Well, for one E4X was never adopted by other browsers that mattered, only Mozilla.
Second, it's main use was for working with XML, which the web is glad we got rid of.
Had the browsers properly adopted XHTML and the accompanying standards, and today we could already enjoy something like XAML on the browser instead of having only Chrome supporting WebComponents with workarounds like ShadowDOM to preserve local changes.
I mean, I get that it's extensible thanks to XML, but that doesn't mean browsers would actually have an incentive to create and implement such extensions.
https://www.w3.org/standards/xml/components
Basically Events, Modularization, Fragments, XForms, XQuery and possibly other parts that were still being worked on when the HTML 5 hype started.
And were we are, yet to have a Web UI designer that can match Blend, Qt Creator, Netbeans Matisse, Scene Builder, Glade, Delphi, C++ Builder,....
When I do web development I always feel like I am stuck with something not better than Notepad for GUI programming.
It's sad and surprising how much time and energy was wasted on XHTML. Back in february I met Steven Pemberton (XHTML spec lead back then, and of ABC/Python fame). He's just such an inspiring guy to talk to, but unfortunately, there was no time for recapitulating the XHTML situation.
Huh, my experience was all those things are terrible (spent some time with Qt and researching Glade, also inherited a Delphi codebase full of spaghetti code at one point).
Visual UI tools are great for prototyping, awful for maintainable code. Anyone I know who's spent any time with them ends up wanting to go back to the code.
There certainly are good Web UI designers [1], it just seems no-one wants them. There are successful visual Web UI prototyping tools tho.
[1] http://macaw.co/
Trying to manually hack GUI generated code is an anti-pattern.
One has to leave the code generated by the tools to the tools, everything else should live in other code files.
Sadly I never heard of Macaw, specially on the enterprise circles I move on, in any case they seem to be gone now.
If you see value in XML, you really should look into XML's big sister SGML. It gives you downward compat with XML, tag inference, type-aware (injection-free) variable and macro expansion, custom Wiki syntaxes (markdown and others), an integrated stylesheet language without new ad-hoc syntax, full HTML 5 parsing with short forms for attributes etc., and more. You might actually like it.
However this is meaningless if the browsers keep being a document engine, full of workarounds to translate a mix of document tags, coupled with some generic <div> and <span> into some kind of general purpose GUI layout engine.
Having to find the proper incantation of CSS3 transforms to portably trigger GUI acceleration is a good example of such hacks.
Even if it is yet another hack, webcomponents + shadowdom looked it would be the solution, even if a bit hackish, but it is Chrome only.
My favourite of the failed web technologies was VRML. Back in the 90s I used to create epic 3D landscapes that would render inside the browser much like how some experiment with WebGL these days. But this was long before it was 3D accelerators were common inside PCs (they were still very expensive so most graphics cards software rendered 3D models). VRML I think was a victim of being too ahead of its time.
The web was exciting back then. Nobody really knew what you could or couldn't do with it. These days I feel we've lost something with all these bloated frontend frameworks which act like magic boxes in that an alarming number of frontend developers don't understand how their code executes. But I guess that's what happens when science or technology becomes business.
FrontPage was a great educational tool, sure it produced crappy HTML, sure it was IE biased, but it made you learn, it was fast and came bundled in Office. Countless numbers of hobby web pages were build thanks to it.
Mostly because of cargo cult -- as Dreamweaver produced very clean HTML output.
That was my point. If you're familiar enough with HTML to use Arachnophilia (which, if anyone isn't familiar with that particular editor, was basically just the Notepad++ of it's day) then Dreamweaver would just seem excessive. Hence why I'd recommend it for designers rather than developers ;)
> as Dreamweaver produced very clean HTML output.
Cleaner than Frontpage and most of the other design tools out then, sure. But my experience of Dreamweaver around that era was that it's HTML output still needed a lot of manual refactoring in a text editor afterwards. While it was better than most GUIs it was still a long way off "very clean HTML output".
I'm sure things have improved significantly in the last 15/20 years though. But that was certainly the state of things back in the late 90s / early 00s.
As an aside note: I won a Source Code Planet (anyone remember them?) competition one month with an entry I made in Dreamweaver. It was a mockup of a Window 95 or 98 (I forget which) desktop with functioning start menu. I always felt naughty for winning based on something I built in a tool that auto-generates a lot of the code for you but then other projects I released were a lot more credible (eg DirectX games) but far less popular. I guess it just goes to show how little most people cared for code quality even back then.
GitHub is the current thing, PSC/Sourceforge/GoogleCode are still technically (at least read-only) around.
HTML 4 Transitional used to be very popular back in the 1999/2000. The XHTML movement was a weird hype, I tried it out but TinyMCE/WYSIWYG-editor spit out old HTML3 code at that time, so I tried XHTML 1 Transitional, but used .html extension as IE6 didn't support .xhtml. Anyway XHTML2 was a trainwreck, they were nuts to propose an completely incompatible syntax, a failed ideology and I switched back shortly after to HTML4 (2004). There was also a weird short hype around a new incompatibility Javascript version 4, short E4X. XML used to be everywhere, and some nuts tried to make XML part of JavaScript syntax - Mozilla and Adobe were into this crazy land. Thankfully it died, and there was never a JS4, but the JavaScript 5 strict was superb. CSS was okayish since IE5.5 and especially IE6 and MozillaSuite. Then there was the hype to not use tables but do everything with CSS divs. I quickly learned that a few tables and using CSS for everything else worked well in practice and gave one a responsive design long before it had a name. Loading data asynchronous used to be possible with XML on both IE56 and Firebird/earlyFirefox already in 2003, I found only two sites back then that had documentation for that obscure API back then, but it worked great - made a CD-ROM based elearning software written as one-page-app (HTML4, CSS and JS) with loading data from XML files and showing text, pictures and multimedia videos (Flash MX, before there was FLV format), that was two years before it became well know as AJAX (now XHR).
https://en.m.wikipedia.org/wiki/XHTML#XHTML_2.0
https://en.m.wikipedia.org/wiki/ECMAScript_for_XML