Under-investment

Print

It's almost absurd to me that printing and print related APIs have the problems and concerns that they still do given that so much of enterprise and government are web based. For example: Will your images be loaded? Who knows! Did you know there is a .print() and it doesn't act the same in several respects as choosing print from the menu? Shouldn't the browser support many of the CSS based features that print pioneered? Like... paging? Or at least actually investing in considering it in the browser at the same time could have helped us determine if those were even good ideas or shape APIs.

Accessibility

In theory all of the processes are supposed to help create standards and browsers that are accessible - in practice, we miss on this more often than is comfortable to admit. This is mainly because - for whatever reason - so much of this, from reviews to testing to standards work in designing APIs in the first place, is largely done by volunteers or people disconnected from vendors themselves and just trying to keep up. My colleague Alice Boxhall wrote a piece that touches on this, and more.

Internationalization

Probably in better shape than accessibility in many ways, but the same basic sorts of things apply here.

Testing Infrastructure

The amount of things that we are incapable of actually testing is way higher than we should be comfortable with. The web actually spent the first 15 years or so of its life without any actual shared testing like web platform tests. Today, lots and lots of that infrastructure is just Google provided, so not community owned or anything.

SVG

After some (like Amelia) doing amazing work to begin to normalize SVG and CSS, the working group effectively disbanded for years with very little investment from vendors.

MathML

From its integration in HTML5 until today, almost none of the work done in browsers has been by the browser vendors themselves. Google is the only vendor who has even joined the working group, and not so much to participate as an org as much as to allow someone interested on their own to participate.

Web Speech

Google and others were so excited to do this back in (checks watch)... 2012. But they did it too early, and in a community group. It's not even a Recommendation Track thing. I can easily see an argument to be made that this is the result of things swinging pretty far in the other direction - this is more than a decade after W3C had put together the W3C Speech Interface Framework with lots of XML. But meanwhile there is simple and obvious bugs and improvements that can and should be made - there is lots of be rethought here and very little invested from then till now.

Study of the web (and ability to)

The HTTPArchive and chrome status are about the best tools we have, but they're again mainly Google - but even other data sources are biased and incomplete. Until 2019 the study of elements on the web was just running a regexp on home pages in the archive. Until just a year or two ago our study of CSS was kind of similar. It just feels like we should have more here.

Polyfill ability for CSS

A number of us have been saying this for a long time. Even some small things could go a long way here (like, just really exposing what's parsed). After a lot of efforts we got Houdini, which should have helped answer a lot of this. It fizzled out after choosing probably the least interesting first project in my opinion. I don't know that we were looking at it just right, or that we would have done the right things - but I know that not really investing in trying isn't going to get it done either. To be really honest, I'd like a more perfect polyfill story for HTML as well. Once upon a time there was discussion down that road, but when <std-toast>-gate happened, all of the connected discussions died along with it. That's a shame. We are getting there slowly with some important things like custom media queries and so on, but a lot of these things we were starting to pitch a decade ago.

Protocols

The web has thus far been built a very particular way - but there are many ideas (distributed web ideas, for example) which it's very hard for the web to currently adapt toward because it's full of hard problems that really need involvement from vendors. I'd love to see many of those ideas really have an opportunity to take off, but I don't see good evolutionary paths to allow something to really do that. We had some earlier ideas like protocol handlers and content handlers for how this might work. Unfortunately content handlers were removed, and prototcol handlers are extremely limited and incomplete. Trying to imagine how a distributed web could work is pretty difficult with the tools we have. Perhaps part of this is related other items on my list like powerful features or monetization

"Web Views" / "Powerful features"

A ton of stuff is built with web technology as apps to get around some of the things that are currently difficult security-wide or privacy wise in the browser itself. Maybe that's how it should be, maybe it isn't. I'm not here to say "ship all the fugu stuff" or something, but it it definitely seems silly that there aren't some efforts to even think "above" the browser engines and standardize some APIs, a bit in the vein of what is now the Winter TC. What people are doing today doesn't seem better. I guesss there is a common theme here that I'd like to really invest in finding better ways to let the platform evolve a bit on its own and then pick it up and run with it.

"monetization"

I mean, this is a really tough one for so many reasons, both tehnical and political, but I just don't see a thing that could have bigger impact than a way to pay creators that isn't limited to ads and a way to fund new ideas. It just seems at the very core of a lot of things. I put it in quotes because I don't mean specifically the proposal called web monetization. There are lots of other ideas and a wide range of attempts happening, some of them seem less directly like money and more like ways to express licencing agreements or earn discounts.

Maps

We seem to have mostly just written off maps entirely as something which you just rely on Google Maps or Apple Maps for. That's a shame because there has been interest at several levels - there was a joint OGC/W3C workshop a few years ago, and many ideas. Almost all of them would benefit more than just those few proprietary map systems. There are even simple primitive ideas like adding the concept of pan and zoom to the platform, maybe in CSS. Surely we can do better than where things are right now, but who is going to invest to get it there?