One of the reasons they gave for the delay of switching to wasm was that nacl had threading support while wasm didn't (wasm supports running in workers, but communication can only happen asynchronously through message channels, no advanced multi threading capabilities like sharing memory etc). Then spectre happened and delayed the wasm threading support of browsers even further. Chrome has experimental wasm threading now, but other browsers don't. So now the non-Chrome launch happens without usage of wasm threading.
Because Apple still doesn't support WebGL2 (similar to OpenGL ES 3.0) after all these years. Apple is involved with WebGPU, but they still won't even support WebGL2, which kinda pisses me off.
No matter what people say, Chrome is not the new IE, Safari is; and thank god it's so unpopular.
You misunderstand the term Chrome is the new IE. IE used to be the dominating browser on the web. Whatever standards IE supported were reverse engineered by other browsers such as AJAX. IE eventually got overshadowed by the open web and fell behind. There used to be websites that only worked on IE, well there still might be. There's too many websites coded and tested with Chrome now using features only in Chrome, breaking other sites, or worse checking if Chrome is being used and ignoring if the browser supports the feature, or worse Chrome implements it before a formal spec, which breaks other browsers who comply with the spec.
or Microsoft Silverlight, and soon Flash will be finally deprecated too, and I'm sure it will still be present in enterprise training modules and other nonsense
I think both Chrome and Safari have negative connections with memories of IE. For Safari it's the stagnation of IE6: installed everywhere but rarely updated with new features that other browsers have.
That's not so much of a problem on desktop because people can always install another browser, but on iOS you're literally not allowed to use any other browser engine.
The issue is that major updates to Safari follow the same cycle as updates to the MacOS and iOS operating systems, meaning once per year at the current rate. Adding major features once per year (and even then, only once they have been in use on other platforms for at least a year or two) means that Safari can lag by three years in terms of features (case in point: WebRTC).
> means that Safari can lag by three years in terms of features (case in point: WebRTC).
In the case of WebRTC, it's even worse than that: public stable releases of Firefox and Chrome supported WebRTC by early 2013, but Safari took until late 2017. That's more than five years for an extremely popular feature (I had three professional engagements where we had to integrate a native WebRTC client component into a Cordova application, and couldn't support that feature on iOS through the browser, before they finally released it).
As my123 kindly pointed out above, it is possible to enable experimental support for WebGL2 in the current version of Safari. I presume this is something they are cooking up for the next release of MacOS.
> it is possible to enable experimental support for WebGL2 in the current version of Safari. I presume this is something they are cooking up for the next release of MacOS.
No, they have had the same level of support (it fails almost a quarter of the conformance tests, lacks a majority of the new APIs entirely, and doesn't even reliably take in GLSL ES 3.0 shaders without crashing even more than the GLES 2.0 implementation) for four years, without further activity. It is not conformant in the TP, and I don't think they plan to ever make it conformant or actually release it.
They have the resources and the expertise in house to make this a thing, they could literally just use ANGLE, but they choose not to.
At least in regard to WebGL, "X is the new IE" isn't the best way to look at it.
WebGL2 is based on OpenGL patterns. I don't think it makes sense to invest new development in that, whether you're Apple, a framework developer or an app/game developer. I think the writing was on the wall for OpenGL even before WebGL 2 was finalized, so a reluctance to embrace it makes sense. I can see how it would be negative in the short term, but I think we're at the point where web apis should be considering the long-term strongly, and in the long term webgl 2 support isn't good.
To me the Google Maps and Earth might be the most impressive piece of software in existence. To see it working on standard web technology on competing browsers is awesome.
It was so amazing when XMLHttpRequest was a somewhat obscure thing, and they had fully interactive maps while the rest of the world was Mapquest, and "click" "new page" "click" "new page" to be able to scroll and have new images load. I think google jumping on the AJAX train and trailblazing is under credited in the history of their rise.
Thank you Microsoft and OWA teams for inventing XMLHttpRequest. Anyone coming into web/development now missed the evolutionary leap from HTML to DHTML/AJAX. Google owes the OWA team a thank you for sneaking that feature in under the radar.
As someone on here aptly described it, AJAX was a "historic own goal" by Microsoft, who was desperately attempting to separate IE from all other browsers by giving it unique features.
+1. I've used it before going on trips or journeys. When I'm using VR, it causes my brain to remember it as if I've been there. When I'm finally there it just feels like revisiting, and I know where I'm going.
Wait, what? I switched completely from Chrome to Firefox in mid November (so happy about the move btw) and had Google Earth on Firefox from day one. Is this article missing something or is it me?
This is not facebook, this is HN. You can't possible conceive that a user of HN does this confusion, do you? Because in that case you are not only insulting my intelligence, you're insulting yours as well. Also a simple click on my name and reading a few answers I've given here could tell you I am definitely NOT the usual facebook user who somehow got lost and ended here.
Let me tell you again, so there isn't any confusion: Starting mid-November last year when I switched from Chrome to Firefox, I had google Earth from day one on Firefox. And yes, I am very much aware the difference between not only google earth and google maps but also the difference between google maps 2D and google maps 3D possibilities. I have kids, I use all these 3 tools when my kids are doing their Geography homework. Is it clear now?
Well, my apologies for trying to help, I guess. I try not to make any assumptions about your background knowledge, and am not going through everyone's comments before replying to them.
I know this is unasked for, but I do think it would save a lot of frustration and make for more pleasant interactions if we try to assume other people have good intentions.
I'm serious though: I've seen people refer to Google Maps as Google Earth, even though they're separate products - and it's not clear to me whether you're not aware of the former or of the latter.
When Google Earth web used NACL, it was so silky smooth, it wasn't even like it was the web. Even on relatively weak hardware (dual core laptop), it was so nice, didn't skip a frame. First time I tried this in stable Chrome it locked up my entire computer. Took about 5 minutes to be able to reboot. On reboot, when I run it, it's like a slideshow, unusable (Ryzen 3900X 12-core beast with RTX 2060 Super). I really want to get excited about WebAssembly, but this doesn't do that.
Today I tried it for the first time (the web version) on my work laptop T490 (Intel UHD 620), Ubuntu with Firefox 73. Everything is super smooth. I'm surprised it isn't for you.
It's also smooth on my home Macbook Pro (2015) w/ Fx73.
I'm surprised too. I've tried on Firefox as well, and on different machines, dual core laptop (Intel), quad core laptop (Intel), iMac 27" Late 2015. All were silky smooth with NACL.
This is my usual reminder that Google Earth for desktop is still available and is still wildly superior in terms of features and performance to the web version.
(scroll to the bottom, select 'download earth pro on desktop')
Edit: Though to be fair, I just checked out the web version and it does appear to be much better than it was when it was first released. It does now appear that they're actually trying to re-implement most of the features of the desktop app on the web. The lack of historical imagery and the replacement of the 'layers' model with a couple of 'map style' options are the most glaring omissions.
Performance is still remarkably better on the desktop app and will be for the forseeable future. I used to run the software over a decade ago on a modest dual core computer with 4gb of RAM without struggle; try doing that with a modern
browser experience.
45 comments
[ 4.5 ms ] story [ 96.1 ms ] threadYou will have WebGL2 in the menu.
edit: Didn't explicitly say it but you should choose the single-threaded WASM option when asked.
http://trac.webkit.org/browser/webkit/trunk/Source/WebCore/h...
Search for "NOT IMPLEMENTED". You'll see 80+ WebGL2 apis are not implemented.
No matter what people say, Chrome is not the new IE, Safari is; and thank god it's so unpopular.
Work in a large corporation and you will see plenty of these.
That's not so much of a problem on desktop because people can always install another browser, but on iOS you're literally not allowed to use any other browser engine.
In the case of WebRTC, it's even worse than that: public stable releases of Firefox and Chrome supported WebRTC by early 2013, but Safari took until late 2017. That's more than five years for an extremely popular feature (I had three professional engagements where we had to integrate a native WebRTC client component into a Cordova application, and couldn't support that feature on iOS through the browser, before they finally released it).
No, they have had the same level of support (it fails almost a quarter of the conformance tests, lacks a majority of the new APIs entirely, and doesn't even reliably take in GLSL ES 3.0 shaders without crashing even more than the GLES 2.0 implementation) for four years, without further activity. It is not conformant in the TP, and I don't think they plan to ever make it conformant or actually release it.
They have the resources and the expertise in house to make this a thing, they could literally just use ANGLE, but they choose not to.
WebGL2 is based on OpenGL patterns. I don't think it makes sense to invest new development in that, whether you're Apple, a framework developer or an app/game developer. I think the writing was on the wall for OpenGL even before WebGL 2 was finalized, so a reluctance to embrace it makes sense. I can see how it would be negative in the short term, but I think we're at the point where web apis should be considering the long-term strongly, and in the long term webgl 2 support isn't good.
Thank you Microsoft and OWA teams for inventing XMLHttpRequest. Anyone coming into web/development now missed the evolutionary leap from HTML to DHTML/AJAX. Google owes the OWA team a thank you for sneaking that feature in under the radar.
https://adobe-flash.github.io/crossbridge/
https://www.youtube.com/watch?v=81BrnktI4Gg
Let me tell you again, so there isn't any confusion: Starting mid-November last year when I switched from Chrome to Firefox, I had google Earth from day one on Firefox. And yes, I am very much aware the difference between not only google earth and google maps but also the difference between google maps 2D and google maps 3D possibilities. I have kids, I use all these 3 tools when my kids are doing their Geography homework. Is it clear now?
I know this is unasked for, but I do think it would save a lot of frustration and make for more pleasant interactions if we try to assume other people have good intentions.
Google Maps: https://maps.google.com/
Google Earth: https://earth.google.com/web/
https://www.google.com/earth/versions/
(scroll to the bottom, select 'download earth pro on desktop')
Edit: Though to be fair, I just checked out the web version and it does appear to be much better than it was when it was first released. It does now appear that they're actually trying to re-implement most of the features of the desktop app on the web. The lack of historical imagery and the replacement of the 'layers' model with a couple of 'map style' options are the most glaring omissions.