Hacker News Reader

Comment by rlv-dan | Hacker News Reader

A big problem is that the web standard is too complicated. Browsers are pretty much a virtual os that can do almost everything. For anyone to make a browser from scratch today is not feasible. We will move towards a monoculture, because there is no alternative unless we make a drastic shift in rethinking how the web should work.

View on HN

Yeah. In the early days of internet, web won, because of its relative expressiveness and lack of -rigidity- structure, compared to gopher for example. Now we see that the same unbounded expressiveness led to a monolithic system within a system, that is only feasibly maintainable by large organizations. One program should do one thing, (and do it well): that's what leads to healthy systems.

For me, it's not the bloated browser centric computing experience that's the biggest problem. In my eyes, it's eclipsed by the armies of web developers whose only professional purpose is to work with this technological mishap. If our only document delivery system, i.e. web, wasn't such a mess, the same could be accomplished by a fraction of the current workforce.

But the web has also allowed for things which are way more than document delivery; imageboards, for instance, weren't around before the web, and the fact of carrying images made them unsuitable for previous technology.

Yeah, it became hard to define. That's pretty cool for users in the short run, but pretty bad in the long run, from a systems point of view. In other words, web surely had its moments, but unfortunately the biggest one is this peak-web thing we're talking about.

This. The pain of developing apps for the web could be tolerated as long as there was a vision for browsers becoming a kind of universal runtime. But that has failed miserably.

Think about what Chrome's dominance means: Chrome can push users and developers around as it wants. There is no long-term viability, and anything is at stake. Chrome already has QUIC support, will soon gain Grid Layout Level 2 support, etc. At the end of this decade, we'll be left with a generation of developers lost in the cloud.

For too long, the situation on the web has been that the inmates run the asylum, including so-called standard bodies financially dependent on Google. As they say, the road to a dystopian future is littered with good intentions.

This!

I'd be very happy with a fixed, lower tier of web standard that is easy to implement against. It should contain enough standards to build attractive, solid (eg) online banking interfaces, blogs and news sites, commenting etc. In other words, a minimum for media consumption, rather than apps.

Building a browser that supported only AMP would be one way to do this and have a significant number of pages you could handle.

(Disclosure: I work for Google on AMP ads. I also know how much people here dislike AMP, so this is not entirely serious.)

Not actually a terrible idea if it can be done without tracking. Thank you for disclosing.

I'd quite like an RSS+reader view browser. At a guess, RSS+reader view+AMP could cover pretty substantial amounts of existing content.

It could be done entirely without external requests: the browser would be an amp interpreter, no need for the amp js as all.

The hardest part is that amp probably isn't restrictive enough: you can still use a very large subset of HTML and CSS.

Very interesting thought. Probably enough HTML for tracking though right?

> Probably enough HTML for tracking though right?

I'm not sure what you mean?

That is true it is a problem. Also the problem is use of web pages for stuff where it isn't a good idea. And also, any browser seems pretty terrible to me, I have found (I wrote some ideas about how to make a better one, but also there is the use for other kind of programs, including telnet, plain text email (without HTML), IRC, NNTP, gopher, etc). Even HTTP is misused.

Maybe Chrome can compete with itself by swapping out the rendering and Javascript vms.

> Maybe Chrome can compete with itself by swapping out the rendering and Javascript vms.

Not that I have any insider information but from the complaints I've read about how the committees that handle promotion requests at Google, I'd be surprised if there aren't Google engineers already trying their hand at this.

What I'd like to know is how much does Chrome cost Google? I've read that Mozilla spends just under half a billion dollars a year so I'd expect Google likely spends at least a billion a year on Google Chrome (and associated tasks). Not sure how we'd come up with a number though because I wouldn't know how to put a price tag on things like webp, webm, depot tools, or the Google infrastructure that Google Chrome team gets to use (presumably for free).

The Mozilla Foundation is doing a whole lot of stuff besides developing Firefox, whereas not all development for Firefox is paid for by Mozilla.

I remember estimates (so, take them with a laaaaarge pinch of salt) on ads promoting Chrome, which were expected to be ~4x the total budget of Mozilla.

I would be surprised if (without counting ads), the Chrome paid devs were not at least 3x the number of Firefox paid devs.

> without counting ads

Oh, good point. I didn't think of that. I think the fact that Google promotes Chrome on Google Search and so on is worth more than the ads I see all over the web as well.

Reply on news.ycombinator.com