Whenever I see these old W3C docs mentioning old standards (MathML, in this case), it reminds me of how short-lived and little supported VRML was in browsers.
I remember a scant few months were it would run in maybe Netscape? And then there were a couple years it was supported via Java Applets before being completely phased away. At least, in the web space; maybe it found purchase elsewhere.
If you're looking for something similar still you may want to checkout https://aframe.io/ -- it's JS library(s) and you load them, then use custom tags in HTML to create VR worlds. it's rad.
I remember seeing VRML as a kid and I was mindblown. If I recall correctly, it ran in IE 4 with an add-on. I really thought, only a few more weeks and I can build games like Doom myself running in a browser.
Well, with my assessment I was off over a decade lol.
Last week I went to some meetup and met a young guy raving about every website is going to be VR in about two years. I didn’t even know about VRML.
I recall the biggest hindrance to VRML, aside requiring a plugin, was that most computers didn’t have hardware acceleration for 3D rendering. So performance was too poor to do anything useful for wide audiences.
Whenever I hear VRML, I am reminded of this Aptiva commercial from the 90's with the "3D San Francisco" site. Was always bummed as a kid that I never actually could find it online. https://www.youtube.com/watch?v=ROOSUiLEesA
Assuming you’re talking about MathML, there’s nothing outdated or short-lived about it. Browser support is reasonably good nowadays[1], MathJax always supported it, KaTeX by default outputs MathML alongside HTML, etc.
I didn't mean MathML specifically was unused, just that it reminded me of other niche/auxiliary standards that have been supported, many of which have been dropped.
I remember being like 15 and being excited about vrml. I think I knew one page with a working demo, and I used to tell people it was going to be the future.
That’s my VR dream. A federation of independent sites where each can have its own unique properties sounds like the way to go. I imagine my site having doors that link to other sites. Maybe when you use one, your browser caches a backlink so you can turn around and walk back.
I’d vastly rather play with this than any system operated by a single company.
Adobe actually tried something like that 20 years ago. The 3D worlds were even scriptable with Javascript and it all ran in the browser (with a plugin).
I still come across people using the term HTML5 as if it's "a thing". For example "our system has been updated to HTML5".
Whereas HTML5 spans more than a decade of continuous browser improvement and extension. It is an "era" or "mindset" not a "specification".
HTML5 was a way of discarding the mindset of "versioning the web" - now browsers implement specific functionality as they choose. The Web Serial API might be in this "HTML5 browser" but not in that "HTML5 browser".
The number is misleading in that sense. It implies a fixed moment in time. And it does that but only in a "before and after" sense.
HTML5 is actually an acronym, it stands for "Cool Shit That Won't Work In IE6". Unfortunately someone typo'd it a bit, so like Referer we're stuck with it now.
Actually, W3C published a number of HTML specs under the "HTML 5" moniker beginning in 2014 ([html5]) until 2017 ([html52]) based on (the loose group of people on github mostly working for Google on Chrome called) WHATWG's HTML specs. From 2020 on W3C again attempted to publish WHATWG snapshots, without redaction this time around; the January, 2020 snapshot ([html2001]) is the only one that made it into a W3C recommendation under that process, while later snapshots were rejected and simply no new snapshot was reviewed by W3C, Inc.'s HTML working group since (whose chapter is closing down these days last I checked). [is-this-html6] has the details, but in a nutshell, the so-called "HTML5 outlining algorithm" spec, one of Ian Hickson's innovations going along with "main", "header", and "footer" elements and the concept of sectioning roots that never was implemented/honored in browsers was removed by Steve Faulkner going to great lengths editing the upstream WHATWG HTML spec.
So in this sense, as is argued by the linked blog post, "HTML 5" refers to those W3C specs, while "HTML 2020" could refer to later specs.
Ian Hickson’s steadfast commitment to web standards and rejection of the xHTML bandwagon who did not understand the difference between catastrophic failure and silent failures and its implications accordingly build the foundation of everything we have now that is called web.
While others evolved into spec warriors, Hixi put users and usability first by looking beyond specs.
He is an unsung hero whose contributions made the difference. He faced so much hostility, yet was the only one who maintained and understood every aspect of these specs. One after another he won open standard evangelists over.
It is like Linux vs Windows and not Google vs the rest.
"The reality is that for all of the work that we've put into HTML, and CSS, and the DOM, it has fundamentally utterly failed to deliver on its promise.
It's even worse than that, actually, because all of the things we've built aren't just not doing what we want, they're holding developers back. People build their applications on frameworks that _abstract out_ all the APIs we build for browsers, and _even with those frameworks_ developers are hamstrung by weird limitations of the web."
We can pretend that the W3C eggheads were up there in ivory towers with no understanding of how the web should move forward, but the reality is that a lot of modern web dev is tree like data structures being sent across the web to be rendered with client side templates, exactly what they were aiming for with XML and XSLT.
Namespaces, custom elements, client side templating and more, the web had it all and threw it away so that we can instead spend the next century reimplementing worse versions of what XML provided out of the box.
Yes and no. Yes Chrome (and Chrome derivatives) have a lot of market share. But Safari is important if only because of iDevices. And Firefox is still I the mix - single digit market share still adds up to a lot of users.
HTML5 was the last time it had a marketing version number. Henceforth, it was agreed to be just HTML. Now, it is like people saying Web 5, Industry 7, or something like that. It is no longer “the way” and is more of a cringe. Saying Web 2.0 was kind of like more of a big upgrade signifying a leap and a lot less about versioning.
I’ve fond memories of the term “HTML5”. I sold the domain html5.in to an agent who was buying for Microsoft. The Microsoft part was discovered later after I sold it for single digit thousand dollars.
No, the evergreen WHATWG HTML is the living standard. HTML5 is no more the living standard than is HTML4, both of which were by W3C. And after years of battle W3C officially forfeited the race, thus making the WHATWG living standard to be the de facto winner.
This is somewhat true, but where do we stand a decade after we got HTML5?
Most browsers use Webkit/Chromium under the hood (because nobody wants to compete in the race anymore), and building a new browser from scratch seems impossible.
I wonder if there is a causality between ditching versions and browser vendors exiting the development of their own engines.
For a bit I think HTML5 was a bit of a short-hand for 'we're doing modern, javascript heavy app'. Now I think HTML5 on a job description is just a default, with the '5' void of meaning. There's only HTML5
Does anyone know where I can find old HTML4 templates that used to be all the rage back in the day? (Similar to the few that can be seen here https://www.bryantsmith.com/template/, or used to be provided on hpage.com)
The terms you would be looking for are "PSD layouts", "sliced layouts", "website templates PSD", "table-based website templates", etc.
Basically these are all websites that were fully designed in something like Photoshop or PaintShop Pro and then manually sliced into graphics that would then be finagled into a table layout manually (with zero borders). Usually with coord-based anchor links that were super finicky, especially if you wanted hover/click effects (usually achieved via HTML+JS versus CSS). Certain portions of the layout were designed to be repeatable for the content-based section of the website, so that would be a <TD> with a fixed width but non-fixed height and background-image with an x-repeat (again, usually in HTML tag properties versus CSS).
Nowadays, it would probably be easier to find a graphic designer to just make you one to your specs and "slice"+table it yourself. It would certainly use better/more modern HTML/CSS.
Also integral to this kind of design was the spacer gif. A 1x1 transparent gif that you’d give different width and height values to in the img tag, in order to create spacing in the page.
I built an early site using those techniques in oh, 1996 I think? (stomped.com) Using zero border/padding tables was a clever way at the time of accomplishing a fancy graphics tab/menu without resorting to calling a CGI script.
Oh this brought back memories! Getting a PS file, using px for everything to "match the designs", random jQuery snippets to get a hamburger menu. Not a damn clue what I was doing but loving every minute of it.
I remember from that time those who where trying to do the "semantic web", some kind of complexity abomination... and in the end we got the current scripted gigantic and grotesque DOM abomination locked up/financed by big tech. Yep, I considered that was a loss/loss.
I recall too how the major web browsers did everything to make xhtml dev a massive pain (that was clearly sabotage). In the end, you have to choose doing html "à la xml", or the html "à la sgml" (which is seriously ugly to parse).
The only way to cleanup that big tech lock-down/mess: regulation on critical(utility) web sites with noscript/basic (x)html (like they mostly were a few years ago).
Can we get a clean "better"? Well, if it is as simple (complexity _and_ size) to implement than html parsing, and as stable in time, yeah. As far as I know, it does not exist.
They didn't implement it as anything resembling a system widget.
What every web developer wanted from 1997 on was the common nav bar of the era: a set of dropdown menus that more or less behaved like the native dropdowns every other piece of software on the machine had. What we got was a <ul> with a funny name, so you had to at least toss a bunch of CSS at it to remotely resemble a "menu", and if it was on something like IE, also JavaScript to breathe it to life.
I always believed if we had added a few more native-widget tags in the 1990s anything-goes era (a <menu> that looked the part, <td sortable="type">, maybe <richtext>, an <include> that was better than iframes), we could have had web based apps earlier, less Javascript to provide trivial functionality, and steered the web into a direction where the north star was the look, feel, and standards of native software.
If quality web apps followed the OS Human Interface Guidelines and looked and active as native as possible, it would have created a completely different evolutionary environment. Would we end up prioritizing things like structured data because standard design practices were already displaying things in familiar structured widgets?
I remember a scant few months were it would run in maybe Netscape? And then there were a couple years it was supported via Java Applets before being completely phased away. At least, in the web space; maybe it found purchase elsewhere.