I was browsing the internet the other day and came across a link to the first website on the Web: http://info.cern.ch/hypertext/WWW/TheProject.html. Go ahead, click it. Believe it or not, it still works perfectly in modern browsers a thousand years later!
The early WWW has some interesting browsers to play with, e.g. Amaya (https://www.w3.org/Amaya/) that had way more edition options than modern Browsers.
At the time TBL was working on the WWW proposal SGML/DSSSL was a pretty big item to rival HTML for markup. You also have a lot of other markup in common use in the target audience of physics at CERN, for example LaTeX, which is still very common for the superior math support. So a lot of the existing documents might have been TeX. Even markup like groff wasn't unheard of. So keeping open for multiple markup formats was politically wise. This was the time before even PDF existed, so at best you might have Postscript or EPS or the source markup.
The modern web is not all HTML either, you have CSS, the universe of Javascript, all the embedded XML sub languages like SVG, MathML and others. They have just been absorbed under the ever changing HTML label.
In some parts, WWW was actually worse than the already existing Gopher/Veronica hypertext and search engine system, as the search stuff was an afterthought here.
"The idea that the web is not hierarchical is important in the proposal. Seems obvious now, but I wonder if this we a key part that no one had figured out before."
Not really, the fundamental idea of HyperText as a non-hierarchical "web" of information predates this by many years, and is probably why TBL refers to it so frequently in the proposal.
The early WWW has some interesting browsers to play with, e.g. Amaya (https://www.w3.org/Amaya/) that had way more edition options than modern Browsers.
At the time TBL was working on the WWW proposal SGML/DSSSL was a pretty big item to rival HTML for markup. You also have a lot of other markup in common use in the target audience of physics at CERN, for example LaTeX, which is still very common for the superior math support. So a lot of the existing documents might have been TeX. Even markup like groff wasn't unheard of. So keeping open for multiple markup formats was politically wise. This was the time before even PDF existed, so at best you might have Postscript or EPS or the source markup.
The modern web is not all HTML either, you have CSS, the universe of Javascript, all the embedded XML sub languages like SVG, MathML and others. They have just been absorbed under the ever changing HTML label.
In some parts, WWW was actually worse than the already existing Gopher/Veronica hypertext and search engine system, as the search stuff was an afterthought here.
"The idea that the web is not hierarchical is important in the proposal. Seems obvious now, but I wonder if this we a key part that no one had figured out before."
Not really, the fundamental idea of HyperText as a non-hierarchical "web" of information predates this by many years, and is probably why TBL refers to it so frequently in the proposal.