5 ms·

Sorry but this vision is totally unrealistic. You're right there's a chicken/egg problem: no-one's going to use that fantasy browser, because it can't browse t

by roca 7y ago

Sorry but this vision is totally unrealistic. You're right there's a chicken/egg problem: no-one's going to use that fantasy browser, because it can't browse the Web. We know this because faster experimental browsers exist, and essentially no-one uses them. Even browsers with tens or hundreds of millions of users struggle to get Web developer attention, even when they're almost completely compatible with the dominant browser and don't require fundamentally rearchitecting, and arguably crippling, one's Web site. Also, from a technical point of view, ripping out JS and DOM APIs does not magically mean you get a "wicked-fast, private, secure browser". It would be a lot more secure, but building a fast HTML/CSS rendering engine that still has to handle window resizing, zooming, incremental loading, editing and other interaction is still incredibly hard, in fact not hugely easier than supporting JS. (Source: I'm a former Mozilla distinguished engineer who worked on this stuff.) By far the easiest path to such a browser would be to take an existing engine and rip out stuff, and I don't think you're going to end up with something much faster. Disabling JS in a regular browser is going to give you most of the wins.

anderspitman 7y ago
Everything is unrealistic until it becomes real. I naturally tend towards cynicism sometimes, but I try to practice dreaming a little. I find I'm happier when I try to imagine the world as it could be and what I can do to push it that direction. True, general web layout is a hard problem. But there are a lot of knobs to turn. What if we removed CSS layout altogether, and only allowed linear pages, a la reader mode? That's a big tradeoff, but maybe it would work. If that's too much, maybe only allowing flexbox would be enough. As for editing, at that point you're making a web app.
- roca 7y ago
  > As for editing, at that point you're making a web app. Are you getting rid of Web forms now? Also, remember that at the very beginning Berners-Lee's WorldWideWeb browser supported HTML editing, so it's definitely part of some vision of the "good old days". Yes, if you cut down or eliminate CSS, or just redesign it, you could build a faster renderer. But these questions point to one of major defects of the "cut-down Web" vision: lots of people have a limited set of Web features that they like, but they tend not to agree on what that set is. You like flexbox. Some people really like grid. Some people really need ruby, or floats, or columns, or editing, or vertical breaking (pagination). Pretty soon you've got something nearly as complicated as real CSS.
  - anderspitman 7y ago
    I think getting rid of editability would be great. The whole point is a minimal browser. I don't really want to cut down the web. I want to start from scratch but in a manner compatible with our current JavaScript VMs.
  - jancsika 7y ago
    > But these questions point to one of major defects of the "cut-down Web" vision: lots of people have a limited set of Web features that they like, but they tend not to agree on what that set is. Democracy to the rescue: 1. Start a FLOSS mailing list thread to get as much input as possible on the best starting point for this reduced set of functionality. Gopher, resurrecting Xanadu, revisiting Hypercard, and so on. Let the conversation flow. 2. Invite the people who chose Firefox Reader View to private gitlab instance and get to work.
  - imtringued 7y ago
    I agree but a lot of the pre-flexbox stuff in CSS is just plain crap. However, I don't think removing obsolete CSS constructs will make anything faster or better because you're not forced to use them.
- roca 7y ago
  If you want to push the world in your direction, join a browser company. You may or may not be able to push them in your direction, but you'll at least understand the problem space a lot better to see where it's worth pushing.
  - anderspitman 7y ago
    You're talking about this from a value-proposition point of view. Whether something is worth doing doesn't depend only on the final value. It also depends on the amount of effort to do the thing. If the amount of effort is tiny, then it's probably best to do the thing and see what happens.

by roca 7y ago

anderspitman 7y ago

Everything is unrealistic until it becomes real. I naturally tend towards cynicism sometimes, but I try to practice dreaming a little. I find I'm happier when I try to imagine the world as it could be and what I can do to push it that direction. True, general web layout is a hard problem. But there are a lot of knobs to turn. What if we removed CSS layout altogether, and only allowed linear pages, a la reader mode? That's a big tradeoff, but maybe it would work. If that's too much, maybe only allowing flexbox would be enough. As for editing, at that point you're making a web app.

roca 7y ago
> As for editing, at that point you're making a web app. Are you getting rid of Web forms now? Also, remember that at the very beginning Berners-Lee's WorldWideWeb browser supported HTML editing, so it's definitely part of some vision of the "good old days". Yes, if you cut down or eliminate CSS, or just redesign it, you could build a faster renderer. But these questions point to one of major defects of the "cut-down Web" vision: lots of people have a limited set of Web features that they like, but they tend not to agree on what that set is. You like flexbox. Some people really like grid. Some people really need ruby, or floats, or columns, or editing, or vertical breaking (pagination). Pretty soon you've got something nearly as complicated as real CSS.
- anderspitman 7y ago
  I think getting rid of editability would be great. The whole point is a minimal browser. I don't really want to cut down the web. I want to start from scratch but in a manner compatible with our current JavaScript VMs.
- jancsika 7y ago
  > But these questions point to one of major defects of the "cut-down Web" vision: lots of people have a limited set of Web features that they like, but they tend not to agree on what that set is. Democracy to the rescue: 1. Start a FLOSS mailing list thread to get as much input as possible on the best starting point for this reduced set of functionality. Gopher, resurrecting Xanadu, revisiting Hypercard, and so on. Let the conversation flow. 2. Invite the people who chose Firefox Reader View to private gitlab instance and get to work.
- imtringued 7y ago
  I agree but a lot of the pre-flexbox stuff in CSS is just plain crap. However, I don't think removing obsolete CSS constructs will make anything faster or better because you're not forced to use them.
roca 7y ago
If you want to push the world in your direction, join a browser company. You may or may not be able to push them in your direction, but you'll at least understand the problem space a lot better to see where it's worth pushing.
- anderspitman 7y ago
  You're talking about this from a value-proposition point of view. Whether something is worth doing doesn't depend only on the final value. It also depends on the amount of effort to do the thing. If the amount of effort is tiny, then it's probably best to do the thing and see what happens.