And by Web server I mean actually web server (not BBS), meaning it had to run on the actual IP internet and be browsed with a dedicated browser program.
So what kind of tools were used in the late 80s/early 90s when some common households started browsing websites?
How was the HTML generated? Were there any tricks to save bandwidth that have been lost to history?
What are some fun anecdotes?

  • 2
    The web server is a different thing from the tools used to create the web site. At base, all a HTTP 1.0 web server has to do is to receive a request and return the requested document, which in those days was generally a file from a filestore. You could write one of those in a day, especially if a single-threaded server could handle your expected load. Feb 15, 2020 at 18:36
  • 12
    One trick to saving bandwidth, since lost to history, was not using megabytes of javascript and css and images... Feb 15, 2020 at 18:41
  • 1
    I hand-edited my first webpages. For repeating content, I wrote a bit of automation to generate the webpage from several sourcefiles (using make), then hand-edited the source files. This technique still works today; makes for small webpages with no Javascript at all. I also wrote a "forum" using CGI and a few shell scripts. Comments by users were just stored in a plain text file (where I could edit them, or delete them, if necessary).
    – dirkt
    Feb 15, 2020 at 19:02
  • 1
    As far as how HTML was generated, wasn't the original browser concept supposed to include and HTML editor? IIRC, Tim Berners-Lee's browser did just that, as did early versions of Netscape. I'm not saying they generated very good HTML code, but a talented amateur could get a page to look like what they wanted, no HTML knowledge required. (I'm not sure why this was asked in a question about servers but thought I'd comment on it anyway.
    – RichF
    Feb 15, 2020 at 23:37
  • 2
    Also, you ask about webservers but then mention generating HTML which has absolutely nothing whatsoever to do with webservers. Also, at least in Germany, browsing the web was definitely not common in the early 1990s, not even in the mid-1990s. Of the very few people who were "browsing" anything at all, they would have used CompuServe, AOL, MSN, or even interactive Teletext over protocols such as Datex-P (based on X.25). We were a very technically advanced household, but we didn't have Internet or WWW until 1996 or even later. Feb 16, 2020 at 9:16

3 Answers 3


I guess this is fairly straight forward. Early web servers were using Tim Berners-Lee's original web server code as it was published in 1991.

I myself downloaded the code and ran web servers on our institutions computers from that start in 1991/92. I found and fixed a few bugs in the code also.

How was the HTML generated? I like others, just wrote the web page HTML in a text editor. Myself I used emacs with a suite of emacs macros I wrote.

There did not need to be tricks on saving bandwidth as the HTML pages were mainly small text files, but we knew how to optimise the images for display. There was little in the way of multimedia as we experience it today. I had dynamic and responsive pages using the CGI mechanism that invoked perl scripts to generate the dynamic content. That fed an early genealogy database with 50,000 entries.

At that time it was not a radical innovation as there were also gopher servers and ftp servers and other information sharing protocols (such as WAIS, Archie, Veronica) that we were all using at that time.

What made a change was the creation on Netscape browser that was much easier to use than the text based WWW browser that came with the original CERN code.

  • 4
    It’s probably worth mentioning NCSA httpd (which eventually became Apache httpd) along with NCSA Mosaic, and NaviServer (which eventually became AOLserver); IIRC CERN, NCSA, and NaviServer served most HTTP traffic in the early days. Feb 15, 2020 at 18:31
  • 1
    Between text-based and Netscape there was Mosaic. Feb 16, 2020 at 11:25
  • @MartinArgerami Yes: you are correct. It was my memory that was failing here... Feb 16, 2020 at 13:29

What were some of the earliest common webservers and how did they work?

By delivering a file? A basic webserver (at that time) did simply take the path supplied with GET, appended it to some base directory, tried to locate that file and transferred it (as is) if successful. Otherwise an error was generated. It was common to write simple servers as programming assignments (*1)

Early servers were open

And by Web server I mean actually web server (not BBS), meaning it had to run on the actual IP internet and be browsed with a dedicated browser program.

Interesting way to cut it sown, but then again, this description would as well work for other systems of the same time like GOOPHER and WAIS, wouln't it (*2)

So what kind of tools were used in the late 80s/early 90s when some common households started browsing websites?

There isn't much web, as we know it (aka HTTP based HTML delivery), in the late 1980s, HTTP/HTML wasn't introducred until 1990. Of course there was the internet, mainly build up of FTP, UUCP and TELNET.

The web also didn't become a thing even in IT circles until Mosaic became available in 1993. And it wasn't until Netscape started to do the Navigator in 1994/95. Common households a that time used proprietary systems like AOL and Compuserve. And it wasn't until the ver late 1990s that the web did make a notable inroad into the 'common household' market.

How was the HTML generated?

Using an editor? Like Notepad? Yes, serious. That or any other editor, after all, HTML is simple on purpose.

Netscape introduced a first WYSIWYG type editor as part of it's Navigator 3.0 Gold in 1996. Not anything home users would afford.

Were there any tricks to save bandwidth that have been lost to history?

Why? At that time pages were mostly (text) content - not 99.8% HTML - there wasn't much to save. They come quite fast even on a 14.400 modem. And basically instant on anything faster than that.

What are some fun anecdotes?

Want to hear one of today? It's still possible to do such lean pages. Use Notepad and don't try t force the browser - let it just do its job. Funny, isn't it?

*1 - And it still is, IMHO, a great little task to get familiar with TCP/IP programming.

*2 - In fact, for some short time WAIS was much more successful on the internet than the WWW

  • I'd argue that more webpages at the time were written using vi or emacs, not notepad.... (and that might very well be the case today) Feb 16, 2020 at 10:37
  • @RadovanGarabík Won't argue against. So, as I have used ed for my very first, take it as a placeholder for the most primitive editor you can think of - after all, everything that can produce a bare ASCII file could and most likely was used.
    – Raffzahn
    Feb 16, 2020 at 11:34
  • 1
    I wrote my first web pages using the EVE editor (without any HTML customizations, just plain-text mode) on VAX/VMS :-) Feb 16, 2020 at 15:31

Early web servers include:

  • CERN httpd, written by Tim Berners-Lee, Ari Luotonen, et alia (Version 0.1 was released in June 1991)
  • Plexus, written in Perl 4 by Tony Sanders of BSDI (early 1993)
  • GN, which was both a web and gopher server, written by Dr. John Franks at Northwestern University
  • NCSA HTTPD (later also known as W3C httpd), written by Robert McCool, et alia at NCSA/UIUC (First released in 1993)

There were surely other early HTTP servers but the above people were named dropped as contributors to the initial CGI spec.

Not the answer you're looking for? Browse other questions tagged or ask your own question.