Some random asides on this slice of nostalgia:

> Note that the typical display size at the time was 800x600 pixels, so this and other websites would likely have been designed for those dimensions.

This was before responsive design existed. First we designed for 640×480, then we designed for 800×600, then we designed for 1024×768. Bad developers would design for wider viewports and leave people with smaller screens to scroll horizontally to see everything. Slightly better designers would design for the narrower viewports and leave huge gutters down either the right side or both sides for people with wider screens. Best practice was “fluid design”, where you would define widths in percentages to adapt to the screen width, but it was difficult to get designers on board.

> But if the web was a “consumer playground” now, it was still one with many constraints. As Zeldman told budding web designers, “the accepted wisdom is to use as few images as possible, and make them as small as you can (small in file size, though not necessarily in height or width).”

It wasn’t just file size. The early web was limited in terms of colours too. There were 216 “web safe” colours.

> the book advocated for “hacks” to HTML in order to make websites more visually appealing. The primary hacks were using invisible tables and single-pixel GIFs to help control layout.

There were a lot of weird hacks. One was to put many <title> elements in your document, and Netscape 2 would flip between them in the window’s title bar to make a crude animation. The title bar because browsers didn’t have tabs back then.

> CSS support from the two main browsers at the start of 1997 was patchy at best. Internet Explorer 3.0 was the closest to supporting the W3C standard for CSS, but it was buggy and inconsistent.

It was basically nonexistent apart from very minor things. Internet Explorer 3 didn’t even understand the em unit and just treated it as pixels, so if you set something to font-size: 1.5em, it wasn’t 50% larger than the parent element’s text, it was invisibly small.

> As for Netscape, its 3.0 browser had poor CSS support. In fact, the company even tried to create an alternative to CSS, with a JavaScript-powered styling mechanism called JavaScript-Based Style Sheets (JSSS).

Netscape 4 transcoded CSS to JSSS on the fly, which had the side-effect that when you disabled JavaScript, it also disabled CSS.

> For all their differences, CSS and Flash did have similar goals: both aimed to expand the state of web design on the web.

Before web fonts were supported by browsers, one fairly common technique was sIFR, which looked for specially marked up text on the page and replaced the text with Flash applets rendering the text in an embedded font. It was pretty ugly loading and caused bunch of problems, but the designers didn’t mind as long as it let them use custom fonts.

It was a pretty hellish time to be a web developer, but exciting as well. The browser bugs and incompatibilities were a thousand times worse than they are today and could really ruin multiple days at a time on the most trivial stuff, but it was also a period of great inventiveness and variety.