News for hackers. Stuff that matters.
Re:Microsoft Full Circle by WorldMaker on Tuesday October 18, 2022 @12:34PM (#33250067)
Microsoft is "early" to VR because it was already a player in VR and has very long been "early". Microsoft has been investing in AR/MR/VR research since before the original Surface, not the tablet one but the giant table that got rebranded and then sold off, and since before the original Kinect. Microsoft is still making plenty of HoloLens devices (some on huge contracts) despite them not being "sexy". Windows 10 has had AR/MR/VR apps bundled for a long time despite most users never using them.

It's hard for it to be a "full circle" when for the most part it looks like a straight line. The surprise shift here isn't in Microsoft but in Meta finally asking for help and partners after seeming to believe they could do it all on their own.

Re:Microsoft Full Circle by karmakaze on Tuesday October 18, 2022 @01:31PM (#33250998)
Initially I thought MS might have learned from missing out on mobile to get ahead in AR/MR/VR and be a device and platform leader. It seems what they learned is to give up before entering and focus on partnering instead. That is something they've always succeeded at PC+DOS, IBM+WindowsNT, Apple+Office, JavaVM.

To me it does seem a low target to aim for. VSCode shows they can execute, but for some reason when the stakes are higher they don't do as well Windows Phone, Edge browser, etc. I suppose I could call that corporate self-awareness.

Re:Microsoft Full Circle by WorldMaker on Tuesday October 18, 2022 @02:00PM (#33251428)
Here specifically too: Microsoft has such a weird lead in AR/MR/VR and have mostly just been sitting on it. For example, because of early Kinect work, they own the patents that would have made Meta's weird journey between "no legs" to "everyone gets legs if they want them! excitement!" irrelevant or much faster had the partnership happened sooner. They built the first "metaverse cloud" on Azure and they've used it and they've made something as exciting in theory as the HoloLens boring Enterprise products.

The other possible takeaway is that after Windows Phone, Microsoft really has been truly risk adverse when it comes to consumer products, even in areas where they had a lead and could have delivered something no one else could.

> VSCode shows they can execute, but for some reason when the stakes are higher they don't do as well Windows Phone

Windows Phone was arguably very well executed at the software level. Microsoft's troubles with Windows Phone were at the hardware levels and (worse) the economic levels where the phone carriers had already decided they prefer a duopoly and selling three things is one too many things to sell.

Re:Microsoft Full Circle by pjmlp on Tuesday October 18, 2022 @02:56PM (#33252190)
Windows Phone was bad executed at development experience level, with the WinRT, UAP, UWP enforced rewrites.
Re:Microsoft Full Circle by WorldMaker on Tuesday October 18, 2022 @03:15PM (#33252457)
We've had this argument more than once and I still disagree with that take. I took a .NET application through those transitions and there was never a "from scratch rewrite" step at any point. A lot of cheese moved and there was sometimes quite a bit of XAML shifts to meet new UI expectations, but most of the business logic code stayed the same or barely changed. On the other side, I've seen Apple enforce equal or worse "enforced" rewrites between major iOS versions with not even a fraction of the complaints Microsoft got across those product versions.

(I had more inconvenience moving apps back and forth between WPF and Silverlight in some of the years prior, and even those moves were never from-scratch rewrites. Admittedly that may have left me with more skills to handle XAML changes and write to lowest common denominator XAML, so I am aware of my biases.)

I realize you aren't ever going to agree with me on this, and I don't expect you to. I do think it is still useful for other people to see that there are counter-opinions here.

Re:Microsoft Full Circle by int_19h on Tuesday October 18, 2022 @04:09PM (#33253159)
Apple can get away with a lot more given its dominant market position (esp. when it comes to delivering paying customers to app authors). But when you're the new scrappy contender with minimal market share, so you need the devs to make apps for your store rather than the other way around, it better be smooth sailing all around.
Re:Microsoft Full Circle by pjmlp on Tuesday October 18, 2022 @04:44PM (#33253621)
A rewrite is a rewrite, even if only one line changes.

Not going to bother to list everything that changed, it is public available for anyone to judge.

Re:Microsoft Full Circle by karmakaze on Tuesday October 18, 2022 @07:48PM (#33255625)
The difference of course is that iOS already has enough adoption to get away with this. Windows Phone and Windows UI development forked for no good reason while keeping very similar names and structures while trying to get initial adoption. It was clearly a mistake at that stage. I personally looked into the tech and as soon as I saw how fragmented it was realized MS didn't have a good plan and didn't venture developing for it.
Re:Microsoft Full Circle by WorldMaker on Tuesday October 18, 2022 @08:13PM (#33255808)
Windows Phone 7 was pretty clear [1] it was a temporary platform (bootstrapped on Silverlight) while they waited for the engineering projects that were to become Windows Phone 8 finished. Microsoft helped pay to migrate some projects across that hump and there was a migration story.

That's the only "hard fork" between the versions. I know a lot of people (including respondents in neighboring comments) saw the changes between WP8 and WP10 as a big deal "soft fork", but very little actually changed other than a UI library upgrade from 1.x to 2.x. (In .NET; JS was dropped in 2.x and C/C++ developers saw some other shenanigans. But JS and C/C++ didn't exist in WP7 either so none of those developers saw all three "migrations" either.)

I hear a lot from developers that it seemed very fragmented, but working in it at the time, it seemed very "incremental" at best, mostly just upgrading library major versions and dealing with semver major breaking breaks. Sure, that's frustrating when that's not your primary development environment and also you aren't getting paid enough for it because adoption is low, but that is a very far cry from "Microsoft didn't have a good plan". They were iterating hard on it. They made some semver breaks that other library authors in other ecosystems could get away with no problem, even with equally low adoption (looking at random JS libraries here). It's easy to blame low adoption for making those semver major breaks seem more painful and "fragmented" than they actually were in practice, but sometimes it just also looks like that was an excuse to ignore the platform more than a problem for development experience reality. (As you seem to state that you skipped the platform entirely, possibly because you were looking for an excuse to skip it and latched onto the first one you saw.)

[1] Obviously not clear enough for some.

Re:Microsoft Full Circle by karmakaze on Tuesday October 18, 2022 @11:14PM (#33257141)
IIRC, there was also dev kits for Windows RT around the same time with the same or similar names which added to incompatibilities and confusion. The only sharable source was non-UI core code, or perhaps hand-altered variations of UI file type contents.

It's one thing to develop versions for Android, iOS, and Windows Phone. It's another if the UX source elements can't be shared within the MS ecosystem Phone, RT, x86 metro apps.

Re:Microsoft Full Circle by sho_hn on Tuesday October 18, 2022 @01:43PM (#33251163)
That's not the point of the title. The point of the article (and their coverage of MS for years) is that Microsoft used to be all about Windows: An operating system that abstracts hardware and enables first- and third-party apps. And now they making themselves about a system that abstracts hardware and runs apps again, but across a spectrum of devices and front-to-back. The common element is now the user and their account, not the single PC it runs on.

I also think it's obvious, but then salient things are. Consumers today don't decide between PC vendors and operating systems, but between entire company-branded computing ecosystems. Are you on the Google or Apple or Amazon train?

Things Stratechery doesn't discuss enough and this article is poor about:

- Unlike in the Windows days, in the new competing-ecosystems world, geopolitics plays a much grander role. Everyone used to be on Windows, not everyone is on Google. China, Korea, Russia all have their own ecosystem juggernauts. In our globally interconnected world, computing has become somehow less globalized recently.

- Microsoft doing VR stuff on Facebook's platform is totally in line with their observations on Microsoft's strategy going forward, but Microsoft and Facebook also compete on the ecosystem front, and the implications of this are just brushed over here as that's not the article, but it would be a more interesting one.

- The really interesting thing that sets these different companies apart is their revenue sources. Apple sells hardware (and content and lifestyle), Microsoft sells enterprise software subscriptions and cloud, Google sells ad space and cloud.

Re:Microsoft Full Circle by WorldMaker on Tuesday October 18, 2022 @02:10PM (#33251574)
> Microsoft used to be all about Windows

This feels like a Citation Needed assumption both in this comment and in the article. Again, there's too much of a straight line here: Microsoft has always been one of (if not the) largest software development companies for the Macintosh. That has never changed.

We all know Ballmer was a huge Windows cheerleader and "home team supporter" and wanted to make sure that the best experiences were always on Windows. But he never killed Macintosh support. He championed the iPad versions of Office that Nadella is credited for (which the article even points out!). The closest Microsoft ever came to being "Windows only" was maybe the brief glimmer of optimism under Nadella where the Windows 10 team briefly imagined that every device could run Windows (and then felt the crushing disappointment to realize that most wouldn't).

Again, it doesn't look like a full circle to me but a mostly straight line with a few zigs and zags if you zoom in micro-enough. I understand where some of this is coming from that Microsoft did have years, especially under Ballmer, where they seemed to care too much about Windows, but Microsoft was never "just" The Windows Company in any era.

Re:Microsoft Full Circle by sho_hn on Tuesday October 18, 2022 @02:12PM (#33251598)
Yes, this is fair.
Re:Microsoft Full Circle by MichaelZuo on Tuesday October 18, 2022 @01:55PM (#33251353)
> Everyone used to be on Windows

Not true. IBM was one of the top 10 most profitable companies in the world as recently as 2015.

After adjusting for real inflations rates, inclusive of asset inflation, they actually made greater profits some years then either Google or Meta or Amazon do today.

Re:Microsoft Full Circle by sho_hn on Tuesday October 18, 2022 @02:02PM (#33251468)
You're of course right, I was exaggerating for effect. What I mean is that the PC stole a lot of other people's lunches, and PCs across the world had an almost equal likelihood across the world to be running Windows, and that mattered when the PC and what you could run on it (i.e. the app format) mattered.

Now it doesn't matter all that much anymore if you're on a Mac or a Windows or a Linux computer. You can largely use the same apps and services across them (with the exception of gaming, which, as an interesting aside, is now much better on Linux than on Mac). Anyone expending the effort to make a new app will usually build them that way, while in the past, you had a lot of platform exclusives.

But what is new and interesting is that a Google account doesn't deliver you as much value in South Korea because a lot of its services are unavailable or bad there, while it's nigh on impossible to live without a Kakao account, which in turn is entirely optional in, well, most everywhere else.

I find it fascinating that the global reach of user-centric/service-oriented computing is less than the reach of the previous device-oriented computing. It's certainly also not true that geopolitics have never mattered before in computing and reaching users, but I'd say they matter more than ever before.

Re:Microsoft Full Circle by MichaelZuo on Tuesday October 18, 2022 @02:49PM (#33252109)
To me this difference is to be expected.

When things were device-oriented whoever had the better devices naturally got easy footholds in every country. Since any country that tries to outcompete the world leading products would have needed to invest vast sums of money. And even then 90% of the supply chain would still be in the typical locations.

It simply made sense for everyone to leave the multi-billion dollar investments to a few multinationals, at least for consumer products.

The only real exception being Japan pre-iPhone, when they had the manufacturing diversity to make nearly everything domestically and enough wealthy consumers who also didn't care much for international interoperability. And even then the major Japanese firms were making very slim profits.

Whereas since the entry bar for creating competitive software is so much lower, many countries can afford to nurture genuinely homegrown products, and have a reasonable expectation that they will be competitive.

Re:Microsoft Full Circle by soundnote on Tuesday October 18, 2022 @06:29PM (#33254936)
> Whereas since the entry bar for creating competitive software is so much lower, many countries can afford to nurture genuinely homegrown products, and have a reasonable expectation that they will be competitive.

I don't think it's that. It's that eg. Kakao has KakaoTalk, a chat app. And chat apps tend towards monopolies and oligopolies because of network effects. The value of the app is less in how good it is, and more in who else uses it.

If you're just buying a music player or an office suite for a common file format, what you use doesn't really matter to anyone else, and their choices to you.

Chat apps, collaboration platforms? Suddenly, who else uses them matters a lot, since it's about the people, not the physical function. And once you get big enough, the network effects turns from a hurdle into a gigantic moat.

Re:Microsoft Full Circle by MichaelZuo on Tuesday October 18, 2022 @07:14PM (#33255359)
> If you're just buying a music player or an office suite for a common file format, what you use doesn't really matter to anyone else, and their choices to you.

The question of how many people use any given physical product, such as music players, certainly matters to supply chain executives and factory managers at the very least.

Because the same network effects applies to manufactured products, notably in economics of scale.

More generally, it's true that KakaoTalk benefits from network effects, but Facebook Korea could have benefited from the same.

It's just that a homegrown Korean chat app is viable against Facebook Korea because the entry bar is lower, whereas a homegrown Korean iPod would have been laughable against Apple, even in Korea.

There's very few people that would have paid 5x more for an all Korean iPod, but most are fine with being slightly inconvenient by Kakao's smaller international userbase.

Re:Microsoft Full Circle by chaostheory on Tuesday October 18, 2022 @04:16PM (#33253254)
> The surprise shift here isn't in Microsoft but in Meta finally asking for help and partners after seeming to believe they could do it all on their own

This isn’t a surprise given the sales numbers of hardware and DAU of meta online properties. While they are the clear leader in VR, given how great the Quest 2 and its original pricing were it should have had much higher sales numbers and usage. That result also wasn’t a surprise given that the Facebook brand was so toxic to the masses, that even changing names to meta hasn’t mitigated it yet in the short term. The name change will pay off eventually for younger generations who aren’t very familiar with Facebook, but that day is about a decade away