So... It's a beta and some limits aren't up to native spec because of things like the UIWebView currently being backed by an OpenGL ES 2.0 surface (instead of 3.0, which is only available on the iPhone5s)?
Did not get that apple was finally unleashing webGL on mobile Safari, that's awesome! Thanks for the investigating on the beta and helping in making it better. I can't wait to be able to have Three.js working on iOS.
Though, I feel disappointed that the promises of "web app" were not really held so far (just look how you can't even open an external link in the browser from a webapp now), in favor of native apps, and hope that webGL in the browser isn't gonna be just another bittersweet thing.
I guess we can simply wait to see how the tickets you linked to will go to have a good sense about that.
Apple's support of OpenGL has been notoriously bad performance and feature wise for quite a while (at least on OSX, not sure of the mobile story), so I'm not that surprised that their webgl implementation isn't brilliant. It would be the most un-apple thing ever, but I wish they'd just leave it up to the device manufacturers to provide a GL implementation directly in the same way microsoft (generally) does. Apple has so far proven to be awful at it.
Apple has made it known through its chip design (and subsequent underclocking), app development limitations, etc. that battery life is of utmost importance. I suspect that some of these limits are an attempt to keep websites from draining mobile devices of energy.
Is WebGL over a browser typically as performant as native code? I would suspect no, but I'm not familiar enough with WebGL to know.
It's unlikely to be battery concerns. If anything, lack of extensions and low limits means webgl code would have to perform multiple passes of a scene to achieve the same visual effects, thereby increasing the battery drain. Also, why allow these features for native apps if it's not good for the battery?
And if you click through to the webkit bugtracker, it looks like almost all the issues listed already have patches committed.
Performance wise it's about the same. The work is mostly done by the graphics card, so if you're able to feed it enough data without stalling it it doesn't particularly matter what platform you're invoking it from.
While, Firefox has been crashing for the past few versions, both Chrome stable and canary have been completing the WebGL Conformance Test without crashing. There is probably something wrong on your end, if it won't run on your rig.
I think its fascinating that the embrace and extend (and wall-off in a garden) strategy being applied by Apple is still producing, nevertheless, a very pliable audience even in this day and age.
I see the 'heading off of open standards' play of Apple in a cynical light, because I believe Apple are playing platform games here. These breaks/incompatibilities are another reason for developers to not bother with other platforms, or at least limit themselves in a way that makes platform-dependent tooling inevitable. Basically, it's developers getting screwed by hardware guys, over and over again.
Cross-platform is not the be-and end-all in software.
Cross-platform has the covenience (and bigger market), but it's also crippled by catering to the lowest common denominator. You won't get stuff that takes full advantage of a platform if it's cross platform.
Some people like native dedicated apps. I, for one, like cross platform stuff for common infrastructure (servers, unix userland, etc), but prefer platform specific and integrated desktop apps that take full advantage of my OS (whether OS X, Windows or Linux).
It means let Apple do all the GUI work for you. And then let them change their designs, and the design of your app of course, according to their whims.
Its okay, I understand I'm getting downvoted for not following along with the groupthink that Apple can't do anything wrong in the GUI department, and therefore native is far superior to anything else, but I've still got the opinion that the native GUI toolkits/frameworks for each platform are really designed to lock users - and developers - into the platform. Far better to avoid this issue completely, and do cross-platform development with custom GUI controls, if you want to have an app that works wherever your users will run it ..
And look terrible on all platforms while at it. Yes, a much better solution indeed. This is not to say Apple UI is good or bad. This has nothing to do with it. Users expect native look and feel from apps. I assume you are a Linux user, where there is almost no consistency to software (outside of teams purposefully designing software to look consistent, such as Elementary OS), but this is not how it is on Windows and Apple operating systems, and for the most part on Android. If you don't like the ecosystem, that's fine - don't develop for it, but don't dismiss something that users expect from this ecosystem.
You're not getting down voted for not following groupthink. You're getting doe voted because you're making a bunch of assertions of your opinion as though it's fact, without actually trying to justify those opinions.
On the substantive point, a couple of years ago I wrote a graphics framework. It used JavaScript core and directfb. I could have just told the app developers to use WebKit, but the performance just wasn't there. I'm pretty sure that's also the reason that apple, Google and Microsoft write native GUI sdks, and not to lock in developers.
>Its okay, I understand I'm getting downvoted for not following along with the groupthink that Apple can't do anything wrong in the GUI department
I think this "persecution complex" is more groupthink that what you accuse as groupthink. There are tons of discussions on HN, and highly voted comments, about Apple's errors in GUI work (and elsewhere). They just happen to do a better job than 80% of the developers out there.
Furturmore, I disagree with the two basic premises of your comment:
1) Native doesn't mean "let the GUI to Apple". You conflate native with "follows a platform's GUI guidelines religiously".
You can create a great native app, using Cocoa APIs et al, that still looks totally different from Apple's apps. Take for example Paper for iOS. Or Procreate. Or Elements. Or tons of stuff, really.
You don't have to follow the GUI look at all, you can even have your own widgets for stuff. Just don't try using Swing in a Mac app, or some shitty Cocoa port on a Windows app. And let the common behavior of items like dropdowns and text-entry boxes be the same (e.g with regard to keyboard shortcuts).
But for those that do want to "leave the GUI" (well, look) to Apple, it's because it saves them tons of fucking work. Redundunt work of having to implement common stuff for themselves and their app.
2) Second, I don't think "native GUI toolkits/frameworks for each platform are really designed to lock users - and developers - into the platform" makes much sense.
What DEFINES a platform is the very presence of such things (native GUIs/frameworks, etc).
Without those, what you ask amounts to OS landscape where everything is the same generic base, sort of like FreeBSD/NetBSD/OpenBSD/etc but also at the desktop level.
What's the sense in having "platforms" for desktop OSes at all, if that's the case?
If you just want one OS to rule them all, for everybody, just ask for that. It makes no sense to want to have different platforms but still be bound to not have unique APIs and frameworks for each.
>It makes no sense to want to have different platforms but still be bound to not have unique APIs and frameworks for each.
The point is, that this is an artificial arbitrary, a line drawn in the sand by market - not technological - forces, and to play along with it is to pander to the divide and conquer strategy begin applied in the effort to capture developer mindshare. Why deal with it at all, when .. with a little effort .. a cross-platform solution that works effectively on all platforms is a tangible, viable product? Just consider the amount of work that had to be done to update apps for the recent changes in iOS' look and feel, and consider what life would be like if instead the effort was made on a cross-platform solution that put the application-standard GUI together, driven by the application-developer, not OS/hardware vendors whose motives are entirely driven by keeping developers away 'from those other platforms'.
I've built big Android and big iOS projects. Neither one of them is better than the other, in any sense whatsoever; both platforms have their thorns and pitfalls, and maintaining two separate codebases, with all the pitfalls, is just counter-productive. Better to have one codebase that fits all and keeps the productivity where it needs to be: on the application, and not on keeping up with the joneses ..
>The point is, that this is an artificial arbitrary, a line drawn in the sand by market - not technological - forces
That's wrong. Platforms are different, and offer different APIs and frameworks, because of technological forces too, not just because of some corporate conspiracy for lock in.
Case in point, even Open Source OSes offer different platforms and incompatible frameworks. Not all are just a UNIX/POSIX clone.
VMS has a different philosophy to UNIX. Lisp Machines even more so. Windows. BeOS. Etc. And that's the core OS. The same goes for platforms and desktop level stuff.
Having the same stuff (APIs, GUIs etc) work in all platforms the same, kills diversification and competition between technological implementations and design philosophies.
>I've built big Android and big iOS projects. Neither one of them is better than the other, in any sense whatsoever; both platforms have their thorns and pitfalls, and maintaining two separate codebases, with all the pitfalls, is just counter-productive.
Well, that's like your opinion man. I find iOS vastly superior for most of my uses cases.
I surely don't wish iOS and Android got merged into the same codebase.
