I wrote HTML in the 90s. Modern standards like flexbox are objectively better than the float hacks and tables we used before. The geocities aesthetic is cute but it is extremely limited.
The web is now a competitor for native apps. That would never have been possible without the fast pace of innovation. Don't knock it.
I was last a "web developer" almost two decades ago, but dipping back in on a few occasions I am always appreciative of how much innovation has happened since then.
The world before the huge investment in browser technology was dark. Tables and spacers for meaningful layout and flash or shockwave for anything interactive.
I remember a time when css based drop down menus were seen as some sort of state of the art.
Yes, presenting a large catalog of products (a few hundreds), for discovery purpose an efficient menu is still a big challenge in term of UX and technical implementation all the more when portability, accessibility, and cross-devices is taken into account.
Things that definitely look like trivial banality at shallow level often end up to need a lot of attention on many concurrent details.
Uh, a guess is that 1+ billion people are already good at using "drop down menus" along with check boxes, radio buttons, single line text boxes, multiline text boxes, push buttons, links. So, when those user interface controls are sufficient for the purpose, using something else might reduce the collection of happy users. The Web site of my bank stays close to such now classic controls.
I'm talking about a time when investment in browser development and web standards was so lacking that being able to achieve things like this blew everyone's mind:
Maybe! My thoughts were, say, tangential or incidental.
A guess is that a central issue is how much in new features should we develop and use?
I see a dilemma: (A) I mentioned the old controls that go back to early Windows and even IBM's 3270 terminals. An advantage of these controls is that lots of software tools implement them and billions of people already understand them. (B) Being too happy with the old stuff or even the present risks progress that is possible and worthwhile.
Your post seemed to illustrate (B).
But generally in the industry, with smartphones, laptops, desktops, Apple, Google's Android, Windows, browsers, apps and extrapolating, we could have an explosion of new features that would complicate work for everyone and fragment the industry.
Ah, maybe Darwin would explain: Lots of mutations with only the best lasting??
For my work, I'm thrilled with the tools and technology available now that I get to exploit.
CSS grids are pretty nice, flexbox is ok, float hacks were fine and an improvement over table shenanigans. On the other hand I quite liked the simple hbox/vbox explicit elements that things like ActionScript + MXML had (Flex). I liked Flex overall quite a bit, even if it was just another ill-fated attempt at freeing us from the browser strangleholds like Java applets and the rest. Having native platform functionality and a bunch of other nice things readily available now (barring Safari, especially mobile Safari, holding everyone back worse than IE6 did) is nice, but it doesn't quite feel like innovation when much of that was available via plugins back in the day.
We have to compare apple to apple here. What was the state of native applications back then?
The main point that we could derive from this is that it's hard to make predictions, especially about the future, and all the more when geopolitics is involved. But still it's fun and sometime inspiring.
I've never understood the hate for table layouts. They literally just make sense. And now all they advanced css frameworks have basically just recreated table layouts via divs with row and column classes. I get the need for responsive designs but I still think we could have gotten there with tables.
It's like people got mad that tables were being used to for something other than strictly tabular data, so they recreated the idea behind table as a layout tool with "css grid" and made it 50x more complicated.
I wish web design could follow like woodworking where the most focus is on using the base tools very effectively. The introduction of new tools is mostly frowned upon. Of course that's all because of the inherently dangerous nature of using power tools. Regardless of tech stack you aren't to likely to lose a finger from coding.
The move from table layouts to divs+css was based on the idea that mixing markup and layout/styling was bad. Then a few years later as everyone moved to React et al. it was good again.
I understand the sentiment, but disregarding the nuance does the situation a disservice. HTML/CSS serves two separate but related use cases: a document layout and display language, and as a display layer for applications. I remember Pete Hunt's talk "React: Rethinking best practices" [1] explaining in 2013 why the styling separation of concerns for document display doesn't make sense for applications. Has opinion on best practice se-sawed back and forth? No, we've merely gone from web content being document centered to being massively application centered, and the discourse on best practices follows that proportional shift.
Would it be better if there was a different web application display technology, not retrofitted on top of HTML/CSS? Like maybe, but HTML/CSS is... fine. Even separated from the success of Javascript, it's an archetypal example of "worse is better" [2] leading to market success.
Yeah, a switch in mode between "make an XML document and then make something else to present it" (I think suffered massively from never actually being completely achievable with CSS, let alone comprehensible) and "just make a UI toolkit for javascript" (where in a full webapp world, the underlying documents are usually JSON and the presentation layer is javascript using HTML+CSS to talk to a layout and rendering engine).
Layouts implemented with tables make sense when you need to place elements once, and not a ton of them at once. Making even slight changes to your tables within tables within tables is such a nightmare, that throwing out all of your code and starting from scratch might just be preferable. Tables are not a basic tool for this task, it's a severe misuse of the tooling, caused by a lack of alternatives at the time.
> I've never understood the hate for table layouts.
Then I would say you have never understood HTML. Using tables for layout conflates content/structure/semantics with presentation and this is a problem in general with the HTML standards prior to HTML4 Strict. The reasons why this is bad have been expounded upon for near thirty years and are well known. HTML tables are not a “base tool” for layout.
A good litmus test for bad Web-design is “does it fail on screen readers?”. Indeed this does; it makes pages near unnavigable.
The web tried to be a competitor for native apps by offering technical parity but it wasn’t enough. Web versions of serious apps tend to be broken and have a banner asking you to download the native app. You can argue about why it happened, but it happened.
It most certainly is not. Web apps still suck ass compared to native. It's just that users are willing to accept even the crappiest solutions because they don't have very significant needs.
The web is now a competitor for native apps. That would never have been possible without the fast pace of innovation. Don't knock it.