Main » Discussion » Web Browser Discussion » New reply
    Alert
    You are about to bump an old thread. This is usually a very bad idea. Please think about what you are about to do before you press the Post button.
    New reply
    Post help

    Presentation

    [b]…[/b] — bold type
    [i]…[/i] — italic
    [u]…[/u] — underlined
    [s]…[/s] — strikethrough
    [code]…[/code] — code block
    [spoiler]…[/spoiler] — spoiler block
    [spoiler=…]…[/spoiler]
    [source]…[/source] — colorcoded block, assuming C#
    [source=…]…[/source] — colorcoded block, specific language[which?]
    [abbr=…]…[/abbr] — abbreviation
    [color=…]…[/color] — set text color
    [jest]…[/jest] — you're kidding
    [sarcasm]…[/sarcasm] — you're not kidding

    Links

    [img]http://…[/img] — insert image
    [url]http://…[/url]
    [url=http://…]…[/url]
    >>… — link to post by ID
    [user=##] — link to user's profile by ID

    Quotations

    [quote]…[/quote] — untitled quote
    [quote=…]…[/quote] — "Posted by …"
    [quote="…" id="…"]…[/quote] — ""Post by …" with link by post ID

    Embeds

    [youtube]…[/youtube] — video ID only please
    Thread review
    Screwtape
    Posted by sureanem
    Both work by the same mechanism: change the way you write software so that you're forced to hand over the control to Microsoft/Google in exchange for "performance improvements"/"new features"

    They're similar, but from opposite ends.

    EEE is all carrot: "change the way you write software, in exchange for these benefits". EEE never (openly) threatens anyone; "extinguish" is supposed to come as a surprise.

    This is (almost) all stick: "change the way you write software, or we'll break it". The fact that Google is talking about benefits as well is just to sweeten the deal a bit, but the fact that Google has already openly talked about a timeline for removing Manifest V2 support proves this is not EEE.

    If one reads their annual report they will find hardly anything about browser development, but a lot about their pet projects and political views.

    You mean the annual report of the Mozilla Foundation, the non-profit organisation that doesn't actually do any browser development?

    What about the odd empty YouTube div tags to cripple Edge?

    It's a long way off from CORPORATE ESPIONAGE.

    Theoretically, GCC needs no comprehension of the standard library at all, but in practice it very definitely wants to know what strlen is and how it differs from any other size_t(char*) function.

    Pretty sure the C standard is very carefully designed to make this possible, it's not just something the GCC team invented on their own and decided to try out.

    They could even pull a Judo move and do special optimizations for some stuff that's otherwise extremely slow, then bait web developers into using them.

    Lots of CSS attributes are exactly that. For example, CSS has "rotate in 3D" attributes you can apply to a particular subtree. It's defined in such a way that browsers can just render that subtree to a texture and draw it to the screen with OpenGL, rather than implementing perspective transformations in pure software.

    I'm sure someone who's better at thinking could come up with some even better idea, but it's abundantly clear they're not using the resources at their disposal effectively.

    That's the trouble with trying to keep the moral high ground.

    Writing them by hand would indeed be a lot of work, but surely you could at least do something like PGO to get a rendering engine specially optimized for nytimes.com whenever you load nytimes.com?

    For any optimisation, you have to ask: what are the tradeoffs? On one hand we obviously have performance, but the other hand is more obscure. Complexity is a cost (both for browser vendors, and for web developers trying to understand why their site behaves the way it does), inflexibility is a cost (deploying a change is now much more complex than reloading the page), fragility is a cost (do we fail to use our optimised version if NYT adds or removes a redundant semicolon?), and I'm sure there are many others.

    Is it really going full Palpatine to ignore what some obscure group nobody cares about has recommended? Joe Q. Public won't care and it won't make CNN, that's for sure.

    How many news sites have already reported on Google's Manifest V3 changes, just based on the possibility that Google is abusing their power? I bet they'd all publish another round of articles if somebody could *prove* Google was up to no good. And Google obviously cares about their public perception, or they would have just straight-up made the change without bothering to give anyone advance notice.
    CaptainJistuce
    Posted by tomman


    Also, I was of the belief that the modern solution for "custom UI for 500 of the most popular websites" was "install our cellphone app!!!!" instead of browser hijacks...
    I think it is called "instant apps". The website can serve up the files for a pocket PC application, and the device can download the application to RAM and run it there. Leave the site and the app dies.
    Nicholas Steel
    Posted by Nicholas Steel
    Posted by CaptainJistuce
    MS gave up on making their own rendering engine because Google was actively sabotaging them by embedding anti-Edge code in their websites. It sucks, but when you only supply the browser, the company that supplies the browser and content gets to dictate everything.


    The sabotaging is still happening, Google is delivering versions of Youtube websites that rely on depreciated beta functionality while either... not doing that if you access the website with Chrome or Chrome supports the defunct functionality when it really shouldn't be.


    See: https://www.theverge.com/2018/7/25/17611444/how-to-speed-up-youtube-microsoft-edge-safari-firefox
    tomman The piracy problem is solved for me: I simply don't consume your content, legit or pirate.

    I'm out of the system for all that matters.

    Hollywood can go choke on a flaming dick of fire, for all that matters. But then, that's just me, an extremist asshole that tries its hardest to get isolated from the world at large.

    Also, I was of the belief that the modern solution for "custom UI for 500 of the most popular websites" was "install our cellphone app!!!!" instead of browser hijacks...
    ‮strfry("emanresu")
    Posted by Screwtape
    If EEE is "stealing something from you", this is "making you return something you borrowed". Just because they both happen slowly doesn't make them the same thing.

    I'm sorry, I don't follow. They both take power from the add-on developers and give it back to you, the people! move it to the browser development team. Both work by the same mechanism: change the way you write software so that you're forced to hand over the control to Microsoft/Google in exchange for "performance improvements"/"new features"

    It seems very optimistic to say "final stages". "Next stage", certainly.

    Well, I reckon their end game is to solve the piracy problem once and for all, and with it the problem of hate speech, just like they tried to do in the early 2000's. And stage one of this end game is to start quelling ad blockers. But stages plural, definitely.

    Delivering ads via DRM sounds weird. Does that mean I can hide ads by using a DRM-free display transport like DVI instead of HDMI, just like that breaks Blu-Rays and Netflix?

    Yeah, but it'll take the page with it. You might think that it sounds unrealistic, but you'd have said that about YouTube livestreams requiring DRM 5-10 years ago too. The first step will probably be newspapers' paywalls getting DRM-enhanced.

    I think Google is pretty confident that they've got the browser market on lockdown and they can now afford to spend effort on what *they* want to do, instead of what their *users* want them to do. You know, like Microsoft did after IE6 was released.

    Well, they can't mess it up too badly. They can on mobile. They don't have complete dominance, since something grossly repulsive (e.g. worse than DRM) would get Mozilla to object.
    And while they do control Mozilla, pushing through anything too objectionable would probably get it to self-destruct, after which it becomes apparent that they have a monopoly, causing them to receive scrutiny from the regulator. Mozilla plays a very important role for them, and is the only check on their behavior there is.

    Well, *obviously* Google has something to do with Mozilla's decline

    I'm talking about Mozilla's internal conflicts. There's not that much shady with Chrome taking the #1 spot, in comparison.

    The thing that conspiracy-minded observers never seem to understand is that things fall apart. It is the usual and standard behaviour of the universe for things to not go the way we want them to; making some positive thing happen requires substantial effort and organisation. On the other hand, making a bad thing happen requires literally zero effort - let your attention slip for a moment, get distracted, or even just fumble, and your world can change irreparably.

    Let's be real here, Mozilla's fall from grace is hardly a case of entropy taking over. They fired their competent people and then put what appears to be loonies in charge, that sounds more like sabotage than incompetence. If one reads their annual report they will find hardly anything about browser development, but a lot about their pet projects and political views.

    I can't find the thread now, but I recall reading a Firefox dev (or possibly an ex-Firefox dev) complaining that even though individual Googlers value cross-browser compatibility and interoperability and all that, Google-owned websites keep working really well in Chrome and wind up sluggish or glitchy in other browsers. It's not because anyone's trying to sabotage other browsers, it's just because if a Googler has to choose between cross-browser bugfixing and something they actually get paid to do, they're going to do the paid work every time.

    What about the odd empty YouTube div tags to cripple Edge?

    youtube-dl gets very regular updates to keep it working, and it only has to handle a tiny fraction of the YouTube API. I can't easily find any stats, but anecdotally it's often a few days between youtube-dl breaking and a new release that gets it working again. If YouTube was entirely unavailable in a particular browser for a few days, that would be a Big Deal.

    Well, okay, fair enough. They could opt for some kind of compromise, like using youtube-dl if it works and otherwise YouTube proper.

    The whole point of a web-browser is that you implement HTML, JS and CSS once, and then you can handle all the websites. As much as people complain about HTML, JS and CSS being over-complicated and inscrutable, it's still vastly simpler to implement a web-browser than it would be to implement a custom native UI from scratch for every existing website. The idea that browsers should implement a custom UI for the 500 most popular websites without their knowledge or consent, and transparently keep those custom UIs up-to-date as the websites change their backend is... not so much a cat-and-mouse game as it is Sisyphean.

    I can't imagine websites explicitly wishing to block Decentraleyes without provocation, but I can't imagine websites explicitly wanting to support it, either. The minute some website gets a support call because Decentraleyes has loaded the wrong script, or the wrong version of a script, or even just interfered with the reliability of their analytics, I guarantee they will figure out a way to fix their problem, and it won't involve reporting an issue to Decentraleyes and waiting for them to release an updated version.

    Couldn't you say this about compilers too? Theoretically, GCC needs no comprehension of the standard library at all, but in practice it very definitely wants to know what strlen is and how it differs from any other size_t(char*) function. And it seems like even if you could never make a perfect compiler, you could still make one which used some tricks to get ahead some of the time. Against an actively adversarial website it may not be possible, but I can't see the others caring about your optimizations. They could even pull a Judo move and do special optimizations for some stuff that's otherwise extremely slow, then bait web developers into using them. Considering web developers do not appear to care much about performance as long as it runs fine on their test rigs, it would be a winning short-term strategy. In the long term, of course it doesn't matter. Mozilla's only ace up their sleeve is that they don't have to suffer under antitrust legislation, but they have been unwilling to leverage this and now it's too late.

    But as a counterfactual example of what they could have done: In the next update, Firefox ships with a built-in opt-out ad blocker. They whitelist "acceptable ads" like ABP, which doesn't include Google but does include Yandex. Collect $X million where X > XGoogle from Yandex, do pass go.

    For reference, Google makes something like $200 billion a year from ads and gives Mozilla about $200m a year. Firefox has a 5% market share. 0.200/(200*0.05) = 0.02. As long as their cut from this Yandex deal would exceed 2%, they profit. As a bonus, Yandex has a vested interest in the failure of the Chrome platform (e.g. willing to subsidize Mozilla and perhaps even turn Firefox into a Chrome), the support of a major national government (like Google), and are not really subject to antitrust/privacy law or the GDPR.

    I'm sure someone who's better at thinking could come up with some even better idea, but it's abundantly clear they're not using the resources at their disposal effectively.

    Writing them by hand would indeed be a lot of work, but surely you could at least do something like PGO to get a rendering engine specially optimized for nytimes.com whenever you load nytimes.com? And Decentraleyes' script repository should of course be updated via the browser's built-in surreptitious auto-update mechanism without user consent or knowledge.

    For extra breakage of non-Firefox browsers competitive edge, it could just go ahead and load jQuery et al into the global namespace each time you open a new page, or at least lazy-load it. Then some pages would start to rely on it, then Google would have to implement it, and Mozilla will have landed a small blow and gained an equally small share of users. Rinse and repeat, in theory.

    The point is that Google is trying to balance their operational goals with preserving user good-will. If independent measurements show that Google's claims are substantially correct (adblockers do slow browsers down, and can be slimmed down without losing their effectiveness) then everybody wins - Google gets to do their thing, users get faster adblocking. If independent measurements do *not* agree with Google's claims, or even contradict them (adblockers aren't very slow, and they really need very large lists), then Google will need to sacrifice their operational goals even further (good for users) or go full Palpatine (probably a very poor move, at least today, so still good for users).

    Is it really going full Palpatine to ignore what some obscure group nobody cares about has recommended? Joe Q. Public won't care and it won't make CNN, that's for sure.

    Or they could just say that they've read the complaints and will review their suggestions, and then stall for a few months and go ahead with it anyway.

    I just don't see the point.
    Screwtape
    Posted by sureanem
    Posted by Screwtape
    EEE is a notable strategy for *acquiring* the power to extinguish something you don't like; you have to do it slowly in stages so that people don't notice how much power you've acquired until it's too late.

    Isn't that what they're doing, though? You introduce this, and some obscure add-on maintainers nobody cares about whine for a bit and ABP goes along with it, and then when the noise's died down you turn the screws another quarter turn, and so on and so forth?

    If EEE is "stealing something from you", this is "making you return something you borrowed". Just because they both happen slowly doesn't make them the same thing.

    To me, it just looks like they're moving into the final stages of "leverage Chrome monopoly".

    It seems very optimistic to say "final stages". "Next stage", certainly.

    I can guess two possible reasons: one is that they judge that they've outmaneuvered Mozilla and will soon have complete control, the other one is that they judge that the desktop will soon fade into irrelevance and it's important to kill ad blocking to set the stage for ads delivered via DRM.

    Delivering ads via DRM sounds weird. Does that mean I can hide ads by using a DRM-free display transport like DVI instead of HDMI, just like that breaks Blu-Rays and Netflix?

    I think Google is pretty confident that they've got the browser market on lockdown and they can now afford to spend effort on what *they* want to do, instead of what their *users* want them to do. You know, like Microsoft did after IE6 was released.

    The real conspiracy-minded observers might say the timing is "interesting" considering they've just managed to kill off the last non-Firefox competitor. They might also speculate that Google had something to do with Mozilla's decline. Corporate espionage just isn't that expensive, you know.

    Well, *obviously* Google has something to do with Mozilla's decline - at least for a while, Google would show "why not try Chrome" ads for every non-Chrome browser that visited google.com; Chrome got bundled with a bunch of third-party tools like Adobe Reader, etc. For all its technical skill, Mozilla has never had a very large advertising budget, so when the largest advertising company on the planet picked an advertising fight, it's not surprising they came out on top.

    The thing that conspiracy-minded observers never seem to understand is that things fall apart. It is the usual and standard behaviour of the universe for things to not go the way we want them to; making some positive thing happen requires substantial effort and organisation. On the other hand, making a bad thing happen requires literally zero effort - let your attention slip for a moment, get distracted, or even just fumble, and your world can change irreparably.

    I can't find the thread now, but I recall reading a Firefox dev (or possibly an ex-Firefox dev) complaining that even though individual Googlers value cross-browser compatibility and interoperability and all that, Google-owned websites keep working really well in Chrome and wind up sluggish or glitchy in other browsers. It's not because anyone's trying to sabotage other browsers, it's just because if a Googler has to choose between cross-browser bugfixing and something they actually get paid to do, they're going to do the paid work every time.

    Youtube-dl isn't blocked, and is most likely significantly harder to block. They could just implement that in the browser with fast updates.

    youtube-dl gets very regular updates to keep it working, and it only has to handle a tiny fraction of the YouTube API. I can't easily find any stats, but anecdotally it's often a few days between youtube-dl breaking and a new release that gets it working again. If YouTube was entirely unavailable in a particular browser for a few days, that would be a Big Deal.

    The whole point of a web-browser is that you implement HTML, JS and CSS once, and then you can handle all the websites. As much as people complain about HTML, JS and CSS being over-complicated and inscrutable, it's still vastly simpler to implement a web-browser than it would be to implement a custom native UI from scratch for every existing website. The idea that browsers should implement a custom UI for the 500 most popular websites without their knowledge or consent, and transparently keep those custom UIs up-to-date as the websites change their backend is... not so much a cat-and-mouse game as it is Sisyphean.

    Also, most websites would probably not mind if you still let them update their layouts. I can't imagine many websites explicitly wishing to block Decentraleyes, for instance.

    I can't imagine websites explicitly wishing to block Decentraleyes without provocation, but I can't imagine websites explicitly wanting to support it, either. The minute some website gets a support call because Decentraleyes has loaded the wrong script, or the wrong version of a script, or even just interfered with the reliability of their analytics, I guarantee they will figure out a way to fix their problem, and it won't involve reporting an issue to Decentraleyes and waiting for them to release an updated version.

    The next step would be for some independent group to try and measure things...

    I don't see what problem it'd solve. Google now knows exactly how much to cripple ad blockers rate to be a net benefit in terms of battery?

    The point is that Google is trying to balance their operational goals with preserving user good-will. If independent measurements show that Google's claims are substantially correct (adblockers do slow browsers down, and can be slimmed down without losing their effectiveness) then everybody wins - Google gets to do their thing, users get faster adblocking. If independent measurements do *not* agree with Google's claims, or even contradict them (adblockers aren't very slow, and they really need very large lists), then Google will need to sacrifice their operational goals even further (good for users) or go full Palpatine (probably a very poor move, at least today, so still good for users).
    ‮strfry("emanresu")
    Posted by Screwtape
    EEE is a notable strategy for *acquiring* the power to extinguish something you don't like; you have to do it slowly in stages so that people don't notice how much power you've acquired until it's too late.

    Isn't that what they're doing, though? You introduce this, and some obscure add-on maintainers nobody cares about whine for a bit and ABP goes along with it, and then when the noise's died down you turn the screws another quarter turn, and so on and so forth?

    Google has always had the power to extinguish adblockers, this is just standard PR where they're trying to find a compromise between what they want and what their customers will accept.

    No, I don't think that sounds reasonable. On mobile, Google has complete vertical integration. On desktop, Google competes with Firefox. To me, it just looks like they're moving into the final stages of "leverage Chrome monopoly". I can guess two possible reasons: one is that they judge that they've outmaneuvered Mozilla and will soon have complete control, the other one is that they judge that the desktop will soon fade into irrelevance and it's important to kill ad blocking to set the stage for ads delivered via DRM.

    To outright kill adblockers might be illegal (cf. Microsoft antitrust case) and at any rate garner immense backlash, but to cripple them for "performance reasons" (again, cf. Microsoft antitrust case) is not. Keep in mind that there's otherwise little to no reason for them to do this, since Chrome beats Firefox in benchmarks and those are done without ad blockers enabled anyway.

    The real conspiracy-minded observers might say the timing is "interesting" considering they've just managed to kill off the last non-Firefox competitor. They might also speculate that Google had something to do with Mozilla's decline. Corporate espionage just isn't that expensive, you know. And Mozilla's decline does look awfully similar not to say identical to what I'd expect Google to engineer if they indeed did have engineered it, with Firefox looking an awful lot like Google Reader ca. 2010.

    Most of the other products Google bought up died, so why shouldn't Firefox?

    I think you missed the bit where websites want to control how they're presented to the user. The only reason Decentraleyes works is because such a small fraction of users use it that it's not worth the effort it would take to undermine. If a major browser started matching "https://youtube.com/watch?v=" and replacing it with some other content, I guarantee YouTube videos would start being served from "youtube.io" or "media.google.com" or something within hours.

    That's a cat-and-mouse game. Youtube-dl isn't blocked, and is most likely significantly harder to block. They could just implement that in the browser with fast updates. They could also recoup their performance loss by blocking all ads by default, and announce their displeasure with YouTube's idiosyncrasies through informal channels.

    Also, most websites would probably not mind if you still let them update their layouts. I can't imagine many websites explicitly wishing to block Decentraleyes, for instance.

    Android already has a system where apps can register themselves as handling particular URLs, so when you browse to youtube.com on your mobile browser, you get a prompt to open the page in the YouTube app.

    Fair enough. You still need the app though, and often it's faster to load the website.

    The point is not that a particular adblocker is inefficient, the point (from Google's point of view) is that if/when an adblocker does such a thing, Google gets the blame, and that's unfair.

    When has Google ever gotten the blame for this though? Very few people discuss switching browsers (they all use Chrome), and the few who do would consult benchmarks and not real life performance. Unless there is a significant difference, nobody would bother. And even then, ad blockers are likely to be as slow in Firefox as in Chrome.

    Google have presented a list of reasons why their proposal benefits end-users. Anyone with two brain-cells to rub together can come up with other, less-altruistic motivations, but that doesn't make those original reasons wrong or misleading, they're just not the whole picture.

    Well, their reasons don't sound very plausible. And if they're not the real motivations, then certainly they're wrong or misleading.

    The next step would be for some independent group to try and measure things like "how much time and battery does an adblocker take" versus "how much time and battery does advertising take", and maybe experiment with reducing the complexity of adblocking lists to see how much affect that has on adblocking costs, and how far you can trim down the list while still being a net benefit over not using an adblocker. You know, actual *information*. Lining up to yell at Google for being oppressive or adblockers for being wasteful might be cathartic, but it doesn't actually improve the situation.

    I don't see what problem it'd solve. Google now knows exactly how much to cripple ad blockers rate to be a net benefit in terms of battery?

    I don't care a particular lot about battery, nor about the 0.1s or whatever longer page load times, however seeing the disgusting and repulsive ads does certainly bother me. This goes for the supposedly non-intrusive and SFW text ads too. It's like the physically repulsive beggars who couldn't even be bothered to learn the language properly hassling you for money. Just piss off! It's not a matter of hygiene or tax laws or whatever, that's just the government's way of politely saying they don't like you.

    (the crucial difference being that ads are ostensibly needed for websites to survive)
    Screwtape
    Posted by sureanem
    Chrome is embracing ad blocking and "extending" the platform, from the ABP filter lists to a crippled implementation. It sounds a lot like EEE to me.

    EEE is a notable strategy for *acquiring* the power to extinguish something you don't like; you have to do it slowly in stages so that people don't notice how much power you've acquired until it's too late.

    Google has always had the power to extinguish adblockers, this is just standard PR where they're trying to find a compromise between what they want and what their customers will accept.

    I'm suggesting that they should have done something like Decentraleyes but for the whole domain. As in, a request for youtube.com/watch?v=asdasd should be treated separately to asdasd.com/watch?v=asdasd, and instead get served a static implementation of YouTube, just like about:config or similar does. And this should be done for, say, the top 500 websites.

    I think you missed the bit where websites want to control how they're presented to the user. The only reason Decentraleyes works is because such a small fraction of users use it that it's not worth the effort it would take to undermine. If a major browser started matching "https://youtube.com/watch?v=" and replacing it with some other content, I guarantee YouTube videos would start being served from "youtube.io" or "media.google.com" or something within hours.

    Obviously, Mozilla would never have done this, Microsoft apparently didn't have the guts, and Google won't bother now that the competition is out of the way. So it won't ever happen. But it should have. It's the next logical step in the development of web browsers.

    Android already has a system where apps can register themselves as handling particular URLs, so when you browse to youtube.com on your mobile browser, you get a prompt to open the page in the YouTube app.

    Surely, they can't be using a linear scan?

    I don't actually know whether any particular adblocker uses a linear scan, but it would be a *lot* simpler than a prefix tree or dynamically building a NFA, and "do the simplest thing that can possibly work" is a common thing that software engineers tell each other.

    The point is not that a particular adblocker is inefficient, the point (from Google's point of view) is that if/when an adblocker does such a thing, Google gets the blame, and that's unfair.

    This is Google, a company which makes money off of advertising and has a nigh-monopoly on browsers. Do you really think they're doing this to save ad blocking?

    Google have presented a list of reasons why their proposal benefits end-users. Anyone with two brain-cells to rub together can come up with other, less-altruistic motivations, but that doesn't make those original reasons wrong or misleading, they're just not the whole picture.

    The next step would be for some independent group to try and measure things like "how much time and battery does an adblocker take" versus "how much time and battery does advertising take", and maybe experiment with reducing the complexity of adblocking lists to see how much affect that has on adblocking costs, and how far you can trim down the list while still being a net benefit over not using an adblocker. You know, actual *information*. Lining up to yell at Google for being oppressive or adblockers for being wasteful might be cathartic, but it doesn't actually improve the situation.
    Nicholas Steel
    Posted by sureanem
    Also, if the 30k URL hard cap becomes put into practice, websites could just use a large enough variation that it wouldn't be possible to block.


    You'd just have to install thousands of adblocking add-ons each with their own 30,000 limit...
    ‮strfry("emanresu")
    Posted by Screwtape
    Except that we're talking about Chrome's own extension API. The "embrace" in "embrace, extend, extinguish" is embracing *somebody else's* platform, so you can improve it or at least destabilise it. If anything, it's the "WebExtensions" effort shared by Firefox, Edge, Vivaldi etc. that are trying to embrace and extend Chrome's extension API.

    Chrome is embracing ad blocking and "extending" the platform, from the ABP filter lists to a crippled implementation. It sounds a lot like EEE to me.

    It's one thing to have a hand-tuned implementation of memcpy for people to call. It's another thing to find chunks of code that *look* like memcpy so you can silently swap them out for your own implementation, and it's yet another thing when programs deliberately obfuscate their memcpy implementations so that you won't be able to swap them out.

    GCC does both, or at least it should. The goal is to produce fast programs/websites. Only results matter, or as the Italian philosopher and politician put it: The means will always be considered honest, and he will be praised by everybody because the vulgar are always taken by what a thing seems to be and by what comes of it; and in the world there are only the vulgar, for the few find a place there only when the many have no ground to rest on.

    I'm suggesting that they should have done something like Decentraleyes but for the whole domain. As in, a request for youtube.com/watch?v=asdasd should be treated separately to asdasd.com/watch?v=asdasd, and instead get served a static implementation of YouTube, just like about:config or similar does. And this should be done for, say, the top 500 websites.

    That would make it impossible for them to mess up their performance by making slight changes to the layout, and it would make them win benchmarks handily, since it could skip the HTTP request and large chunks of rendering. The downside is that design changes might take a bit more time to propagate, but this is hardly something users would be upset about or even notice.

    Another downside is of course that it would make browsers even more arcane pieces of software and break with lots of standards. Then again, standards are descriptive and not prescriptive, so if enough people start doing it, it will become standards-compliant behavior again.

    Obviously, Mozilla would never have done this, Microsoft apparently didn't have the guts, and Google won't bother now that the competition is out of the way. So it won't ever happen. But it should have. It's the next logical step in the development of web browsers.

    Currently, all adblockers are based on big lists of Bad URLs. Once a URL is added to the list, there's very little incentive to remove it: if a user sees an ad, they blame the adblocker, but if the user's browser is slow, they'll blame the browser (and checking a million useless URLs might still be faster than loading the ad). Manifest V3 seems specifically designed to prevent blockers from using the "big list of Bad URLs" approach, so yeah, I imagine all those extensions will need to be at least redesigned.

    Surely, they can't be using a linear scan?

    This is Google, a company which makes money off of advertising and has a nigh-monopoly on browsers. Do you really think they're doing this to save ad blocking?

    Also, if the 30k URL hard cap becomes put into practice, websites could just use a large enough variation that it wouldn't be possible to block.
    Screwtape > No, Chrome isn’t killing ad blockers -- we’re making them safer

    Porqué no los dos?

    > It sounds a lot like a classic Microsoft stategy.

    Except that we're talking about Chrome's own extension API. The "embrace" in "embrace, extend, extinguish" is embracing *somebody else's* platform, so you can improve it or at least destabilise it. If anything, it's the "WebExtensions" effort shared by Firefox, Edge, Vivaldi etc. that are trying to embrace and extend Chrome's extension API.

    > They could definitely have fixed it, just like all respectable C libraries have hand-tuned assembly implementations of memcpy et al. for any CPU architecture of note.

    It's one thing to have a hand-tuned implementation of memcpy for people to call. It's another thing to find chunks of code that *look* like memcpy so you can silently swap them out for your own implementation, and it's yet another thing when programs deliberately obfuscate their memcpy implementations so that you won't be able to swap them out.

    > Will Manifest v3 break uBlock Origin?

    Currently, all adblockers are based on big lists of Bad URLs. Once a URL is added to the list, there's very little incentive to remove it: if a user sees an ad, they blame the adblocker, but if the user's browser is slow, they'll blame the browser (and checking a million useless URLs might still be faster than loading the ad). Manifest V3 seems specifically designed to prevent blockers from using the "big list of Bad URLs" approach, so yeah, I imagine all those extensions will need to be at least redesigned.
    Nicholas Steel
    Posted by CaptainJistuce
    MS gave up on making their own rendering engine because Google was actively sabotaging them by embedding anti-Edge code in their websites. It sucks, but when you only supply the browser, the company that supplies the browser and content gets to dictate everything.


    The sabotaging is still happening, Google is delivering versions of Youtube websites that rely on depreciated beta functionality while either... not doing that if you access the website with Chrome or Chrome supports the defunct functionality when it really shouldn't be.

    There's an add-on for Firefox and Chrome to get Google to deliver their older Youtube website design and experience a significantly more responsive website: https://github.com/xxzefgh/youtube-classic-extension (I use this and a custom dark theme.)
    nyanpasu64 Will Manifest v3 break uBlock Origin? uMatrix (the ability to ban all requests from specific hosts)? The ability for extensions to block all JS?

    I'm speculating uBlock Origin (combats Instart Logic proxied ads) will be the first casualty, though it doesn't seem gorhill has mentioned it?
    ‮strfry("emanresu") They had it coming. They could definitely have fixed it, just like all respectable C libraries have hand-tuned assembly implementations of memcpy et al. for any CPU architecture of note. You already have Decentraleyes, so why not take it a step further?

    CaptainJistuce
    Posted by Nicholas Steel
    Here's a response from Google regarding Manifest V3 and Adblockers: https://security.googleblog.com/2019/06/improving-security-and-privacy-for.html


    Odd, I just get a video of some guy in a suit laughing maniacly.

    Posted by tomman
    Speaking about MS, they're still thinking about releasing their Chromium skin Edge for Linux:
    https://tech.slashdot.org/story/19/06/14/142256/microsoft-edge-might-come-to-linux

    ...why?! At this stage, I would actually want a native MSIE port (yes, that's real evil, but then so is Chrome). Trident and Presto need urgently to come back from the dead, evolve, and give some serious fight to the Webkit/Blink hegemony (since Mozilla is too busy with their art school dropout designers to actually care about Making Gecko Great Again)
    MS gave up on making their own rendering engine because Google was actively sabotaging them by embedding anti-Edge code in their websites. It sucks, but when you only supply the browser, the company that supplies the browser and content gets to dictate everything.
    ‮strfry("emanresu") It's very expensive to make a rendering engine, and with all the HTML5 stuff Trident/Presto are practically back at square one. There's not much profit to it anymore.

    Microsoft did it to get an advantage for their platform, this motive is out. Opera got money from shareware, ads, etc. Later on from propping up Google's monopoly. Firefox also gets their money from propping up Google's monopoly. That's the only revenue source nowadays, unless you want to get real creative (see: Brave, Hola)

    But to do this in the first place, you need a user base. You're not going to get any investments to develop a browser because Google later on might want to pay you for the search engine spot. Most importantly, Google will make sure to eliminate any competitors that pose a serious risk to them. If your project is indeed a commercial one, they can just kill it in one fell swoop whenever you start acting up. This is also why Mozilla are doing such stupid shit: if they didn't, Google would dry up the funding, and then their pet projects (which are what they really care about) would die.

    And when you've secured your user base, you might as well cut your expenses by switching over to Blink and externalize development costs to Google.

    It's a closed loop. Google makes browsers, pays money to license the usage of brands and create the illusion of competition, and in return they get a monopoly in the search engine field.

    Mozilla are kind of a special case, in that they for reasons of pride need to keep developing their engine (and also because it's a pet project), but a profit-driven corporation would obviously not care about this.

    So your only option is a FOSS third-party browser project to create that alternative you want and cater to the <0.1% of the population that actually cares. However, it has a few issues:

    1) you'd need several years of work until you reached any results of note
    2) nobody in their right mind who isn't a web developer would want to work on such a project for free
    2b) it is at least a commonly accepted truism that web developers only do FOSS projects to pad their resumés and are actually unhelpful and do not care about FOSS much at all, but them if anyone
    2c) web developers are not suited for this kind of low-level, difficult work

    You'd probably be making the GNU Hurd of browser engines. I think the fatal flaw is that you'd never reach feature escape velocity. A small but dedicated core of developers wouldn't be big enough to implement features faster than W3C adds them. So at best, you would be making the TempleOS of browser engines.

    At the end of the day, it's far easier to just deal with a blob of a few hundred megabytes, hope it doesn't get worse, and realize that there are far greater threats to freedom and security. Because at the end of the day, the question remains: why waste your time fixing a platform you don't actually like?
    tomman Speaking about MS, they're still thinking about releasing their Chromium skin Edge for Linux:
    https://tech.slashdot.org/story/19/06/14/142256/microsoft-edge-might-come-to-linux

    ...why?! At this stage, I would actually want a native MSIE port (yes, that's real evil, but then so is Chrome). Trident and Presto need urgently to come back from the dead, evolve, and give some serious fight to the Webkit/Blink hegemony (since Mozilla is too busy with their art school dropout designers to actually care about Making Gecko Great Again)
    ‮strfry("emanresu") As we all know, Google are deeply concerned with the privacy of their chattel users.

    It sounds a lot like a classic Microsoft stategy.

    [x] embrace
    [x] extend
    [ ] extinguish
    Nicholas Steel Here's a response from Google regarding Manifest V3 and Adblockers: https://security.googleblog.com/2019/06/improving-security-and-privacy-for.html
      Main » Discussion » Web Browser Discussion » New reply
      Yes, it's an ad.