Hacker News Reader

Comment by optforfon | Hacker News Reader

I feel like I'm not seeing the forest for the trees.

I swear Microsoft's business model was to bring people in to their ecosystem and then charge them for what is hopefully a superior development experience.

Is there a new business model? How do they lock people in and squeeze them for money? Is their cloud solution radically different from the competition that they can put all their eggs in that basket? What's the big picture plan? Or is the new CEO just going rambo and pandering to DIY geeks who want everything for free ? (myself included!)

I feel like there has to be some parallel universe blog where these press releases explain what they're doing to investors beyond "we're going to take your money, and we're going to make things with it. And then give it away for free! Please give us more money"

Like awesome dude... but what the hell is going on?

View on HN

TL;DR: Azure.

What's going on in a new strategy.

Microsoft wants people building applications in Azure. They want people using ASP.NET technologies. Given that Azure competes with AWS and Google Compute; that ASP.NET and C# compete with Elixir/Phoenix, Clojure, Scala, Koitlin, etc, there is absolutely no way they can maintain any form of lockin. Even in the conservative companies, they compete with Java. The pitch to the investors is that by opening up their toolset, it gets more people on their cloud services and boosts recurring revenue. Which is what everyone else is doing.

> there is absolutely no way they can maintain any form of lockin.

They still have a very strong lock-in in graphics though. Why didn't MS join the Vulkan working group? And that's their attitude in general - they didn't become any better than before. They just drop lock-in under pressure by competition. Other than that, they pursue lock-in all the same.

Microsoft has changed, they now know what will bring the antitrust hammer down. They no longer sabotage or hamper competition actively like they have in the past. Instead they're trying to be everything to everyone and cater to everyone's needs.

Does that mean they're altruistic and want to save the world at their own expense? Definitely not. But they also recognize when when they're not going to win a fight and have started to compete in markets they can't own.

That being said if they own a market (e.g. PC Gaming) they aren't going to exchange money for warm fuzzies.

Vulkan is not "warm fuzzies". It's something that can make life easier for everyone. But it reduces their lock-in, that's why they ignore it and want to force developers to do double work if they want to target their closed platforms like Xbox.

To Microsoft it is. Vulkan and DX12 set out to achieve many of the same goals but DX12 has the added benefit of momentum and market lock-in. DX12 has been available since July where as support for Vulkan is spotty at best. Why drop DX12 for Vulkan besides warm fuzzies?

> Why drop DX12 for Vulkan

Because they can make life easier for developers and also benefit from the shared effort. But their whole goal is to make life harder for developers (i.e. those who don't want to be limited to MS walled gardens) as a way of taxing them for not falling in line. A usual crooked monopolistic behavior that MS is so accustomed to.

Prior to Vulkan I would argue that it would make little to no difference. OpenGL is different on every platform you look at and each hardware vendor has produced a slew of proprietary extensions to support features not found in OpenGL. What you end up with is render code written per platform even with OpenGL and on Linux you have different implementations per driver. At least with DX you're only writing 1 renderer for Xbox/Windows.

With Vulkan they finally dropped backwards compatibility, something Microsoft had been doing with each successive version of DX, so it would have been a good time for Microsoft to transition away from DX. But what motivation do they have at this point? OpenGL adoption isn't yet threatening their platform, and with Apple pushing Metal they have even less to worry about. Vulkan has only further fragmented the market because anyone that wishes to use it must also write fallback renderers because of the limited support.

In any case if/when Vulkan does become the industry standard, Microsoft will only need to make an Xbox One API because Windows already has support for Vulkan, better support than any other platform at that.

> But what motivation do they have at this point?

The motivation can be pretty simple - don't be a jerk. But it's MS. Lock-in is practically in their DNA already.

Waiting to see Vulkan on the console SDKs....

I hope it will happen. But I guess only when competition will be strong enough. I doubt MS will suddenly decide to play nice.

I was thinking about Nintendo and Sony...

I suspect they are in "wait and see" mode. Nintendo did join Vulkan group, unlike MS. They as well might start supporting it. They are smaller, and for them supporting standard technologies is an easy way to attract more developers. Sony also can do it, as a move for reducing MS influence.

Apple is also on the group, yet it is going with Metal. It is the slowest vendor to adopt OpenGL versions and it has a very bad OpenCL implementation, a standard that they created..

Being on the group means zero if the actions don't follow.

When are you going to properly start bashing Sony, Nintendo, Apple alongside MS?

There is some confusion there. Apple actually aren't in the Vulkan group if I understand correctly.

> When are you going to properly start bashing Sony, Nintendo, Apple alongside MS?

When their influence will be comparable in damage. Apple is probably closest, but not in graphics APIs. Metal is non existent. Apple causes more damage in browsers these days. Remember all the recent articles about Safary is the new IE?

Sony and Nintendo have much more influence in the consoles and game developers than MS.

Nintendo was the company that introduced the concepts of gated development on the aftermath of 1983`s crash.

Also why not bash NVidia given CUDA or their influential part in DX designs, to the point their cards are almost a mirror of the API?

Google so praised for Vulkan, pushes its own computing API instead of OpenCL.

Intel spends more resources in their DX drivers and their OpenGL support is still pretty shitty. One just needs to delve into the OpenGL developer forums. They were quite known in the games industry for having the driver lie about supported hardware features.

Or by announcing at GDCE 2009 their new graphic analysers tools, with DX only support. To my question what about OpenGL I was given the email of some random dude at Intel.

Yet, only MS gets the bashing.

How does Vulkan make life better for developers today? Isn't that more on hardware devs?

> How does Vulkan make life better for developers today?

Assuming Vulkan drivers are ironed out (now they are still fresh), it makes life easier by removing the need to implement different APIs for each platform. By sabotaging such option, MS taxes developers which want to target their Xbox for instance requiring them to use DX12 instead.

How has Microsoft sabotaged Vulkan? They haven't, Windows is one of the very few places that Vulkan will actually run. They offered a competing technology, DX12, that works today... and 6 months ago. Vulkan might work in 6 months... or not.

> How has Microsoft sabotaged Vulkan?

By not supporting it on Xbox which is a walled garden. Since Xbox has a significant market share, they essentially force developers not to use Vulkan (and use DX12) if they want to target Xbox. Compare it to browser wars of the past. MS did the same thing exactly, and the only reason they magically started supporting shared Web standards is because they lost those wars. Today it's a real wonder - they even decided to support free video codecs in the browser (while Apple still refuses).

I'm failing to see how that sabotages Vulkan. As far as I know, Vulkan has only recently reached 1.0, and nVidia JUST started supporting it. Until recently, there was nothing to support.

Your entire complaint is that developers have to do more work if they want to support multiple platforms, which has always been the case. I don't recall the original Final Fantasy getting released on the Sega Genesis. There's never going to be a magical catch-all graphics API because technologies like Metal and DirectX take advantage of known hardware specs and OS features to make the games look good and run fast.

> I'm failing to see how that sabotages Vulkan.

In order for cross platform solution to offer value of avoiding duplicate work, it should be really cross platform. MS prevents prevents it by not supporting it on Xbox.

> which has always been the case

Exactly because of such lock-in.

> technologies like Metal and DirectX take advantage of known hardware specs and OS features

Let's not try pretending that MS and Apple do it for any technical reasons. There are none.

I think you just blindly champion Vulkan, an unproven and not read for primetime technology.

The two largest platforms, MS and Apple have their own stacks, Metal and DX12. And they have been maintaining those from years, especially DX12 which comes from decades of DX development.

They should throw all this away (and lose any competitive advantage for their platform in the process) just because there's finally a competent new stack (that's not even ready yet)?

>Let's not try pretending that MS and Apple do it for any technical reasons. There are none.

Would wanting to offer the best possible experience for your platform, to control how and when any new feature is implemented according to your long term plans, and to not mess with lower-common-denominator cross-platform design-by-committee APIs counts as a "technical reason"?

Let's not pretend you ask for this for any technical reason. It's all about developer convenience at best.

> I think you just blindly champion Vulkan, an unproven and not read for primetime technology.

It's the only cross platform solution. So yes, it should be pushed forward to make it work. It's not a "blind" push - it's the only reasonable thing out there. Neither Metal nor DX12 plan to do anything comparable. Petty greed of Apple and MS prevents them.

>Petty greed of Apple and MS prevents them.

Wanting control of your platform and how it evolves is not "petty greed".

If anything it's the developers that want to save money, by only coding their apps once.

I don't believe platforms should have lowest common denominator standards (and evolve in lockstep) just to make porting easier.

Graphics APis should be common. Not making them common is about petty greed. Control of the platform here is done for the sole reason of making cross platform development harder. I.e. hindering competition.

You're asserting that DirectX was developed as a competing API to OpenGL and a lock-in mechanism to hinder competition which is a ridiculous assertion that completely ignores history.

OpenGL started development in 1991 by Silicon Graphics and it's primary focus was on CAD/CAM utilizing expensive hardware accelerators. It was initially only available on IRIX and was never intended for video games. It was not open and the Architectural Review Board that maintained the API was controlled by Silicon Graphics.

Direct3D was developed by a company called RenderMorphics starting in 1992 and was bought by Microsoft in 1994. Direct3D was added to the DirectX suite in 1996 specifically for video game acceleration. At the time Microsoft was shipping their own OpenGL Drivers and would continue to do so until Windows 2000 was released.

Microsoft and Silicon Graphics actually attempted to unify the two APIs under Project Fahrenheit in 1997. Project Fahrenheit consisted of low, medium and high end drivers with Microsoft providing the low-end and Silicon Graphics providing the mid and high end portions. Microsoft abandoned the effort in 1999 because DirectX adoption had taken off and they were focusing resources on the original Xbox console. Silicon Graphics was in financial trouble at the point and it no longer made sense to pursue a project heavily reliant on them.

Silicon Graphics trouble's were reflected in the stagnation of OpenGL development with only incremental releases up to version 1.5 in 2003. SGI would eventually go under and the API would get turned over the Khronos in 2005 but it wasn't until 3DLABS got involved that anything happened and it wasn't until 2007 that OpenGL 2.0 was released. At the point the Xbox 360 has been out for 2 years and Microsoft was heavily invested in DirectX.

What you are asking for may become a reality, but Vulkan is not something that any reasonable primetime product should have even considered trying to support until what, like... was it 2 months ago now or not yet?

The XBox is not a tech test platform. Neither are the mobile handsets.

Claiming this reality amounts to "sabotage" is somewhat absurd.

> Claiming this reality amounts to "sabotage" is somewhat absurd.

It's not. MS is not new to sabotaging portability. Learn from history.

I could use that exact logic to point to many other disingenuous working groups and say Vulkan is an attempt to create a new lock-in scheme under the guise of an open technology.

It wouldn't be the first time. "Learn from history."

How exactly is it lock-in if unlike MS it's cross platform and cross hardware? Just saying "it's the same logic" doesn't make it the same when it contradicts the facts.

Being cross platform and cross hardware isn't contradictory to lock-in.

Lock-in is about whether, after adopting technology X, it takes substantial effort to go to an alternative technology Y -- not about whether you can take X with you in a different platform/hardware.

>In order for cross platform solution to offer value of avoiding duplicate work, it should be really cross platform. MS prevents prevents it by not supporting it on Xbox.

Vulkan doesn't need to be cross platform because there are tools like Unity which abstract away the graphics API and let developers target a number of platforms.

However, you are lodging a complaint against Microsoft that has been the case for every console device since the first ones were released. However, that being said, not supporting Vulkan on the XBox harms Microsoft more than Vulkan since all that means is some games will not be available on the XBox. Just like there are XBox games that are not available on the PS4, the PC, or the Wii U. There's nothing Microsoft is doing that's out of line with the rest of the console makers. It's the console industry in general.

> Vulkan doesn't need to be cross platform because there are tools like Unity which abstract away the graphics API and let developers target a number of platforms.

Unity developers will be doing that double work in such case. And not everyone is using third party engines anyway. Don't think that if something is abstracted, it somehow magically happens for free.

> Unity developers will be doing that double work in such case.

I'm pretty sure they're well paid for the efforts.

> I'm pretty sure they're well paid for the efforts.

And that cost is passed along the way reaching those who use it including end users ;) Someone will feel the tax.

The same can be said for Apple, Metal, and iOS. I'm also certain most companies use a tool like Unity which does that double duty for them.

Agreed. Apple isn't any better in this aspect.

> I'm also certain most companies use a tool like Unity which does that double duty for them.

That doesn't justify the lock-in tactics of MS, Apple and the like. If that burden shifts to engine developers, it's still a burden and effort that could be spent on something more useful like improving actual features of the engine. In the end, it's simply a tax that everybody will be paying.

Mind share and Market share are the justifications. Once you have a developer comfortable with a platform, it can be hard to impossible to get them to switch.

These are businesses with an interest in making money, not people with an interest in making the world a better place. I'm not saying it's right, just that it makes business sense.

> Mind share and Market share are the justifications.

Rather herding their market share and mind share, and making it hard to go outside the herd. I.e. if you want to be cross platform - pay the tax. That's the whole purpose of their lock-in.

> Once you have a developer comfortable with a platform, it can be hard to impossible to get them to switch.

Yes, and it's disgusting to use development tools for lock-in. MS do it like forever.[1]

> These are businesses with an interest in making money

Lock-in is a crooked method of making money.

[1]: https://en.wikipedia.org/wiki/Criticism_of_Microsoft#Vendor_...

Right because that isn't the case with every other vendor or platform out there?

It isn't. Not everyone is a jerk.

Name one major technology company that doesn't try to lock you into their recurring subscription platforms.

Anything that uses interoperable technology. Lock-in implies something that you can't replace for no valid reason. Standards are aimed at such interoperaibilty and luckily we have a lot of standard technologies which are interchangeable. It would have been a nightmare, if literally everything would have been locked-in MS style.

I.e. imagine you bought a light bulb and can't use it because it's XYZ lightbulb that can't work with fixture ABC and requires you to rewire the whole house. What a wonderful MS style experience. Luckily it's not the case most of the time.

> Anything that uses interoperable technology

Such as? Sure you can replace your iPhone with an Android device, but it's not pretty.

Better ask Apple why never could use standard chargers and USB cables ;)

https://www.khronos.org/members/

Really? Might want to check that again.

Although, I am curious, considering Apple does the same thing with Metal that Microsoft does with DirectX. The difference is Apple has no dedicated console device they sell.

> Apple has no dedicated console device they sell.

ahem ... http://www.apple.com/tv/ :P

Khronos in general - yes. For example for WebGL (competition and the fact that they lost browser wars caused that). But Vulkan? MS clearly avoided any participation. So you only confirm what I said - MS drop lock-in only when competition starts burning their heels.

Yes, "cloud" is changing everything, for better or worse. Microsoft realized that enterprise, their primary market, LOVES cloud and that Windows technologies were much less capable in the cloud than open technologies for a variety of reasons, not the least of which is licensing.

MS is opening their platform to run on Linux because they want people to continue using Windows, Visual Studio, and .NET as development platforms. The core of MS's business has always been Windows, and secondarily, Office. Since so many people want to use the cloud, and on top of that since so many people want to use the cloud via sexy vendors like AWS, Microsoft needs the heart of their Windows lock-in strategy, which is .NET, to work on it.

These moves are defensive of .NET's place in enterprise development, and consequently Windows's place on the enterprise workstation. Once MS feels that .NET is far and away the clear choice for development in the cloud-centric world, they'll work to bring more people onto Windows as the cloud OS of choice (Azure, etc).

https://github.com/Microsoft/dotnet How exactly is releasing the .NET framework on Github a lock-in strategy?

Like many people, you are missing the point. The operating system is irrelevant. You have people editing Excel spreadsheets in an iPad on iOS, and then switching to Excel on some form of Windows. They've been trying to get Windows Server and IIS as the platform of choice, and they have lost to Linux, Apache, and NGINX.

From a technological perspective, they can't afford lock-ins, not when there are so many alternatives. They want you using their services.

>How exactly is releasing the .NET framework on Github a lock-in strategy?

In the way described in my original post. Open-sourcing the .NET framework means people will help them port it to new platforms. It means that people will be able to get the .NET framework running on whatever is running in their corporate cloud. This means that users will continue to develop in C#, using APIs developed by Microsoft that will always work best on Windows, using development tools provided by Microsoft that will also always work best on Windows.

If your company switches from .NET to Java because .NET doesn't work well "in the cloud", then your company starts to ask other questions, like "Well, we don't use Visual Studio anymore, so we don't need to keep paying Microsoft $xx,xxx per year for that. Everyone likes Mac a lot and we don't have to use MS's dev tools anymore, so why don't we move everyone to that? We can start to phase out our Windows Server and SQL Server licenses too since we're not using any other Microsoft stuff and then we won't be paying them another $yy,yyy and we'll be one of the coolest companies around because everyone will be using a Mac."

MS's strategy of opening the foundations of their platform so that the applications can run on any backend makes it so you instead say "Oh, we still run .NET apps on our AWS cloud, so we still need Visual Studio, better keep paying MS $xx,xxx." And "oh, SQL Server runs in our cloud infrastructure, so we can stick with that; that's another $yy,yyy for the license this year."

If you stay on MS's platform, they can then call you up at any time and say "Hey, switch to Azure, it's a lot cheaper than AWS because we don't charge ourselves license fees." On the other hand, if you move off their platform because MS tech doesn't work well in the cloud, they lose a lot of existing revenue, a lot of potential revenue, and a lot of mindshare and exposure that decreases their relevance in general.

The key to MS's lasting success has been .NET. That's what's been propping up Windows and consequently Office. If they lose .NET, they're in deep trouble. They are going to do everything they can to keep it alive, even if it means making it so your ASP.NET web app runs on a Linux server.

I can't get over the fact that Microsoft's cloud product is the color of sky without clouds in it.

So they've decide that the business model of the past two decades is doomed? Are investors onboard with that? Sounds like some serious downsizing is coming to come up in the next 5 years...

We might be looking back at this CEO with incredulity in 10 years when MSFT is struggling to keep the lights on ("they did what?! They just gave away all their products for free??")

Have you paid attention to the change in behaviors that have come about due to mobile devices? Web services are huge in helping to store user data and keep it in sync across the diverse set of computer people use.

Right now, I can't think of any vendor that forces a lock into their platform for development purposes. Even Apple made Swift free and cross platform. Lock in on Visual Studio and C# makes no sense. C# is a nice language, but there's a wide variety of alternatives that are just as good. Visual Studio is really nice, but people have been building stuff with Eclipse, Sublime, and Emacs for quite some time.

They are making a long-term push to subscription oriented services.

> So they've decide that the business model of the past two decades is doomed?

Windows is a commodity, a) it won't pay for its own engineering team and b) unlike other competitors, Microsoft hasn't been able to capture as much value from the current Windows ecosystem and marketplace.

Developing for Windows is becoming a liability: a) your team can't develop only for Windows anyways; b) your team can be more efficient by using a wider set of tools and technologies than Microsoft's own developers, and c) if Windows is not profitable and your product depends on Windows, it better pay for part of Windows' cost.

Internal OSS adoption, investing on cross-platform development and in Azure seem all strategic investments in managing the risks above while remaining a software (product) development company.

Every other gain from the same strategies (the ones affecting their customer base) are cherries on the top of the cake.

> Are investors onboard with that?

Not sure they realize what's going on. If they did, I couldn't explain the current P/E ratio in days of declining revenue, a prospect of shrinking revenue streams (overall; medium-term at least) and large bets on lower margin businesses.

> They just gave away all their products for free?

They probably wish they had invested in Office for the Web, Android and iOS much earlier on and with much more intensity, and still have given it for free. They were lucky none of the competitors really stepped up to fill the void. In a cloud-first and mobile-first world, there are multiple revenue stream available once the data information workers process is more closely tied to a provider; specially with the synergies in their product line.

They used to live of windows licenses but their market share is shrinking if you take tablets and mobiles into account. The new strategy seems to be making money on cloud services, which are platform agnostic, and search engine, of which Windows is becoming just a natural extension, like Android and Chrome OS are an extension to google.com, a probe given for free that Google wants to see in each of their users pockets to keep an eye on where they are, who they talk to, what they do. Hence the extensive tracking of Windows 10, and Windows 10 upgrades being given for free.

We may like it or not, or agree that it makes sense or not, but at least it looks like a rational strategy.

The alternative is to wait while change passes them buy and then issue a Nokia "burning platform" memo.

(Nokia are now a wholly owned subsidiary of Microsoft. If Microsoft collapsed, who would buy them? Apple?)

Microsoft bought Nokia's mobile phones but the rest of Nokia survives (marker cap $22 billion).

> So they've decide that the business model of the past two decades is doomed?

Isn't it? How much growth do you see around the PC and Office business?

Is growth really necessary? They have cornered the corporate desktop pretty well, businesses aren't going to stop buying new computers for all employees every 3 years with Windows and office and the associated backends with SharePoint/Exchange etc.

> Is growth really necessary?

I actually agree with you. I think Microsoft's evolution is mostly done. The XBox is their consumer device and everything else is all about enterprise and small business users.

Microsoft is a boring, mature, very profitable company. It isn't such a bad thing to be.

Their hardware is in an interesting place though. They are really the only company making decent Windows hardware (and I say that as a longtime ThinkPad fan). I would be so happy if they took the Surface Book and simplified it into a traditional laptop with battery life as good as or better than similar Apple machines.

Sun 2.0?

I think Microsoft is recognizing that locking people into an ecosystem is not a long term solution when there are near equivalent ecosystems available for free. It almost seems like they're positioning themselves to get mindshare ('Hey, this free product from the Evil Empire isn't that bad! Maybe they're just another empire, and not so evil.') and transition that into a paid service (Azure is what immediately springs to mind, but Office is pretty essential when you get past a certain point (I say this as a long time OpenOffice/LibreOffice user - there are some things MS just gets right).

Or maybe Microsoft is just creating a bigger and better ecosystem....

Who else is trying to cover all your needs (1) both on-premise and in the cloud; (2) across platforms from USB thumbdrives to smartphones, tablets, laptops, desktops and servers; and (3) has cross-platform apps on Windows, Android and iOS?

Google is pretty much cloud-only and Apple is pretty much Apple only, and neither of them supports anything like the same range of devices.

It's a long play, but Microsoft has been executing this strategy for a few years now....

> Maybe they're just another empire, and not so evil.

They are just another empire. Enter the new world! This is the smartphone and cloud era!

> Like awesome dude... but what the hell is going on?

Survive now, figure out how to make money later.

If Microsoft does nothing, very little new development will be done to run on Windows. You don't need Windows to access Gmail of Facebook. Their backends don't run on Windows. Why would you deploy your software to Windows if it's cheaper and more effective to deploy it to Linux? Shortly, all their technologies will become irrelevant.

It's true they are doing interesting stuff in the JavaScript space, but what does it do to improve their bottom line? It's not unlike Sun with their Java ecosystem, but from an even worse position than Sun held.

I think it's about capturing new marketshare or previously lost marketshare. There's a large group of developers and their associated companies that were never likely to hop on the MS toolchain. In many ways MS == .NET for years. Now, it's more agnostic with Azure and even Visual Studio Code.

So, marketshare and marketing.

> Microsoft's business model was to bring people in to their ecosystem and then charge them

Yes. But for about the last 10 years, that ecosystem hasn't been doing very well. Most Windows hardware is garbage and (except for games) there's virtually no excitement around software. What's the last Windows application that went viral?

Their developer strategy (universal Windows) is interesting and if I were in charge I would probably go down a similar path, but as a developer it doesn't interest me at all. Applications that can run on a phone or a desktop is an interesting trick, but in reality I normally want a substantially different UI on each. That would lead me to think about separating the UI from the application and then I might as well make a web app.

Popcorn Time, I believe

That's a very good answer.

Reply on news.ycombinator.com