As part of a larger story about Apple’s plans to allow third-party app stores on the iPhone and iPad in EU countries, Bloomberg‘s Mark Gurman claimed that Apple is also considering removing its requirement for iPhone and iPad web browsers to use WebKit, the open source browser engine that powers Safari.
Well, well, well. The EU might actually force Apple to turn iOS into a real operating system.
This is all good news, but I do wonder if we’re already in “too little too late” territory. Timing is everything. It’s significantly harder to re-balance now that the market is already encumbered by duopolies, Even assuming regulators do manage to curtail anticompetitive abuses going forward, the market share for small competitors will still continue to reflect years of an imbalanced playing field. The new/small competitors who lost the opportunity to grow during the nascent years can no longer grow with the market, instead now they must chip away at the market share of larger competitors.
Stamping out anti-competitive practices is hugely important to competitive markets, but I still think apple is going to have an advantage for years to come simply due to momentum.
It would be nicer if they also included a provision for Apple to open up obsolete hardware to alt OSes, but I’ll take alternate browsers.
Flatland_Spider,
Of course. And I’d want that provision to cover android devices too. Proprietary mobile hardware has been bad for alt-OS in general.
Alfman,
And gaming consoles, too. Especially the last two generations which could be considered good entry level desktop machines. (Xbox One X, and PS4 Pro). This could even include obsolete hardware like Steam Link, or OnLive.
For Android there are “external hackers”, but having a “device unlock after end of security updates” policy would be great 🙂
Great news for Google. Can’t wait for our new Chrome monopoly.
No but, um, you see, a Chromium-centric rendering engine monoculture across all platforms is, um, totally not like the bad old days of IE6, because, um… choice! And, um, freedom!!! Apple bad!!! They don’t make REAL computers for VERY SMART people like meeeee!!!
Finally, the one thing I always wanted from Apple: the freedom and choice to wreck my iPhone’s battery life and performance with Chromium-based apps. Thanks, EU!
that guy,
Sarcasm aside, I think it would be very interesting to run all the browsers side by side and measure their relative performances on identical hardware!
Do you have more information about chrome battery life? I searched and found this…
https://github.com/MicrosoftEdge/MSEdgeExplainers/blob/main/MediaCacheReduction/explainer.md
Chrome’s caching mechanisms are/were continually writing data to disk and keeping it from sleeping. This is especially troubling for streaming media use cases because watching an hour of video translates into continuous disk writing the whole time. This would not only contribute to shortened SSD lifespan, but it also kills battery life. It seems like a plausible explanation to me. MS addressed this in their MS edge fork a few years ago. Google may have fixed it too.
If you are referring to some other problem could you give a link to it specifically? I’m just curious.
Where have you been this past decade? Chrome has historically been a notorious RAM hog, to the point that even non-techies are aware and circulate memes about it. Electron apps (using Chromium) are likewise despised for being needlessly bloated, resource-intensive, and poorly integrated with host OS services and features. It’s like a modern-day Adobe AIR or Microsoft Silverlight, in that it’s a quick way for developers to crank out a cross-platform app with minimum effort and maximum suck.
On a positive note, it seems Google finally got Chromium to suck considerably less earlier this year, so there’s that. Not gonna pat them on the back for something that should’ve been fixed during the Obama administration.
that guy,
Memes are one thing, but I’m not interested in those at all. It’s too subjective and every browser is surrounded by bias. Sometimes the echo chamber contradicts real-world data, for instance…
https://www.tomsguide.com/news/chrome-firefox-edge-ram-comparison
https://www.businessinsider.com/google-chrome-vs-firefox-performance-memory-2018-7
These all suggest firefox actually uses more memory than chrome doing the same task.
Without data, this is more of a hypothesis that still needs to be tested and compared directly. And to be clear, I’m not trying to convince you that one browser is better than another, it’s your prerogative to choose for yourself, I was just curious about the underlying claim. Regardless I think users deserve the freedom to switch instead of being forced to always use one.
This all depend on what you consider “normal” behavior.. I am still not sure having 100 heavy sites open and loaded in memory at the same time is normal.
Carewolf
I’m not sure where you got that from? Regardless I tested this just now on my linux computer evaluating the memory.max_usage_in_bytes of a cgroup.
I opened both firefox and chromium without plugins, went to youtube.com in a single tab and played a single video (also the preroll ad).
I accept this is anecdotal, and ideally I’d run more trials on different systems with different sites, etc, but for what it’s worth…
firefox = 888MB
chromium = 809MB
Firefox used around 10% more than chromium, which is similar to the results I linked to earlier. So I don’t know where this idea that chrome based browsers specifically are memory hogs comes from? Do you? Maybe it could be because most people are using chrome and they’re noticing (correctly) that chrome is using a lot more memory than it used to even though this trend is not specific to chrome.
There may be several reasons behind the high memory usage, including optimizing for performance, handling significantly larger resources, bloated javascript libraries, etc. We’d need to dig deeper to be sure, but back to my point: we really should be looking at evidence instead of memes to make conclusions.
The Electron app thing is true.
I tried using an 8GB RAM machine for work, and it was fine until I opened all the electron apps I needed for work. Doing nothing, it would start swapping.
Anecdotal evidence, but it holds true across various machines.
To be fair, it’s probably badly written JS triggering tendencies of Chromium.
Flatland_Spider,
Ah that’s your problem right there. 32GB is the new baseline, haha 🙂
That’s kind of what I’m asking: is chromium actually the root cause, or is it more the victim of bloated software written for it? Obviously we need to look at specific case study to get a better idea, but I find it similar to the accusations made against flash. There were tons of poorly written flash applications (and I despised them as much as anyone), but as flash disappeared it did not fix the problem. Instead things continued to get worse and what happened is more and more bloat simply moved into JS browser software. Now people want to blame the browsers, but once again I’m not sure they’re the root cause. I suspect poor or careless software development methodologies may be responsible for much of the bloat we’re seeing everywhere. Now that RAM and CPU cycles became plentiful, we as software developers have become guilty of taking them for granted and neglecting even basic levels of optimization.
Alfman,
I know I am biased, but RAM is there to be used, not to stay idle.
Even when not actively used, it should be caching something (or pre-caching). That is why I found complaints against Windows using up “all the RAM” counter-productive. (I think they now just hide those values).
To say it bluntly: “I paid for the RAM sticks, and they should be utilized as much as possible.”
But of course “who decides what is the best use” is a good question. Except for disk caches which can be coordinated with the OS and apps, I am not sure there is an API for this purpose. “Hey, kernel! I have this pre-rendered HTML image, but you can discard it if necessary. Just let me know if you do it” would be a nice abstraction. And preferably with some “priority” tags. Like first discard “off screen” parts, then the visible bitmaps, than DOM tree, than raw HTML, etc.
(Maybe there is such an API, and I don’t know).
I think there are now web APIs for running code in the background. And those not registering as media players, etc, are lowered in priority.
And of course fine tuning is not as straightforward, since sometimes people really want the background app with full resources:
https://support.google.com/chrome/thread/160627926/is-there-a-way-to-disable-automatic-tab-discarding-for-specific-sites-in-google-chrome?hl=en
sukru,
Or how about this idea…let all processes request more than 100% of ram and then kill them off as necessary using semi-random heuristics. Sounds like fun right?
(sarcasm in reference to linux’s OOM killer if you didn’t pick up on it).
On a serious note though, I agree. There should be OS hooks/events for voluntarily reclaiming memory. QEMU has the memory bubble driver, which serves this purpose for reclaiming memory from within virtual machines, but operating systems need something at the application layer. Most applications will ignore it of course, but browsers could support it, and programs written in managed languages could do garbage collection.
I agree with the OP in your link that chromes behavior may not be desirable. Discarding tabs without prompting the user could result in some unfortunate data loss if people aren’t careful. Maybe “tab hibernation” would be a better option.
Alfman,
Yes, I have fond memories of “OOM killer” craze. They fixed the “fork() bomb” issue though.
and..
spoke too soon. Just killed the system with “:(){ :|:& };:” [warning: don’t try this at home]
Indeed. I’m consulting a lawyer now. They’ll be making battery killing applications available and not warning me. 🙂
I support this in principle, but I worry this is the end of non-Chromium browsers.
torb,
I also worry about the lack of diversity. IMHO the lack of browser diversity is strongly influenced by the lack of platform diversity. Consider that only a minority of users will change their browser at all. Bundling has been the major driver for browser usage by the masses. Therefor I strongly suspect safari will continue to be the dominant browser on iphone regardless of whether apple’s owner restrictions get eliminated.
Here’s the thing, apple repealing owner restrictions is obviously great news for those who want control over their own software experience…owners deserve to make these choices for themselves. However it doesn’t mean that most people are going to change their behaviors at all, I predict a transition will be fairly uneventful. It will take several years for competitors to really make a dent into apple’s market share. Apple still has the bundling home field advantage even without coercive restrictions.
100% correct. Bundling is the reason Edge is far and away the most popular browser on Windows!
that guy,
Yes, although microsoft’s circumstances are somewhat unique with antitrust cases putting a wrench into microsoft’s own bundling. This helped alternatives become competitive against IE. Google in particular paid vendors to bundle google chrome. on new computers instead of IE.
https://www.searchenginejournal.com/google-chrome-dell-sony/13055/
https://www.eweek.com/news/google-confirms-chrome-web-browser-is-bundled-on-sony-laptops/
Google was also paying developers to bundle google chrome into the installers of popular software. Together these bundling strategies worked to increase google chrome marketshare across windows, android & chromebooks devices.
Mozilla lacked an independent business model to pay for bundling, and so they’d continue to loose marketshare to google. I think the antitrust conditions imposed on microsoft are expired, and microsoft’s browser could make a comeback as people buy new hardware with edge bundled on it. This is a slow process because people don’t buy new hardware all that often, but assuming you have bundling rights, it is a proven long term strategy, at least assuming one can stay clear of regulators, haha.
This is good news for the FF mobile users. 🙂
Or bad, depending on the opinion of the FF mobile dev team.
I mean, they use their own engine on Android without issue, so if anything it seems like this would make their job easier, since they’d no longer have to maintain a build that strips out the engine.
Apple. Too little, too late as usual.
Lost the Eu market by launching a 2G phone in the 3G era and never caught up (except to none techy people).
May consider if the start running android apps and is possible to install google play (via usual sideloading like amazon I would expect). F-droid by default would also be good.
This should’ve happened a decade ago. At this point, I’m not sure if it makes a difference anymore. I think Apple is trying to appease regulators with minimum effort. Other than me, does anyone use Firefox?