Goodbye, Siri

It was the longest happy relationship of my life.

We slept together every night, and she woke me up every morning. She caught me up on the news, and whenever we drove somewhere, she gave me directions. If I missed a turn, she didn’t complain, just recalculated the route.

She was so easy to live with, her voice was pleasant, and she had a great sense of humor.

My heart breaks as I write this.

We broke up over virtual reality.

Siri on iPhone

I was a big fan. Siri — not so much.

In fact, aside from some very lackluster support for a limited number of google Cardboard apps, there has been no word out of Apple at all about support for virtual reality.

So, on Monday, I went out and bought the Samsung Galaxy S6. And oh, the places I’ve been able to go. So many more apps, built-in browser support for virtual reality, and the bigger screen and faster processor makes all the apps fly,.

maria@hypergridbusiness.com'

Maria Korolov

Maria Korolov is editor and publisher of Hypergrid Business. She has been a journalist for more than twenty years and has worked for the Chicago Tribune, Reuters, and Computerworld and has reported from over a dozen countries, including Russia and China.

  • So you base your selection of phone on the ability to wear it mounted in a cardboard box strapped to your head a few minutes a week? ;-))

    • Well, yeah — doesn’t everybody? Also, I just upgraded from a cardboard box to Sunnypeak.

      • Sorry I don’t even know what Sunnypeak is 🙁 I am sure it is good!

        As you know with Apple they never, ever announce anything they have in development before it is ready to go. Last year they took the development community by surprise after having worked for 4-5 years on the completely new programming language Swift and this year they announced Swift to go open source later this year, in addition to being available also for Linux directly from Apple.

        They do, however, talk about technologies at each year’s developer conference, and to get some insights into where they are going over the next year or so is the announcements for Gameplaykit, and updates to SceneKit and Metal. Metal is interesting because developers are able to get 2-5 times better graphics performance on the same hardware over OpenGL.

        The combination of the capabilities in Metal, Swift, Gameplaykit and SceneKit makes a full SL/OpenSim viewer running on hardware with the A8 or better processors feasible.

        Check out the videos at https://developer.apple.com/videos/wwdc/2015/ under Graphics and Games section.

        Swift: https://developer.apple.com/swift/blog/

        • Yeah, I am well aware of the universal law of contrariness… after holding off on switching to Android for years because I just liked my iPhone so much, right after I make the switch, Apple will announce its headset.

          Well, if they do, it will probably take a couple of iterations for their product to get good and wipe out everything else, and by that time I might be ready to upgrade again, anyway. 🙂

          • That’s the thing about Apple products – never buy version 1.0 or 1.1 even, and for the iPhone waiting to version 2 was a good choice for most.

            One thing working in product management teach you is that your customers are going to (try to) use your products in ways you never had thought about in the design or test phase – but that is also how products evolve and gets better if you respond sensibly to that challenge. Otherwise chances are they move on.

            … if you are going to watch any of the videos watch the Intro to Gameplaykit. It basically has the majority of building blocks to reimplement the OpenSim server completely in Swift.

          • Do you know if anyone is working on that? And by implementing the OpenSim server in Swift — does that mean that the viewers could be running on iPhones?

            Of course, there’s still the network lag problem of trying to use wireless communications to send messages back and forth — but I’m hoping, as people watch more and videos on their mobile devices, that the carriers will be speeding up the networks anyway. The fact that there are competing mobile companies bodes well!

            The Sunnypeak is the HMD I just got for the Android for $40. I really like it and just posted a review of it:

            http://t.co/AVgsRcfTty

          • IDK that anyone is implementing the OpenSim server in Swift (yet) – the tools are still in beta but I can see it is feasible. It would be a good opportunity to clean up outdated and inefficient code also.

            The viewer code is pretty much independent from the server code as long as they have a protocol set to communicate with each other. It might be there is scope for delegating more of the logic to the viewer to decrease the bandwidth need. SceneKit already supports bullet physics on the handset, so it is possible that even physics in combination with pathfinding can offload the servers significantly.

            I am interested in making a Swift viewer simply because the code sharing between OS X and iOS version would be significant (not completely identical) and you could most likely tuck on that a Linux version too.

          • Cinder Biscuits

            Everything that can be done in swift was already possible with Objective-C. The reason nobody has done it is there’s not really anybody who wants to tackle the entire linden stack on their own. Hell, I’m basically the only one willing to put any real effort into Mac viewer development at all. If I could get my hands on an iPhone there’s a pretty good chance you’d see something FOSS pop up sooner or later, but it wouldn’t be written in swift or with SceneKit.

          • I agree that everything that can be done in swift was already possible with Objective-C as far as application development, but Swift has the potential of attracting developers who would not go through the pain of learning Objective C. The upcoming Linux release will also put it in front of a lot of people who would not touch Objective C.

            Swift being classified as a system level language also ensures it can be used for a very wide range of tasks, including (one day) compiling OS X in Swift. This is also why I think it is possible to redevelop OpenSim server in Swift getting much better performance out of it. Apple also tends to optimize the heck out of their tool-stack, and already now Swift will get you much better performance than Objective C for certain operations.

            As for getting your hands on an iPhone, you can do a very large section of the development and testing in the emulator in Xcode, and it is only when you need to test and optimize performance you have to work on real hardware. I ma not sure I would start an iOS development project targeting the iPhone because of the limited screen real-estate, but it might be possible to use auto layout for a unified iPad/iPhone interface.

          • Cinder Biscuits

            Swift really isn’t that great of a language, and I don’t know why you’d want to use cocoa for a server. That’s pretty silly. At YrGrid we’ve written several of the robust services in golang and C++, and I know Inworldz has their Whip asset server written in C++.

            And yes, I’m well aware of how to write apps in Xcode, but there’s really no point in me spending several thousand hours writing a piece of software I can’t enjoy myself, is there? Us FOSS developers are just so selfish that way.

          • I think you are a bit misguided; Swift is a system level language that can be used to write utilities, terminal based applications, servers, operating systems or apps that use Apple’s own frameworks like Cocoa Touch.

            The Linux version will not ship from Apple with Cocoa, but will more likely be integrated in IBM’s Eclipse as a general purpose language for writing server and business applications + code sharing with mobile apps that needs business logic from IBM’s enterprise solutions in the mobile space.

          • Cinder Biscuits

            I know exactly what swift is. I know how to use it and have done so, but there’s not really any sense in using it to design high-concurrency server applications when there are other “system level languages” that fit the bill a lot better.

          • Why even engage in a discussion on something you don’t have any intention to do, use or get involved with? Is it just to protest or make some other “important” statement as a FOSS developer?

          • Cinder Biscuits

            Why not?

          • they call it opinion, and sometimes people cast opinions out of thin air.

          • gabegw11

            I totally agree with you on this.

          • I’m not sure emulator can be good idea if majority of functionality is OpenGL-ended.

          • With a complete rewrite it would not be based on Qt/OpenGL, and the rendering would run on Metal and not OpenGL. You would of course have to test the rendering on real hardware and not in the simulator. Network stack + a few more things can also not be tested in the simulator.

            https://developer.apple.com/metal/

          • gabegw11

            I really hope that we can get some better mac viewers all the current ones are just bleh and have issues.

          • Yes, but it requires the same rewrite/codebase. The current Mac viewers are both based on “alien” frameworks, old libraries and open source functionality such as for handling audio that is not at all optimized for OS X.

            I am helping the Kokua developer to get the Kokua viewer to be compiled with Clang (same as the SL 3.8 viewer), but that does not change the fundamental issue for the Mac version of the viewer.

          • Cinder Biscuits

            The performance problems within the current linden based viewers for mac (not Radegast, which is a whole other world of issues) has little to do with the frameworks used or the age of the libraries and more due to Linden’s render pipeline just plain sucking and Apple’s up-until-now ambivalence towards OpenGL. Metal in the viewer would be *huge,* but it’s going to take some developers with guts to fork far enough away from Linden Lab to get there though, but they’ll never go there themselves.

          • I agree LL is not going to go Metal for the current viewer, but if they are anything but serious about next generation they will.

            It might also be that the current viewer would get a boost from being compiled with OS X 10.11 target as it will to some extent use Metal when available even if not explicitly coded for it. I have not had the chance to test the current viewers in 10.11 yet.

            The issues with the current Mac viewer is not only the rendering pipeline, but also Qt that is a major drag on it and makes it behave in the same sucky way as every other “Mac” version running on Qt. DAZ Studio is a very good example of how bad it gets.

            Even details like keyboard combinations drags it down because you have to unlearn a lot of keystrokes that is programmed into the spine of Mac users being used (almost without exception) consistently across applications.

          • Apple had been unwilling to go from their comfort zone with OpenGL – it is not _their_ play, it is group/consortia play and they’re known to play weak in “team”. In a way, it is old good msft model – my way or highway. Another reason is that OpenGL (and it is twin-but-not-little-anymore sister OpenGL ES) scene is heavily stomped by IPR. And IPR for Apple is not just fleecing but reaping off meat from their cash body. But making two sided scene (OpenGL and Metal) might not work as they expect – after-Jobs Apple is less and less tiger, more and more hippo. And… Think of how will grow more-than-twice budgets of R&D for any serious graphic product just because apple is different. You have studied business side of software/systems business and know by heart – R&D costs are sunken costs, one needs serious grounds for. To sum it up, Apple pulls consortia blanket on itself know, but there is no guarantee they could succeed, moreover my opinion blanket (existing consortia behind OpenGL) will pull them back – they are not interested in Apple’s future. And that industry is big, big enough to asphyxiate and subdue even Apple.

          • A test build of the Kokua viewer for Mac is available.

            It has been built with the Clang compiler in Xcode 6.4 and is based on Linden Lab’s 3.8.1 code base with the OpenSim additions and general improvements in the Kokua viewer. You need OS X 10.8 or higher to run this version of the viewer.

            It has been tested on all the major OpenSim grids both with direct login and hypergrid teleports in addition to SecondLife.

            It can be downloaded from http://blog.kokuaviewer.org/2015/07/16/kokuant-mac-version-is-available-for-testing/

          • lmpierce

            Thanks for the notification and the link!

  • Simply put… Samsung kicks Apples butt! I had a Samsung s2x that could do more things than an iPhone 4 and at 1/3rd the price to start. My friend had a i4 and was impressed by the camera on my phone… and even more so when he found out it was a Samsung. I just upgraded to an s4 btw, love it!!