text-only formatting

By not using a graphical browser I would not say I am "admitting defeat". I have full control over HTTP headers, I only retrieve the data I want, I never see any ads, I do not send data to trackers or telemetry collection points

On the other hand, using a graphical browser does feel like "admitting defeat" as by doing so I allow "web developers" to destroy all preferences I have for how and when I want data retrieved and presented. If I allow the graphical browser unfettered access to the internet, and allow it to run unreviewed Javascripts, I lose all control over DNS and HTTP requests. For me, the experience of using a graphical browser where I have no such control is slow and painful, a horrible "user experience"

I do not see using "old" software as "nostalgia". I see it as being practical. "New" software generally sucks

The loss of "user agency" as some call it only occurs if one uses a so-called "modern" browser and runs Javascript. The seizure of user agency is accomplished by getting people to use a particular "user agent" that is controlled by online advertising companies (e.g., Google) or their business partners (e.g., Mozilla). If we called this a "choice" perhaps some readers might disagree. But so-called "Big Tech" have consistently argued that people "choose" Big Tech's "user agents" that, via their design and "default settings", effectively remove user agency for the majority of people who use them

[anthk]

You can use Dillo and explore the light web, gopher and gemini with no JS or big plugins at all.

Gopher and Gemini are simple scripts, so is the actions plugin, but I have my own written in 'rc', a shell borrowed from Plan9 and a bit improved (readline keys, history).

For audio/video I just spawn mpv+yt-dlp in the spot.

https://dillo-browser.github.io/

[1vuio0pswjnm7]

"(at least one forward proxy now even has its own TCP client)"

Correction: s/TCP/HTTP/

I also use a separate TCP client from the TLS proxy author