But due to an increased focus on production values over character development and story, the king beyond the wall appears only for a short time during a lunch break to dance wildly upon the tables and promptly runs away.
Two perspectives on the recent grand entry of Microsoft into the Linux space:
1. Linux won in the embedded (ARM devices), mobile (Android) and server spaces, and this is Microsoft bowing to that.
2. This is Microsoft's embrace, extend and extinguish strategy from the 1990s in a new wrapper. They just had to do it to an operating system this time.
Not sure if these perspectives are correct, but I'm just throwing them out there as food for thought.
This has nothing to do with the article. This is about openshift, not linux (albiet, openshift runs on linux).
They seem to be embracing openshift as a prominent k8s distro on their cloud, which I think is interesting. Instead of attempting to create their own EKS type service, they're providing an openshift platform.
I'm sure there are no other deals between MSFT, RedHat and IBM. The timing to pressure Linus into "approving" the CoC (MSFT is Platinum sponsor of the Linux Foundation) also fits well into the embrace/destruct-or-take-over strategy.
I see this move as Microsoft not being convinced that backing k8s is the best strategy at the moment. Also, Openshift and Red Hat are "closer" to enterprise IT and therefore a good chance to capture "containerization" opportunities in large IT orgs.
It would seem to be bowing to what their customer base wanted (whether their customers realized it or not). Most of my personal/work devices are Windows and prior to the big change, on each one I'd have a virtual machine with a linux distro in it so that I could do some work that was outside of ASP.NET (which I'd do in Windows).
Now, ASP.NET can be developed in OSX/Linux/Windows and they've cut out layers involved so that developers can work within Linux on their Windows machine.
Is the goal to upsell you on Azure? Most likely. Nothing wrong with that though: every company that intends to make a profit wants you to use the thing that gives them said profit. The fact you can use the two previously mentioned things without any requirement to use Azure is probably what they should be measured on in regards to those two things. On the bright side, Azure is a pretty diverse setup whereas it could've been Windows Server w/ IIS and Windows Server w/ SQL Server instances as far as the eye can see.
Another possible perspective is that Apple has more or less given a passing interest to OSX in recent years and Microsoft felt they could make inroads with developers using Mac products (that weren't there for iOS development. ie web). It's worked to some degree (anecdotally. Though the substantial enhancements to the OS on the Windows side have been at a higher clip than on the OSX side for a bit though) and been a driver for a bit of good faith with development audiences so they probably view it as a win-win. But they're not doing a great job of extinguishing (if that were indeed the goal) because Visual Studio and Visual Studio Code have both found their way onto OSX (VSCode on Linux too). Would seem nutty to bolster OSX's appeal if this were the same old Microsoft just under new management.
> This is Microsoft's embrace, extend and extinguish strategy from the 1990s in a new wrapper
We're all right to be cautious about trusting large companies' motives when they embrace an open standard. But I don't think there's any reason to invoke EEE! until they have introduced a proprietary (i.e. closed-source) implementation of something and then added non-standard extensions.
Since much of Microsoft's recent activity has been released as open source (MIT licensed, no less), I have a hard time believing Microsoft would go down that road again any time soon.
Rather, I think they're building their business on providing the best and easiest-to-use hosting and cloud services provided you're using their excellent and free developer tooling. And I'm OK with that, since they're not doing anything that is unfairly competitive. (Even AWS could start providing VSCode extensions and they'd be on equal footing with Microsoft, though to win they'd have to compete with MS's cloud directly.)
I really agree with your point around "best and easiest-to-use". I like this even more than how Google became defacto choice by being the "better" search engine (I put it in quotes because its pretty fuzzy now if it is better, but when it started it was certainly cleaner and probably better).
I like when a company wins by just being superior (and hopefully in a way where they aren't deliberately building a moat)
If they go defunct or if the C# debugger somehow becomes unavailable, that’s a strong incentive for someone else to build one, given C#’s market reach. Perhaps that person would even be kind enough to open-source it.
Either way, because C#/.NET and Roslyn are open-source, the likelihood of .NET shops being left high-and-dry without a debugger is very slim, since it wouldn’t be all that hard for an experienced developer to build another one.
> But I don't think there's any reason to invoke EEE! until they have introduced a proprietary (i.e. closed-source) implementation of something and then added non-standard extensions.
That would be WSL wouldn't it? Their goal is clearly to avoid having people moving to Linux as a development platform, and keep them on Windows.
The extensions are the ability for Linux processes to integrate with Windows software running on the same machine. In a way, they're extending Linux with Windows services.
> Their goal is clearly to avoid having people moving to Linux as a development platform, and keep them on Windows.
I think it's the opposite. There's little danger that Windows shops are going to move to Linux. However, I know a bunch of devs currently using Macs that would probably switch to Windows with nice devices like the Surface Pro and access to real gaming and such, if only Windows was a suitable Unix environment for their work.
They want devs using Macs and such to move to Windows, and I think it's going to work.
The only implementation built by Microsoft of any standard in WSL is the bridge that provides a Linux kernel interface (it proxies Linux syscalls to the NT kernel). But their only goal there is to support Linux userspace, so it would be rather pointless to add extra syscalls or extend the existing ones. (And besides, they haven't even finished implementing all of the system calls...)
For WSL 2, they're taking the Linux kernel whosesale and dropping it into a VM in Hyper-V. If anything, they're backing away from anything that could be construed as a possible EEE.
I'm hearing this sentiment more and more frequently.
I, however, don't see him doing anything to that "broken model." Since Ballmer left, Microsoft actually got even more dependent on plumpy "corporate" clients. It just people who once self hosted AD now got upsold on running their AD and other corporate software on Azure.
Can't criticize Nadella here. You will be able to do so in 5 years when something evil happens. Then HN will be trying to determine how this could possibly have happened.
Your claim about MS not focusing on their retail side might be accurate, but that in and of itself doesn't really mean anything. The stock market (i know .. i know.. ) seems to have accepted MS's "broken" model, and its stock price has tripled (or something like that) under Nadella. Certainly, its not a company that people seem to think is broken in any mainstream way of thinking about a business.
> The stock market (i know .. i know.. ) seems to have accepted MS's "broken" model, and its stock price has tripled (or something like that) under Nadella
Well, as it did under Ballmer, under same circumstances. He also did pretty much the same — doubling down on milking plumpy corporate clients while streamlining the business in overall + doing some disasterous acquisitions which were nevertheless welcomed by Wall street analysts
Well, I don't know what you are referring to as "it", but the stock price movement wasn't the same under Ballmer...which might be a bit unfair to him since he came in after
the .com boom. https://imgur.com/a/46fnc4P
>He also did pretty much the same — doubling down on milking plumpy corporate clients while streamlining the business in overall
This is what I am talking about ...
Developers, developers, developers : Ballmer was right about that but he though he should only care about .Net and MSFT products and fight everything else.
The Nadella model is we don't care what you use, where or how : if you're making software we want to help you ... with this buying services from MSFT is not the goal per se but a consequence
I think Balmer lost vision of what a "platform" truly is. Microsoft understood this in the 90's with Windows, forgot it in the 00's under Balmer and seems to be recapturing it heading into the 2020's. They could be a huge force again by the end of next decade.
When I discovered VS Code a few months ago (I'm a hobbyist just getting into coding, not a professional) after many recommendations from friends who code and are die-hard open source true-believers, I did a legitimate doubletake after reading the name of the company that developed it.
I then spent an hour researching it, trying to find out what the catch was. Was the wikipedia page entry a prank? Was I getting some kind of malware masquerading as a pro development tool? Was is a temporary trial before big licensing fees were demanded of me? Does it need some sort expensive companion server that only runs on Windows Server? What's the catch?
But no, no catch. It's free, and it's open, and it works great. But every time I launch it I feel like I've walked into some weird alternate dimension. I have to remind myself that yes, this is reality: I'm using a native-Linux, fast, reliable, extensible, MIT-licensed editor made by freakin' Microsoft. So strange, man. So strange.
... except perhaps for the Microsoft telemetry that is active in the distributed VScode binaries [1]?
To quote user ilaksh in this thread: "They will set it up so that it will _seem_ like your are developing for Linux, but due to some extensions or technicality it will only work on _Linux inside of Windows_.
So from my perspective it is amazing that people think this is a good thing for Linux."
I did see and disable the telemetry option in the preferences the first time I used it, and so far it seems to behave itself and not quietly re-enable telemetry.
If you would have told me 10 years ago that by 2019 Microsoft would have first-party support in Visual Studio for cross-platform development on Mac and Linux, created a Linux subsystem so users could run Linux natively on Windows, and their CEO would be in attendance at a Red Hat conference, I would have said you are crazy. Interesting times, indeed.
And it works. I find myself pitching Azure to my web dev clients, partly because I can deploy my C# project directly to an Azure Web App with a single button inside Visual Studio.
Didn't some documents leak showing a Microsoft Evangelist killing Apple centric conferences?
I know Microsoft tried to kill Linuxfest Northwest a few years back, sponsoring it and taking control of 1/3 of the floor space, both approved afterpartys, etc. This year they had a small booth that was staffed for one of the two days...
Most of the people I know stopped going to LFNW after the shitshow that was that year, they still attend SeaGL and other conventions.
The exploding can of WD-40 that MS handed out that year was the cherry on top of the hijacking of the main space, fake Azure credits, and poorly run afterparty that MS put on.
Frankly, MS should not have been allowed to hijack LFNW that year. This year MS was the only booth to not staff on Sunday, I'd hope LFNW doesn't waste table space next year.
What I wonder is, does anyone have OpenShift experience here? If so, what do you think of it? And compared to Kubernetes? I know OpenShift is Kubernetes plus extra features, but that doesn't necessarily mean it's better.
Openshift is a fork of Kubernetes...with a few more things bolted on. I think the distinction is important. Development occurs in a separate repo and the upstream Kubernetes code is backported into that repo. Similar to how Red Hat handles RHEL.
I know. But Kubernetes doesn't require 16 GB of ram on the master node to run, for example. There's a ton of differences between Kubernetes and OpenShift, hence why I'm asking. The fact that OpenShift is a superset of Kubernetes precisely means there's differences.
We have been using it for a few years. Its a complete solution for deploying apps (with logging, monitoring, rolling deployments, etc) and it has a pretty nice UI. Once you have it set up its as easy to use as Heroku. Can't really imagine a better solution for medium sized projects.
I tried openshift 3 when it was new and received apologies from RH employees on IRC because it was so buggy.
It felt like one guy saw kubernetes and decided to make a GUI for it, hence openshift. Felt like it was a true open source project, not bad, but not production ready for end users yet. This was early 2018.
Work with OpenShift a lot and I really like it. Very similar to Kubernetes but solves a lot of business problems including providing:
* security
* source to image builds
* a well designed ui for monitoring and configuration
* advanced routing techniques such as a/b deployments ect
* liveness probes, readiness probes and other needs
* metrics via a prometheus plugin
Azure has become my go-to place to deploy nodejs, I used to use AWS exclusively but Aws is brutally complex while Nadella's team have done an amazing job with the clarity and usability of the azure UI.
It looks like Google is loosing hacker confidence while Microsoft is gaining them.
Am not able to keep count of the number of times I have been burnt because of Google's decision to drop support for their products, change functionality after buying then and inane UI changes.
Contrast that to experience with Microsoft, I can't recall that many instances.
I consider myself to be roughly in the category of "hacker" in the context of that sentence. I don't have confidence in The Google controlling everything any more than I do with MS.
I used to be an MS developer many years ago. I will say one positive thing -- MS makes a lot of very useful development tools.
However, I'm amazed that people think it's not MS anymore. They have not gained my confidence in any way in terms of this supposed altruism or whatever.
Linux in some ways has been one of very few possible Windows competitors on the desktop. The fact they they essentially decided to just swallow Linux into Windows does not make me think they are more friendly to Linux.
They will set it up so that it will _seem_ like your are developing for Linux, but due to some extensions or technicality it will only work on _Linux inside of Windows_.
So from my perspective it is amazing that people think this is a good thing for Linux.
I doubt this will happen. For one thing, the dominant targets for Linux are servers and embedded systems - both of which are cases where resource overhead is pretty important. Running a Windows to run Linux inside of incurs a pretty substantial overhead, so I really can't see any case where developers would be willing (or even necessarily able) to spare the resources to install Windows on their formerly-Linux-only targets just to maintain the convenience of getting to develop for "Linux" on their Windows PCs.
And desktop apps developed for WSL will simply never be Linux-compatible out of the box at this rate, given that WSL does not contain any graphical or UI functions; MSFT has been pretty clear about this right from the start, and so I really can't see a scenario where people mistakenly believe just developing user-facing applications for WSL will make them "just work" on Linux, given that all the UI code is on the Windows side.
72 comments
[ 3.2 ms ] story [ 125 ms ] threadI mean, if you're against cloud hosting that's fine, but what is wrong with Azure specifically?
1. Linux won in the embedded (ARM devices), mobile (Android) and server spaces, and this is Microsoft bowing to that.
2. This is Microsoft's embrace, extend and extinguish strategy from the 1990s in a new wrapper. They just had to do it to an operating system this time.
Not sure if these perspectives are correct, but I'm just throwing them out there as food for thought.
They seem to be embracing openshift as a prominent k8s distro on their cloud, which I think is interesting. Instead of attempting to create their own EKS type service, they're providing an openshift platform.
Now, ASP.NET can be developed in OSX/Linux/Windows and they've cut out layers involved so that developers can work within Linux on their Windows machine.
Is the goal to upsell you on Azure? Most likely. Nothing wrong with that though: every company that intends to make a profit wants you to use the thing that gives them said profit. The fact you can use the two previously mentioned things without any requirement to use Azure is probably what they should be measured on in regards to those two things. On the bright side, Azure is a pretty diverse setup whereas it could've been Windows Server w/ IIS and Windows Server w/ SQL Server instances as far as the eye can see.
Another possible perspective is that Apple has more or less given a passing interest to OSX in recent years and Microsoft felt they could make inroads with developers using Mac products (that weren't there for iOS development. ie web). It's worked to some degree (anecdotally. Though the substantial enhancements to the OS on the Windows side have been at a higher clip than on the OSX side for a bit though) and been a driver for a bit of good faith with development audiences so they probably view it as a win-win. But they're not doing a great job of extinguishing (if that were indeed the goal) because Visual Studio and Visual Studio Code have both found their way onto OSX (VSCode on Linux too). Would seem nutty to bolster OSX's appeal if this were the same old Microsoft just under new management.
We're all right to be cautious about trusting large companies' motives when they embrace an open standard. But I don't think there's any reason to invoke EEE! until they have introduced a proprietary (i.e. closed-source) implementation of something and then added non-standard extensions.
Since much of Microsoft's recent activity has been released as open source (MIT licensed, no less), I have a hard time believing Microsoft would go down that road again any time soon.
Rather, I think they're building their business on providing the best and easiest-to-use hosting and cloud services provided you're using their excellent and free developer tooling. And I'm OK with that, since they're not doing anything that is unfairly competitive. (Even AWS could start providing VSCode extensions and they'd be on equal footing with Microsoft, though to win they'd have to compete with MS's cloud directly.)
I like when a company wins by just being superior (and hopefully in a way where they aren't deliberately building a moat)
Either way, because C#/.NET and Roslyn are open-source, the likelihood of .NET shops being left high-and-dry without a debugger is very slim, since it wouldn’t be all that hard for an experienced developer to build another one.
What good? All that open source code.
The constant goalpost moving here is ridiculous. “Oh, Microsoft released n projects as Free Software? Well by god they should have released n+1.”
That would be WSL wouldn't it? Their goal is clearly to avoid having people moving to Linux as a development platform, and keep them on Windows.
The extensions are the ability for Linux processes to integrate with Windows software running on the same machine. In a way, they're extending Linux with Windows services.
I think it's the opposite. There's little danger that Windows shops are going to move to Linux. However, I know a bunch of devs currently using Macs that would probably switch to Windows with nice devices like the Surface Pro and access to real gaming and such, if only Windows was a suitable Unix environment for their work.
They want devs using Macs and such to move to Windows, and I think it's going to work.
For WSL 2, they're taking the Linux kernel whosesale and dropping it into a VM in Hyper-V. If anything, they're backing away from anything that could be construed as a possible EEE.
Nope it seems their target is not linux but the whole dev(millenials)community
I, however, don't see him doing anything to that "broken model." Since Ballmer left, Microsoft actually got even more dependent on plumpy "corporate" clients. It just people who once self hosted AD now got upsold on running their AD and other corporate software on Azure.
Well, as it did under Ballmer, under same circumstances. He also did pretty much the same — doubling down on milking plumpy corporate clients while streamlining the business in overall + doing some disasterous acquisitions which were nevertheless welcomed by Wall street analysts
>He also did pretty much the same — doubling down on milking plumpy corporate clients while streamlining the business in overall
I take it you think that this is a bad thing?
MS thank you for VS Code - I never thought I would replace my Vim.
I also love how Azure products work so seamlessly with it. Almost magical.
I then spent an hour researching it, trying to find out what the catch was. Was the wikipedia page entry a prank? Was I getting some kind of malware masquerading as a pro development tool? Was is a temporary trial before big licensing fees were demanded of me? Does it need some sort expensive companion server that only runs on Windows Server? What's the catch?
But no, no catch. It's free, and it's open, and it works great. But every time I launch it I feel like I've walked into some weird alternate dimension. I have to remind myself that yes, this is reality: I'm using a native-Linux, fast, reliable, extensible, MIT-licensed editor made by freakin' Microsoft. So strange, man. So strange.
To quote user ilaksh in this thread: "They will set it up so that it will _seem_ like your are developing for Linux, but due to some extensions or technicality it will only work on _Linux inside of Windows_.
So from my perspective it is amazing that people think this is a good thing for Linux."
There are good historic reasons to distrust M$.
[1] https://news.ycombinator.com/item?id=18209082
I know Microsoft tried to kill Linuxfest Northwest a few years back, sponsoring it and taking control of 1/3 of the floor space, both approved afterpartys, etc. This year they had a small booth that was staffed for one of the two days...
Most of the people I know stopped going to LFNW after the shitshow that was that year, they still attend SeaGL and other conventions.
The exploding can of WD-40 that MS handed out that year was the cherry on top of the hijacking of the main space, fake Azure credits, and poorly run afterparty that MS put on.
Frankly, MS should not have been allowed to hijack LFNW that year. This year MS was the only booth to not staff on Sunday, I'd hope LFNW doesn't waste table space next year.
It felt like one guy saw kubernetes and decided to make a GUI for it, hence openshift. Felt like it was a true open source project, not bad, but not production ready for end users yet. This was early 2018.
I used to be an MS developer many years ago. I will say one positive thing -- MS makes a lot of very useful development tools.
However, I'm amazed that people think it's not MS anymore. They have not gained my confidence in any way in terms of this supposed altruism or whatever.
Linux in some ways has been one of very few possible Windows competitors on the desktop. The fact they they essentially decided to just swallow Linux into Windows does not make me think they are more friendly to Linux.
They will set it up so that it will _seem_ like your are developing for Linux, but due to some extensions or technicality it will only work on _Linux inside of Windows_.
So from my perspective it is amazing that people think this is a good thing for Linux.
And desktop apps developed for WSL will simply never be Linux-compatible out of the box at this rate, given that WSL does not contain any graphical or UI functions; MSFT has been pretty clear about this right from the start, and so I really can't see a scenario where people mistakenly believe just developing user-facing applications for WSL will make them "just work" on Linux, given that all the UI code is on the Windows side.