Anyone who followed browser development back then (1994) knows how it worked. Decisions about what went into browsers were based on how easily they could be implemented as much as what others were doing.
Yes, CSS was discussed in 1994. So was VRML (I think) and god knows how many other things. I mean, the Cello browser didn't render JPEGs (!).
Browser authors made stuff up that they thought was useful. The usually did this by adding tags, because that was a lot easier than creating a whole other language like CSS
Let's see. 1994. If you've got a 66MHz Pentium on your desktop, you are living large and cutting edge [1]; most are still on 486s. If I recall correctly, we're still talking single-digit megabytes of RAM being the most you could reasonably expect. Browser authors are still feeling through the hardest text layout problem since TeX, and browsers are expected to do it realtime, on processors that today we'd consider low-medium end embedded processors. A problem still giving them fits, I might add; how many of us still feel our browsers render slowly even with machines a hundred times more powerful? Yes, we don't render the same pages, but that just proves my point more, the desire to do these computationally expensive things was there in 1994 but a page that takes 1 second to render today is, well, not even guessable back then because you probably ran out of RAM.
It wouldn't have mattered if they had suddenly had an attack of the stupid and tried to implement the hopelessly-idealistic-for-1994 CSS spec. It would have sucked anyhow. It would have been slow and buggy and basically ruined the entire idea of CSS for years. It's better that they didn't try! There's no way they could have functioned with anything other than "Can I make this work?" and "Does it completely hork up my rendering engine?" being their two basic top priorities.
Always fun to step back and remember just how bad we had it all back then. How we'd have killed for a $300 netbook!
IE 3 which was released in 1996 had (partial) CSS1 support. I remember it being as fast as Netscape 3. (But that might be because bandwidth at that time were the bottleneck - rendering speed were not an issue, since it were still rendering faster than than the data arrived).
Netscape took a long time to embraces CSS. I guess "not invented here" was a big reason, because they actually designed their own CSS alternative: Something called JavaScript Style Sheets, which was somewhat like CSS, but used JavaScript syntax.
It means that Netscape ignored the idea of separating markup from style, probably setting back the progress of the web by some years. This seems to imply that Andreessen's vision of the web was not very ambitious at the time - mixing FONT and CENTER tags is convenient for small-scale publishing by individuals handcoding one-off HTML, but is unwieldy for larger scale ideas.
Anyone who followed browser development back then (1994) knows how it worked. Decisions about what went into browsers were based on how easily they could be implemented as much as what others were doing.
Yes, CSS was discussed in 1994. So was VRML (I think) and god knows how many other things. I mean, the Cello browser didn't render JPEGs (!).
Browser authors made stuff up that they thought was useful. The usually did this by adding tags, because that was a lot easier than creating a whole other language like CSS