Microsoft Open Sources C# Compiler
roslyn.codeplex.com
roslyn.codeplex.com
Folks, this is a very big deal for Microsoft. Who would have imagined this 10 years ago?
Here is an image that shows what they are putting into the community https://pbs.twimg.com/media/BkT9oBcCQAAHIAV.jpg:large
Visual Studio and their MVC framework continue to impress me (I mostly work in Ruby today, but have to maintain some .NET stuff) and MS takes inspiration from successful open-source tools and frameworks where needed. Many of their recent advancements over the last 3-4 years have clearly been the work of engineers who enjoy and value the best of the open source world.
I'm not trying to tell you that Microsoft is suddenly our warm, fuzzy friend. As even Scott Hanselman would gladly admit, they're ultimately trying to sell you software licenses or Azure services.
But it seems to me that they're using open source the right way in order to achieve that goal, as opposed to the bad old "embrace and extend / embrace and extinguish" days at Microsoft.
Whether or not Microsoft has earned another shot with us or not is definitely up for debate, and with all the cool shit happening in the open source world it's probably even a moot point to an extent.
But they're definitely doing more than cargo-culting in Redmond these days.
But I do indeed think this is better than not having happened.
I wouldn't place bets on any strategy that relied on Microsoft not closing the doors again at some point in the future, on future versions of their development stack.
Like you said, though, this is better than not having happened.
I know a lot of people who went to work for MS after the .com crash in the 2000's. All of them went in with the attitude that MS needed to get religion on standards and open source. Seems to me that some of this is paying dividends now.
They have a bad track-record, but not in terms of "this team" (known as Devdiv). They might have gotten all uppity about IE, Windows or what-have-you - but the .Net team had a relatively quick epiphany about open source. Look at Mono.
Java is always touted as a "more open" platform, Microsoft-haters have no issue spewing falsity that the Microsoft encumbrance is a big risk - however, so far as the framework track-record goes, Microsoft has a "community promise", an ECMA spec and has given Mono unit tests for Moonlight; Sun/Oracle has a track record of aggressively pursuing copyright (the recent spat with Google/AOSP).
I'm not saying that Java sucks, or that people shouldn't use it. I'm just saying that a lot of people need to have a hard look at how much time they spend trying to mud the name of a company trying to make a turn. These opinions are antique.
It may change in the future given new governance in those corporates (just as the e.g. Java, NodeJS situation may change), but as of now and as far as .Net FX goes: Microsoft is likely one of the best living examples of how things should be done.
Any examples? Besides "bundling IE with the OS and pressuring some OEM vendors back in the nineties"?
There are a lot of people at Microsoft who really do understand the value of open source. When I worked there ten years ago, there were plenty who did. When I wrote interop papers as an outside consultant in 2007, I worked with a unit that did.
But Microsoft also has a bunch of real problems that make it hard for them to realize this value, and so their options in going into it are somewhat limited. They can't pivot and become a services company because their whole Microsoft Partner network would be turned upside down. So they enter services in some key areas where their partners can't. It's a very delicate process, and it means for many on the administration side, the value of open source software for Microsoft is not as high as it would be for some other firms.
Nonetheless this is huge. Open sourcing the VB.net/C# compiler effectively means the possibility of an Linux port, which means that .net has the possibility to become write once, run anywhere even if the whole class libraries are not open sourced yet. This is a gamble for Microsoft because it reduces the barriers to moving from Windows to other platforms (Linux, OSX, BSD, etc).
I'm hoping that the tooling and integration gets a lot better in *nix, but nod sure how the Mono license (MIT iirc) and this new C# license (Apache) will jive together.
Personally, most of my new dev for the past couple years has been in node.js (which has its' own niceties and quirks).
They renamed it from Windows Azure to Microsoft Azure to remove the windows connotaiton - specifically mentioning that 15% of all Azure deployments are Linux OS. So I would say that this is where they see the future, although there are still some very byzantine sales processes to go through to buy enterprise azure, because of the existence of the partner network. In that sense, they're almost like the car manufacturers forcing you to buy through a dealer for their direct-to-consumer products.
I don't like Microsoft but Visual Studio is the only IDE I've ever tried that is substantially better than gedit/jedit. The auto-completion (intellisense) is just delightful to use.
Which is exactly what you can now create using the Roslyn technology that is now open sourced. Add these languages and they will get all the fancy support that Visual Studio provides, like (remote) debugging and code completion.
I mostly feel sympathetic to the engineers working there. It's no fun living under an embarassing dictator.
I once saw a presentation of the Visual Studio lead in Belgium, where he claimed that there were over 1200 people involved in the Visual Studio project.
If that's true, i don't think there is any IDE that can top that number and it explains why Visual Studio is the best tool for (almost any) job.
People should realize that Visual Studio is not solely C#, it's NodeJS, PHP, Mono, Python and so much more in one environment.
Non-copyleft licenses can easily align with the "embrace, extend, extinguish" strategy. The value of .NET for most developers is closely associated with Visual Studio. They can make the next version closed source and have Visual Studio only accept that one and we are back to the same old Microsoft. Sure, you'd still have the previous open source version, but who cares when most people will probably just move on to the next.
What? I would think if anything giving everyone else a crippled license while retaining copyright etc yourself that would be an even bigger setup for a EEE.
Remember, MS can do w h a t e v e r they want with that code anyway it is not lke they'd be bound by a copyleft on their own code.
(Unless they also started accepting patches without copyright assignment I guess.)
[0] https://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish
[1] http://www.apple.com/pr/library/2000/04/05Apple-Releases-Dar... [2] http://opensource.apple.com/
Apple built on oss foundations, and most of what they've open sourced is simply a result of them not closing forks of once open projects, which is very valuable, but nevertheless less impressive that releasing important bits of their own making.
How do you reckon that?
The fact that the core OS is still open source doesn't actually mean much when there's a huge layer of proprietary code on top, does it?
And the iPad, which is kicking the hell out of Android tablets (to my consternation, because I'm primarily an Android user and their tablets are terrible).
And the Macbook Pro, which is making pretty major inroads at leading-edge companies. A buddy of mine is working on *nix software at IBM and they asked if he wanted a Thinkpad or an MBP; that blew my mind.
They don't do big iron, but they move a lot of product and very little of it is faddish. They Get It, or have Gotten It historically, and a lot of people respond well to that. They might at some point no longer Get It, but until then, saying that they only have the iPhone going for them is wishcasting.
Apple isn't dominating any market.
In the PC market Apple is absolutely destroying all comers: 45% of the profit share in 2012. HP, Lenovo, Dell, Acer and Asus combine for the remaining 55%. (I couldn't find numbers for 2013) http://www.forbes.com/sites/timworstall/2013/04/18/apples-ma...
They're dominating similarly in the mobile market. http://www.businessinsider.com/apple-smartphone-profit-share...
I mean, if you want to say their products are crap - that's subjective and we can't argue that.
But if you don't think they're doing incredibly well in their key markets, that's diametrically opposed to the truth.
The thing you miss is that lot of people buy shit if it's cheap.
But really.. Apple is like that competitor that start miles ahead of any competitor, but because of some draconian moves and misteps, they end loosing all of their advantages.. it happen with Microsoft.. now its happening again with Google
Remember that time (well within the iOS era) that Apple claimed that FaceTime would be an open standard for interoperability?
I don't agree with RMS many times but in this one way he was 100% wrong when he targeted mono.
C# 2.0 is nearly a decade old now - and mono certainly hasn't limited itself to it. Absolutely none of C# 3, 3.5, 4.0, 4.5 was ever released as an open specification or as open source software until now - yet these are the versions mono and everyone was using.
So the "MONO FUD" was never actually FUD, it was legitimate concerns about the lack of openness of the technology. Those are no longer concerns (about this release version) - and we can be glad that RMS et al were "wrong" about this one (When they were in fact, absolutely right at the time.)
https://www.fsf.org/news/dont-depend-on-mono
So, five years later Microsoft makes another step in the right direction, and you think he's wrong because he didn't want to risk building the open source environment around something that could go away with a simple threat.
Btw, doesn't Microsoft make more money from Android patents than Microsoft phones? Microsoft does exercise their patents.
Edit: why the downvotes? If I started a social movement and everyone associated me with a watered-down alternative that tries to silence my views, I would be a little miffed.
I really like microsoft products and I use them every day, however the business model is too lock people in to their ecosystem so that you have to purchase licenses for their products.
Mono clearly undermines that b/c it allows you to "easily" jump ship to a free platform.
Unless I'm somehow completely misunderstanding their business model - killing Mono just makes business sense.
* I work for MSFT, on the Cloud.
I think what RedHat is doing is nice, but it's really not on the same scale to what Microsoft does; not to mention that they essentially have no monopoly/lockin - b/c anyone can start doing what they do (ie. start providing support for CentOS).
The synergy of microsoft services is closely tied to their ability to lock in developers and customers. Devs like C#, they write code to run on windows, windows gets more OS-exclusive software, more people buy Windows and Office, etc. etc.
Mono means people can run their C# work on free software, which mean they stop looking any better more appealing than the competition
This current move from MS suggests that they may not be being evil in this case, and that's great, but it doesn't make people with concerns about MS FUD-merchants at all. In fact, that's kinda insulting - most people accept that past behaviour is a good predictor of future behaviour so being cautious and sceptical about MS was not an unreasonable stance.
What I hear privately from the people I meet who work on the .Net and Visual Studio side of things is consistent with what the company is now saying publicly.
Basically: "We luvvv Mono" :-)
It's all part of the same thinking. RMS et al think - not unreasonably - think that this is Microsoft's business plan, so they really don't want the open source community investing their time on a platform they think will get clamped down to work only on a proprietary system.
You know, I tried that. I've tried to run about 3 .NET programs, written for Windows, under Mono on Linux. They all have had some library that was never part of the "open spec" that, therefore, ruined their chances of running on anything but Windows. The experience(s) put a nice, neat pin on the board as to what "open" meant when it came to Mono. It was a lot of double speak. And it's exactly this long-suffering characteristic about Microsoft that makes people like me take these current announcements with a large grain of salt.
Not depending on them for anything is, has been, and will be a fine strategy.
Until they really make free enough licenses not only for the core stuff but for the crucial libraries nobody can expect to be off-the-hook of the traditional lock-in.
And even if they do all that, it doesn't mean anybody was "wrong" only that the conditions are changing and allow reconsideration.
well, Microsoft used to say that open source software was anti-american, cancer, communist, anti-competitive, etc... yup I think they changed a bit
No WPF?
Result: we now have to ship a .reg file to our clients before we can launch our shit via ClickOnce. That's only 2000 people we now have to hand hold through this.....
There are no less than 3 similar but different and incompatible UI libraries that can be instantiated by XAML: WPF, Silverlight (WPF/e), and Windows store/phone (WinRT). WPF is dead like WinForms but the only solution for normal windows apps unless somethings changed with WinRT to make non full screen apps with.
If you have ever tried to develop across more than one of these with a shared code base you will know how frustratingly similar yet how incompatible and different they are. They are not supersets or subsets they are forks.
WinJS is the other.
But my guess would be that Ballmer stepping down was sort of symbolic in this regard. With a new CEO it looks more like a "new Microsoft" doing all these things.
And he knew it.
With the code released, can you:
- Run a ASP.Net web server?
- Run a classic C# .Net project better than with Mono?
The open sourcing of Roslyn can only mean good things for Xamarin.
I think you wanted to say that the "change" was made only thanks to the exit of Ballmer, but I think that's an exaggeration.
I think it'd probably be more accurate to say that these particular changes are symptoms of the same corporate change of direction that the change of CEO was, not that the change of CEO specifically resulted in these releases.
P.S. We're the Visual Basic compiler too :)
Man, this is a crazy world. Cats and dogs living together. Open source MS projects. The falcon cannot hear the falconer
This is basically old hat for me -- I'm just glad Microsoft agrees it's the way to go. :)
We're looking at a lowered on-ramp to .net apps that run more smoothly than java+linux+every-support-tool-bit (dev-wise) across devices (and things). It's compelling.
Everything from dev to deployment and ops is getting touched at this conf. This shows where MS's strengths can present differentiation. Their profit source is, and has been, building/selling tools and platforms for others to run businesses. When the thing people hate becomes more like what they aspire to have/be, the tension is palpable.
Google is about the only other player nearly capable. But Google's mission has far less to do with selling shovels vs. harvesting and monetizing.
Apple is just not in this space. If anything, they sure feel like a device-only company in this conversation. WWDC is about the garden. Build is handing out spades, shovels, and, now, dirt.
Xamarin's Evolve, last year, felt a bit like all this does. The are big things happening. If nothing else, at least it feels like change, if not progress for devs and consumers.
The garbage collector for example would be very interesting to peruse. Sure threading and synchronization code would be differ but these have been emulated on Linux before to good effect, I don't see any serious barriers there.
I'm sure the codebase is factored internally to have some sort of platform abstraction.
If you already have a compiler in the language you're compiling, you can update the compiler with new features by the following process, called "bootstrapping". This process is used in gcc for example:
1. Use the old binary version to compile the new source version.
2. Use the binary produced in step 1 to compile the new source version.
3. Use the binary provided in step 2 to compile the new source version.
The results of stages 2 and 3 should be the same (assuming the old and new compilers assign the same semantics to the compiler source code and don't have non-determinism, e.g. using wall-clock time to determine when to stop searching for optimizations).
The bootstrap process can't be used on the first compiler for a brand-new language, because there is no "old compiler" that can be used in stage 1. The only way around this is to write the first compiler in a different language that already exists.
Of course, if a self-hosted compiler is your goal, you can afford letting the very first compiler (the one written in another language) be limited, "dirty," or "hacky". You don't have to implement the entire language in the first compiler; the first compiler just has to support enough of the new language to allow you to write the second compiler.
Anyway, once you have the first compiler, you write the first version of the second compiler in whatever subset of the language the first compiler supports. Once the first compiler builds the first version of the second compiler, the first compiler is no longer needed; you can add all new features to the second compiler only.
New versions of the second compiler (perhaps supporting more language features) can now be produced by bootstrapping. Of course, you can also use the second compiler to compile a totally different compiler implementation written in the new language (this would be the third compiler for the new language).
Bootstrapping a compiler from nothing is harder and involves a lot of incremental steps. http://en.wikipedia.org/wiki/Bootstrapping_(compilers)
Some other interesting contemporary examples:
I think as we speak Go is working on its plan to switch from using a compiler written in Go[2].
Rust[3] has kind of a hybrid approach where the front-end of the compiler is written in rust, but generates llvm bytecode which is then compiled the rest of the way by llvm itself. It takes three "stages" for them to do a full compile, first a binary "snapshot" compiler is downloaded and used to compile the compiler (stage 0), then that generated compiler is used to compile the compiler again (stage 1), and then that generated compiler is used to compile the compiler again (stage 2). Stage 2 is actually just a test - its output should be identical to that of stage 1, and if it isn't, something went wrong.
There is a classic lecture from Ken Thompson[4] that I think is really illuminating on how this stuff works, if you're unfamiliar with it.
It's really fascinating stuff!
[1]: http://homepage.ntlworld.com/edmund.grimley-evans/bcompiler.... [2]: https://docs.google.com/document/d/1P3BLR31VA8cvLJLfMibSuTdw... [3]: http://www.rust-lang.org/ [4]: http://cm.bell-labs.com/who/ken/trust.html
Then they wrote the second C# compiler in C#, and used the first compiler to compile the second compiler. Voila!
http://c2.com/cgi/wiki?StructureAndInterpretationOfComputerP...
And tends to happen to other languages... you just have to bootstrap the first compiler.
Interestingly, no -- but neither can the native C# compiler (that's what we call the old C# compiler that everyone's using in VS right now)!
Why? The C# compiler is bootstrapped, so we actually compile the compiler with the compiler. It just so happens that a while ago we felt we needed a feature so badly for the C# compiler that we introduced it into the C# language (exception filters) and started using it in the compiler immediately. So in order to compile the compiler, you need a C# compiler that supports exception filters, which right now is just the Roslyn C# compiler. Honestly, the feature (and some other small changes we've made) aren't that complicated, so they could be backported to other compilers, but I don't know why you would want to do that.
Lots of people use it
It's also somehow satisfying to witness Xamarin's success and further opening of .NET after all the bashing and hating Mono received from the free software community.
Now that MS has released the stuff under Apache license things have really changed.
C# is a great language, and I hope to see it flourish outside of the MS walled garden. Miguel de Icaza does what he can with Mono, but it can be so much more.
I really like C#. But it's currently pretty close to useless anywhere other than Windows - Mono doesn't even come close.
Hopefully this will close the gap.
But I still hope Microsoft does well, because frankly I like .NET
Competition is great for everybody and Microsoft is making all the right moves!
But not much behind ASP.net.
http://www.techempower.com/benchmarks/#section=data-r8&hw=i7...
And quite a bit ahead of Sinatra
http://www.techempower.com/benchmarks/#section=data-r8&hw=i7...
Then again, so's Perl. The joys of implementation differences.
I find Nancy a lot easier to develop with than ASP.net. You do pay for the expressiveness however.
At work I'm deploying a Nancy based project to parse/process documents into structured data using a 3rd party vendor's SOAP service. There were some unpleasant compatibility issues with the SOAP service and our Perl server's SOAP client and zero logging so I took the service's source code (so glad the Vendor provided it) and replaced the WebService portion of it with a thin Nancy controller to route the calls and return JSON. Since most of the processing time is spent chewing on the files Nancy's overhead isn't an issue and Nancy made it dead simple to code.
There are a number of things people are doing, based on Linux, which are basically using Linux as an OS and then layering on some custom drivers or such into a product. Whether its a web 2.0 company using it as the server OS or an embedded signage company. All of these were "impossible" when you had to have your own OS team to support them, and Microsoft benefited from that. Now the OS "maintains itself" (such as it is) and so businesses get everything they got from employing Microsoft tools but at a much lower effective cost. They don't need to pay big license fees, they don't need to hire programmers to maintain a lot of code that isn't central to their product, and they don't have to spend a lot of money/time training people on their own infrastructure. That is a pretty big change.
Its nice to see folks realize it isn't the software that is valuable, its the expertise to use it that has value. By open sourcing the C# compiler Microsoft greatly increases the number of people who will develop expertise in using it and that will most likely result in an increase of use.
And here's some notes on bridging between Haskell and Unity http://lambdor.net/?p=321
That's not to say Unity is better, it was just easier to use.
Cloning into 'roslyn'...
remote: Counting objects: 10525, done.
remote: Compressing objects: 100% (4382/4382), done.
remote: Total 10525 (delta 6180), reused 10391 (delta 6091)
Receiving objects: 100% (10525/10525), 16.94 MiB | 1.69 MiB/s, done.
error: RPC failed; result=56, HTTP code = 200
edit: Ah nice, Apache 2 license explicitly calls out a patent license is granted for use. I wonder how much cajoling it took to get the lawyers to agree to that!
See Locke1689's comments here, especially:
https://news.ycombinator.com/item?id=7524722
"the native C# compiler (that's what we call the old C# compiler that everyone's using in VS right now)"
Really, this seems better than opening up the old compiler that they're moving away from.
You can view and download the whole .NET source here: http://referencesource.microsoft.com/
Is this the return of the original MS?
Certainly Microsoft wouldn't mind just throwing some millions at them and buying them outright, so are we to deduce that any such offer was rejected?
Teams under his lead have been open sourcing things in expanding significance for years. As Scott has been steadily climbing the Microsoft chain of command and with each step of the ladder it seems larger projects got open sourced.
To use java as an example, this would be like open sourcing javac without open sourcing the JVM. (Just fyi, both javac and JVM are open source under the GPL license. :) [edit - Looks like large parts of .NET platform are being released]
Still, this is a move to be applauded. Microsoft has been easing itself into a more open source friendly company over time by putting more projects under an open license. Microsoft ASP.NET MVC is another good example. There is an incomplete list of open source projects originating from Microsoft here: http://www.microsoft.com/opensource/directory.aspx
[edit above in brackets]
Edit: I confused roslyn with the recently announced AOT compiler.
Makes me think of Sun open sourcing Java.
So while this development is positive, one should take it with a grain of salt. MS still can't be just trusted blindly.
See slide: https://pbs.twimg.com/media/BkT9oBcCQAAHIAV.jpg:large
Here's to hoping that Microsoft goes beyond the compiler and open-sources the VM and the whole standard library too - .NET and C# always had great potential, it's a pity that Microsoft saw it as a way to sell Windows. And I hope they realize they have a long way to go before they catch up.
http://roslyn.codeplex.com/SourceControl/latest#Src/Compiler...
Something I've just learned from looking at the code is you can jump between cases in a switch statement :
switch (a) {
case '1':
...
case '2':
goto case '1';
}
Although there is no reason why they wouldn't use goto. The oft-cited "goto statement considered harmful" was in a very different context and basically just ranted against using goto when there are control structures that make intent clearer.
[1] http://www.microsoft.com/en-us/download/details.aspx?id=1412...
It has actively been pushed by some of Microsoft's evangelists (Phil Haack (ex employee, works at github now i think) and Scott Hansselman to say the more popular names).
I believe they got some playfield to do things and now the community has more and more impact (eg. Nuget and software like myget which is based on Nuget (Nuget for Enterprise))
Also, the CEO isn't Balmer anymore, that probably helps to.
[0] - http://roslyn.codeplex.com/SourceControl/list/changesets?pag...
With Valve pushing their Debian fork and more gaming support for Linux in the last time, Microsoft wan't to appeal to the Open Source community the reduce the "bashing" which ... which could actually loose some force behind it. Not that it could actually benefit Linux with better Mono support etc.
Microsoft, please add unix terminal instead of start button in Windows 8.
You can always use a bash/cygwin prompt for that if you want.
I think this will help .Net devs make smarter decisions about their code now that they can see what's happening in the background.
Generally I've not been very impressed with the quality of Android's documentation and firmly believe a company such as Google can do better. And after having a look at the process around getting a patch approved I deemed it not worth my time trying to contribute documentation fixes and enhancements.
What does this mean for the future of C# on Linux?
It's not an advertising company like Google. Google makes money when you use the Internet; Microsoft makes money when you pay for its software.
Microsoft's endgame is in sight.
I suspect that Microsoft is trying to fight their way into the Server space where Linux installations are prevalent by offering services via Azure.
So, for example, you have an extensive Linux environment running critical back-end services. You can have a BASH script which generates C# code, compiles it and eventually deploys it to Azure. This is I think a fantasy scenario that Microsoft is envisioning.
I don't understand what you're trying to say here. The roslyn compiler can generate code that can be run on the Mono VM. It was even demonstrated by Miguel on stage.
You'd also need something to run the roslyn compiler on linux, and that again would be mono.
Visual Studio will be using the roslyn compiler in the next version also.
Azure doesn't really have anything to do with this. It's an open source compiler that can run on the .NET VM or the Mono VM, and generate output for either.
Also, I read a lot of comments saying this way good for mono ... how is this ? Wouldn't an open source CLR be more useful ?
But, I dunno. I'm extremely skeptical of Microsoft's ability to put long-term momentum into any of their non-core strategies. All these things are one re-org away from becoming basket cases.
Case in point: XNA
git clone https://github.com/microsoft/windows.git