IronRuby dead

Charles said:

exoteric said:

*snip*

I don't agree. Type inference and dynamic typing are not the same thing, obviously. The point of all of this is dynamic runtime behavior and programming languages that support dynamic type systems. I don't want to go down the static versus dynamic rabbit hole. It's too deep and generally doesn't lead anywhere. As far as C#'s dynamic, it requires the DLR to produce expression trees and what not that the CLR can actually do something with...

 

At any rate, we're just having a conversation. I don't want you to think that I am right - you're smarter than I am, gentlemen. Remember that.

C

No worries Charles.

The static types are still there but as a programmer you don't have to "type them down", which is what gives F# the dynamic feel. That doesn't account for the full expresiveness of dynamic languages but it does fool some programmers into thinking that the language is dynamic, at least newbies.

exoteric said:

Charles said:

*snip*

No worries Charles.

 

The static types are still there but as a programmer you don't have to "type them down", which is what gives F# the dynamic feel. That doesn't account for the full expresiveness of dynamic languages but it does fool some programmers into thinking that the language is dynamic, at least newbies.

"That doesn't account for the full expresiveness of dynamic languages"

What expressiveness?

Bass said:

AndyC said:

*snip*

Like what? Another rewrite of WF?

 

I'm just finding it ironic that .NET which seems to have billed as a "multi-language runtime" keeps losing languages. While the JVM (never ment for anything but Java) seems to be where all the language development action is these days. Good work Microsoft. Keep killing your development story.

 

I'm sure everyone in the world will drink the Kool-Aid and use C#.

FUnny ...

Eh Bass:

1) why must microsoft support more than say 2 languages?  the .net runtime and framework allow anyone to create or port a compiler to do whatever languuges they want.

2) i recall that in programming we have seen dozens of languages over the years but folks focus most of the time on working with just one or two of them .... so with C# and the C/C++ trends this seems rather to be expected

Things change and yet they stay the same....

figuerres said:

Bass said:

*snip*

FUnny ...

 

Eh Bass:

 

1) why must microsoft support more than say 2 languages?  the .net runtime and framework allow anyone to create or port a compiler to do whatever languuges they want.

 

2) i recall that in programming we have seen dozens of languages over the years but folks focus most of the time on working with just one or two of them .... so with C# and the C/C++ trends this seems rather to be expected

 

Things change and yet they stay the same....

Because people use more then two languages. Personally I think everyone should STFU and use Java.

Oh wait.

Do you get what I mean though?

You might be a C# programmer and you probably think everyone is using just "one or two of them". But that's because you are pigeon holed into that niche. IE. Oracle DBA will think everyone uses Oracle, because he is pigeon holed by his own Resume into a career as an Oracle DBA. So all he will ever see is Oracle this Oracle that. But that is a false idea. 

Charles said:

exoteric said:

*snip*

It probably doesn't signal the end of Microsoft's commitment to supporting dynamic languages as much as it may signal the end of spending time and resources trying to make the CLR do what it is not intended to do... That said, these are my opinions and not those of Microsoft. DevDiv is well equipped to communicate their thinking - I am not doing it for them as that is not my place (since I don't work in DevDiv and I have absolutely nothing to do with designing or explaining their strategies).

 

In my opinion, Windows should just run any language/runtime incredibly well. This would make it easier for developers to target our platform with the tools they know and love. But, again, I am just speaking my mind, sharing my perspectives which are, by the way, often incorrect and not representative of reality.

 

C

So you think the CLR is made worse by running a dynamic language?

Charles said:

Minh said:

*snip*

Of course not. What makes you assume this? I said the CLR, today, is a static runtime, through and through.

C

it may signal the end of spending time and resources trying to make the CLR do what it is not intended to do...

I think the inclusion of the dynamic libraries in SIlverlight gives it so much potential... and it could not have happened without the work on IronRuby and IronPython. If you count the # of people working on the DLR vs the CLR it's not even funny. It's unfortunate MS decided to cut short term cost and not explore the long-term potential of the DLR.

Charles said:

Minh said:

*snip*

Yes. The CLR team was not shrunk in size, remember. One could conclude that if we are committed to supporting dynamic type systems, well, we would add this capability directly to our de facto managed virtual machine. I don't know... Or, to my point, we would just support and executionally enhance dynamic runtimes we don't also create and ship.

 

C

It's all bits and bytes under the hood, isn't it? Just because the DLR runs on top of the CLR, it doesn't mean it's a 2nd-class citizen, now. VB.net requires an external DLL... and no one considers VB a 2nd-class citizen.

Charles said:

Minh said:

*snip*

I'm not going down this rabbit hole, Minh.

C

In an alternate reality, Alice went home, became a housewife, and lead a perfectly normal life

Summing up my thoughts about the "news" and various comments here.

1. Microsoft is being EXTREMELY stupid and short sighted if they really are dropping the Iron languages. Without dynamic languages, the .NET world becomes just one giant hammer instead of a toolbox.

2. Charles's comment about dynamic languages not fitting in with the CLR means that Microsoft should also be dropping PowerShell. Nuff said, I think.

3. Master Pie's complaints about Ruby are really complaints about dynamic languages not Ruby. While I share some of the complaints, having written a large desktop application entirely in ECMA Script, I will still turn to dynamic languages for many tasks. Hammer and nail sort of thing, you know?

4. I have to strongly disagree about Ruby/Python mostly running on Linux. I've used many applications on Windows that are widely popular that were written in both languages. I've also done a lot of scripting on Windows using both languages. On Linux, on the other hand, outside of those popular programs I mentioned running on Windows you're far more likely to find things written in Bash, Perl and PHP.

5. I've already disputed rhm's assertion that Ruby developers are unlikely to be using Windows, but I'll now argue about "deploying on Windows". That's hogwash. First, if it's strictly Ruby and no dependency on any .NET libraries, you can deploy to Linux using the standard Ruby runtime. Second, if you're using .NET libraries the IronRuby platform runs on any platform that Mono runs on. And the old "Mono isn't complete" argument, that I've always found lacking anyway, really doesn't matter much here, as the "missing" libraries on Mono are libraries that you're probably not using with IronRuby in the first place.

6. Next rhm questions whether or not .NET devs would be interested in Ruby, and the answer is emphatically YES. The .NET OpenSource folks already use tools like Rake for building, Cucumber etc. for BDD testing and there's a huge movement right now to use Gems for package management. Then there's the whole need to allow your application to be scripted by end users and the benefits of allowing them to use any "Iron" language to do so. I could go on, but the point is made. .NET devs are already dependent on Ruby/"Iron Languages", much less have a desire to use these languages.

All in all, if Microsoft is abandoning dynamic languages on the CLR they are screwing up massively. If they are just dropping the current Iron languages, they are still making a huge mistake, both politically and even for technical reasons.

Minh said:

Charles said:

*snip*

It's all bits and bytes under the hood, isn't it? Just because the DLR runs on top of the CLR, it doesn't mean it's a 2nd-class citizen, now. VB.net requires an external DLL... and no one considers VB a 2nd-class citizen.

The DLR is an excellent tool to add dynamic features to static languages, but maybe - just maybe - it's not the optimal solution for a purely dynamic language.

While this doesn't prove anything, it's at least suggestive that the IE team developed Chakra from scratch instead of basing it on .NET and the DLR.

wkempf said:

Summing up my thoughts about the "news" and various comments here.

 

1. Microsoft is being EXTREMELY stupid and short sighted if they really are dropping the Iron languages. Without dynamic languages, the .NET world becomes just one giant hammer instead of a toolbox.

 

2. Charles's comment about dynamic languages not fitting in with the CLR means that Microsoft should also be dropping PowerShell. Nuff said, I think.

 

3. Master Pie's complaints about Ruby are really complaints about dynamic languages not Ruby. While I share some of the complaints, having written a large desktop application entirely in ECMA Script, I will still turn to dynamic languages for many tasks. Hammer and nail sort of thing, you know?

 

4. I have to strongly disagree about Ruby/Python mostly running on Linux. I've used many applications on Windows that are widely popular that were written in both languages. I've also done a lot of scripting on Windows using both languages. On Linux, on the other hand, outside of those popular programs I mentioned running on Windows you're far more likely to find things written in Bash, Perl and PHP.

 

5. I've already disputed rhm's assertion that Ruby developers are unlikely to be using Windows, but I'll now argue about "deploying on Windows". That's hogwash. First, if it's strictly Ruby and no dependency on any .NET libraries, you can deploy to Linux using the standard Ruby runtime. Second, if you're using .NET libraries the IronRuby platform runs on any platform that Mono runs on. And the old "Mono isn't complete" argument, that I've always found lacking anyway, really doesn't matter much here, as the "missing" libraries on Mono are libraries that you're probably not using with IronRuby in the first place.

 

6. Next rhm questions whether or not .NET devs would be interested in Ruby, and the answer is emphatically YES. The .NET OpenSource folks already use tools like Rake for building, Cucumber etc. for BDD testing and there's a huge movement right now to use Gems for package management. Then there's the whole need to allow your application to be scripted by end users and the benefits of allowing them to use any "Iron" language to do so. I could go on, but the point is made. .NET devs are already dependent on Ruby/"Iron Languages", much less have a desire to use these languages.

 

All in all, if Microsoft is abandoning dynamic languages on the CLR they are screwing up massively. If they are just dropping the current Iron languages, they are still making a huge mistake, both politically and even for technical reasons.

re 1: .NET has always been about integrating with other languages/APIs without everything having to compile down to CIL. If the Iron languages go away (which they will only do if there's no interest) then that doesn't leave .NET in an island that precludes using any other language in a .NET solution. CouchDB runs on the Erland virtual machine, but they still use a seperate Javascript runtime with it. IE has a Javascript runtime and also loads code in Java and in .NET if required to. Interop is just as powerful as being able to compile everything down to the same VM.

re: 2. Powershell isn't a dynamic language. It has an elaborate type inference system to avoid having to declare types of variables, but it relies on the standard .NET facility for dispatching function invocations. It doesn't use the DLR (predates it by a long time in fact).

re: 5. Why would the bulk of current Ruby users (the Mac and Linux crowd) take a dependency on Windows or Mono just to link to .NET libraries? There's tons and tons of open source C and Ruby libraries that they can use already. Even as someone who loves .NET, I can't see any reason why a rails dev would start using .NET technologies and I haven't heard of anyone coming across from the Mac/Linux using Ruby world to Windows because of IronRuby.

re: 6. If there's so much interest in IronRuby from .NET developers then there's nothing to worry about. It's already open sourced. If IronRuby developers are, as I suspect, in a minority of a minority, then I can understand the panic.