Future Focus I: Dynamic Lookup

Article
01/25/2008

by Charlie Calvert and Mads Torgersen

What is Future Focus?

This is the first of a series of monthly posts designed to give insight into the C# team’s current plans for future versions of Visual Studio. Each post will highlight one or more key subjects that will impact users of the C# language.

It is important that readers of this column have the right expectations. The information in this column is meant to be a helpful guideline for C# developers, and not a binding commitment. We are not attempting to give a complete list of features in the product, but only to share what we can in a way that will be easily accessible to all C# developers. The Visual Studio schedule, unforeseen technical problems, intellectual property rights and competitive pressures may impact our schedule or our ability to share our plans. We will, however, do our best to keep you up to date on the latest news from the team as they design and implement future versions of the C# language.

Future focus is not designed to present a detailed specification of future features. Instead, its purpose is to outline in broad strokes, and easy to understand terms, the directions that team will take in the future.

Dynamic Lookup

The next version of Visual Studio will provide a common infrastructure that will enable all .NET languages, including C#, to optionally resolve names in a program at runtime instead of compile time. We call this technology dynamic lookup.

Work on support for dynamic lookup was begun in the CLR, but soon became part of the Dynamic Language Runtime, or DLR. The DLR provides the infrastructure on which a common set of dynamic tools can be built. For instance, the DLR provides the infrastructure for both IronRuby and IronPython. It will be the infrastructure on which the C# team implements dynamic lookup .

Support for dynamic lookup is already available in Visual Basic for .NET, where it is often known as “late binding”. The new release of .NET will provide C# developers with similar functionality, while at the same time providing a shared infrastructure for runtime name resolution across all .NET languages, including VB.

Useful Scenarios

There are three primary scenarios that will be enabled by the new support for dynamic lookup:

Office automation and other COM Interop scenarios
Consuming types written in dynamic languages
Enhanced support for reflection

Office Automation

In the next version of Visual Studio Office automation will be easier. Developers will be freed both from the need for using a bulky type library, and the need for including optional arguments in their method calls. The support for Office Automation will be part of a general effort to enhance support for COM Interop and Office PIA.

Consuming Dynamic Languages

Dynamic languages such as IronPython or IronRuby are becoming increasingly popular. At this time, those languages can call C# code, but we can’t easily call into their code. The next version of Visual Studio will simplify the steps C# developers take to call into IronPython or IronRuby classes. This will give developers access to a useful existing code base, and an alternative way to write new code.

Call Reflection

C# developers can currently use reflection to instantiate classes and call arbitrary methods that are not known at compile time. The dynamic extensions to the C# language will make it much easier to make such calls.

Syntax

The syntax that will be used for dynamic lookup has not yet been finalized. The code that I show here is therefore only a tentative sketch that reflect the team’s evolving plans.

The team is currently considering adding the keyword dynamic to the language and using it to demarcate a block of code:

 static void Main(string[] args)
{
    dynamic
    {
        object myDynamicObject = GetDynamicObject();
        myDynamicObject.SomeMethod();         // call a method   
        myDynamicObject.someString = "value"; // Set a field
        myDynamicObject[0] = 25;              // Access an indexer
    }
}

All the code that occurs in a dynamic block will potentially support dynamic lookup; even if the accessed members are not known by the C# compiler to exist, it will allow the code. At runtime the DLR will look at the actual object referenced by myDynamciObject for members with those names. It will access them if they do indeed exist, otherwise an exception is thrown. Outside of a dynamic block developers can only call C# code statically, just as they do today.

The details of the compile time process have not yet been determined. For instance, the compiler might treat all methods in a dynamic block as dynamic and only attempt to resolve them at runtime. Alternatively, it might first try to resolve them statically, and if that fails it will attempt to resolve them dynamically at runtime. As we gain more clarity on our design of this technology we will publish many more details.

Summary

In this edition of Future Focus you have learned about the team’s current plans for enabling dynamic lookup in the next version of the C# language. You have seen one tentative plan for enabling this syntax. You have also seen that dynamic lookup will be enabled for three code scenarios:

Office Automation in particular and COM Interop in general
Consuming types written in Dynamic languages
Enhanced support for Reflection

Mads Torgersen is a Senor Program Manager on the C# team. He has been working closely with Anders Hejlsberg and other key members of the C# team as they develop the plans for the next version of the C# language. Charlie Calvert is the C# Community PM.

Comments

Anonymous
January 25, 2008
You've been kicked (a good thing) - Trackback from DotNetKicks.com
Anonymous
January 25, 2008
What's the usability scope of the "dynamic" keyword? If it's only inside method/property members, that will be quite limiting on its use.
Anonymous
January 25, 2008
Interesting concept. I imagine the runtime will do all sorts of nice optimizations to make sure that this runs nicely. However, given what you have now, what's to stop you from implementing this as a compiler trick? What is integration with the CLR going to give you over having reflected calls emitted by the compiler?
Anonymous
January 25, 2008
casperOne, The biggest thing that this gives you over reflection is ease of use. This is a much simpler syntax than we had through reflection as we know it today. Also, there is new functionality here in terms of COM interop and interaction with dynamic languages. Today those features are either missing altogether, or are very hard or awkward to use. I think it is a little early to start talking about performance. We are really giving you an early look at this technology, and it will be some time before we see it in more depth. Certainly I hope the performance will be great, but it is too soon to start making any claims.
Anonymous
January 25, 2008
The comment has been removed
Anonymous
January 25, 2008
The comment has been removed
Anonymous
January 25, 2008
Hi all, it's me again. For those who don't know me, I'm a tester on the C# IDE team at Microsoft. I already
Anonymous
January 25, 2008
@Charlie Our typical use for late-binding was Interop/Office scenarios. We would create a single class to act as a facade over these COM resources. Sometimes COM classes can be expensive to instantiate. In these cases, we would instantiate them at the constructor and set a private field. Also, COM references would frequently cross method (but not class) boundaries as we would do normal internal refactoring. For anything but dead simple scenarios, it would be difficult to accomplish what we need in one method. As large as the Office interfaces are, there's not a whole lot you can accomplish inside a single block before it gets complex and unmaintainable. In these cases, I wouldn't necessarily need namespace-scoping, but file or class scoping would be sufficient.
Anonymous
January 25, 2008
Hi all, it's me again. For those who don't know me, I'm a tester on the C# IDE team at Microsoft
Anonymous
January 25, 2008
I have been searching for any clue about the next version of C# and what the features to be included,...
Anonymous
January 25, 2008
Great news that you are considering changing the access to Office Automation. I hate the optional parameters and the "missing, missing, missing" throughout the C# code. :-) Klaus
Anonymous
January 26, 2008
The comment has been removed
Anonymous
January 26, 2008
What kind of control do I have over it? That is, do we have a method missing hook?
Anonymous
January 26, 2008
Finally, some news for next version coming up... Interesting ..., i ll give my feedback ... at first sight it seems to provide easy way for reflection...but there can be some interesting talks to it, like scoping or performance related..
Anonymous
January 26, 2008
would be awesome of this worked on anonymous types
Anonymous
January 27, 2008
Charlie Calvert has posted an article that is co-authored by Mads Torgerson on one possible implementation
Anonymous
January 27, 2008
What about the return types of the members you call dynamically. If we write something like: string x = (string)myDynamicObject.SomeStringMethod(); would the runtime call SomeStringMethod and then try and cast the result to a string or check that the result can be cast to a string when it dynamically binds and throw an exception at bind time if it found that it was not possible? Would the runtime attempt to bind the entire dymanic block before it executes it (better since this would be more defencive) or dive in and try its best to perform the calls (and possibly cause an exception after only some of the calls had run)?
Anonymous
January 28, 2008
Adding to the Scoping of the Dynamic Object. We too use the object at the class level. For example we write an export class that has a single handle into Excel. Having the scoping at the method level will severly limit the cross method calls. Unless we can declare the object as DynamicObject in the calling method. private void CreateHeaders(dynamic object excelObject, string[] headers) { dynamic(excelObject) { } }
Anonymous
January 28, 2008
The comment has been removed
Anonymous
January 28, 2008
Mmmm that sounds so VB... does it mean that the F# tunics have left the team and hand been replaced again by the VB Ties?
Anonymous
January 28, 2008
The comment has been removed
Anonymous
January 29, 2008
What I don't understand: If everything inside the "dynamic"-block uses dynamic lookup, then you'll loose type safety for everything inside this block, not just the objects you're really needing it for. This is a much broader granularity IME than declaring a single object as dynamic and leaving all the other objects typesafe. In a typical scenario you'll need every access to a specific object to be dynamic. If you just declare this single object dynamic, you'll have fewer dynamic lookups than having many dynamic blocks in a class that contain dozens of other objects that would normally be just fine with static lookup, but now use dynamic lookup, just because a single object in this block needs it.
Anonymous
January 29, 2008
Thomas, Thank you for this helpful comment. As you know, this is one of the subjects that is still very much under discussion. For instance, the team is considering first attempting to resolve all code in a dynamic block statically, and using dynamic lookup only if the static calls can't be resolved at compile time. The team will take your feedback into account and I'll work to get updates to the community if and when the status on these issues changes.

Charlie

Anonymous
January 29, 2008
Thomas, Thank you for this helpful comment. As you know, this is one of the subjects that is still very much under discussion. For instance, the team is considering first attempting to resolve all code in a dynamic block statically, and using dynamic lookup only if the static calls can't be resolved at compile time. The team will take your feedback into account and I'll work to get updates to the community if and when the status on these issues changes.

Charlie

Anonymous
January 29, 2008
Welcome to the fortieth issue of Community Convergence. This week we have two new releases of note: We
Anonymous
January 29, 2008
Welcome to the fortieth issue of Community Convergence. This week we have two new releases of note: We
Anonymous
January 29, 2008
I agree with Thomas. I think that the typical scenario has us using dynamic lookup for all references of a single object, so it makes sense to declare the object as dynamic rather than a block of code. Something like this: class SomeClass { dynamic object _myLateBoundMemberObject; object _myStaticallyBoundMemberObject; private void SomeMethod() { dynamic object myLateBoundLocalObject; object myStaticallyBoundLocalObject; } } ~Steve
Anonymous
January 29, 2008
How about using -> instead of . for dynamic lookup? I'm only half joking!
Anonymous
January 30, 2008
The comment has been removed
Anonymous
January 30, 2008
If I understand what your doing, this is really just another C# compiler trick where your just going to replace it in IL with a reflection version. If got that right, then I also prefer a new Dynamic Lookup method operator instead. '->' might cause many C++ refuges some heartburn, so how about: myLateBoundLocalObject'Method1() or myLateBoundLocalObject~Method1() or myLateBoundLocalObject@Method1() or myLateBoundLocalObject..Method2() or even reverse it to make them stand out more aka Method1() of myLateBoundLocalObject or Method1()@myLateBoundLocalObject I think if a good operator is found it would make the code much cleaner. I personnally like the '..' version as it would be easy to remember. One period for Static bound calls, Two periods for Dynamic lookup. -Pablo
Anonymous
January 30, 2008
Quick suggestion: could the "dynamic" modifier be implemented as an attribute? That way the attribute could be applied to a variable, class, function, etc. That would address the scope issues presented earlier, and it would be easier to read and comprehend the code.
Anonymous
January 30, 2008
Kinda OT, but since somebody important could be reading this... Can we /please/ have something like ':' operator, so that following could be written: string greatGrandFatherName = person:Parent:Parent:Parent:Name; ...instead of tedious... if (person.Parent == null) return null; if (person.Parent.Parent == null) return null; if (person.Parent.Parent.Parent == null) return null; return person.Parent.Parent.Parent.Name; OK, this isn't the best example, but there are many uses for this (especially in ORM) and it makes syntax so much cleaner, IMHO. LP, Dejan
Anonymous
January 30, 2008
This reminds me of the times when Delphi first added late binding features in order to support COM automation scenarios...
Anonymous
January 30, 2008
I often rethink or have additions to my posts. This topic of what's coming in C# vNext is definitely
Anonymous
January 30, 2008
Hi. Thanks for a great post Charlie and Mads! I'll post some of my feedbacks here. What about a dynamic block and a new late bound operator (.. for example): object myObject = GetDynamicObject(); dynamic { int i = myObject..GetValue(); string s = myObject.ToString(); } This way you spesifically choose to do late binding and show exactly here you want it. But this kind of breaks the nice typesafe static nature of C# I guess...
Anonymous
January 30, 2008
This is so fun I have to put out another idea... What about some kind of inline interface declaration? object myObject = GetDynamicObject(); dynamic { interface (myObject) { void SomeMethod(); int GetSquare(int v); } catch { } myObject.SomeMethod(); int i = myObject.GetSquare(100); } This way you both declare a dynamic section and declare the exact members you need. There's actually no real late binding, just a syntactic sugar for reflection lookups. Another nice feature here I think is the possibility to catch if the lookup fails. You could even start a new interface failover declaration inside the catch.
Anonymous
January 30, 2008
David, Thank you for your comments. The team will consider them carefully. Please note that the proposed C# features discussed in this post are not designed to convert C# into a dyanmic language, only to make it possible to call code written in a dynamic language and to make COM interop simpler. I and others on the team hear and appreciate what you are saying in your second paragraph about the declaration of dynamic variables and I will pass that information on to everyone on the team.

Charlie

Anonymous
January 30, 2008
After posting the -> 'suggestion' I decided that '..' would be a reasonable looking alternative. So there you have it - independent evolution of the same syntax idea from multiple people - it must be a good idea :) The whole thing reminds me a bit of the old VB syntax for accessing fields on a recordset: Sn!FieldName 'instead of Sn.Fields("FieldName")
Anonymous
January 30, 2008
I vote strongly in favour of Thomas Krause's suggestion. I think a tagged block (similar to "unsafe") that changes the language rules for all code inside it would be a bad idea. Firstly, if you made it affect all types within the block, that's too coarse-grained - what if some calls could be static? This immediately leads to the idea of mixing dynamic and static lookup, as mentioned. Secondly, the main reason for using that tagged block approach is because you want to clearly call it to the attention of someone reading the code, but in fact if all the code in the block happens to be statically resolvable then the dynamic block would not make any difference, and so would be completely misleading to anyone reading the code. It's like a worrying comment that might have genuine implications, but might not. Far better to have a pseudo keyword made of two tokens, "dynamic object", which declares a reference type on which method/property access is always dynamic. This can then be used on local variable or member declarations. As for making it clear to the reader of the code, Visual Studio can do a much better job of this by highlighting all method/property accesses on dynamic objects. It would be great if they had a grey background or something like that - they'd stand out a mile and it would give instant feedback to explain why there was no intellisense happening.
Anonymous
January 30, 2008
Thomas, Daniel, Commenter, others, Thanks for these suggestions. This is great feedback and it is very much the type of thing we were hoping to see. I'll make sure all your ideas are passed on.

Charlie

Anonymous
January 30, 2008
The comment has been removed
Anonymous
January 30, 2008
Wouldn't it be cool to support both late binding and intellisense. I think optionally declaring methods and properties in the variable declaration could do. public class Program { // Without intellisense dynamic object dynObject = GetDynamicObject(); // With intellisense dynamic object dynObject = GetDynamicObject() { void SomeMethod(); } } This way you can opt for intellisense according to own needs. And it could also be possible to declare inline: dynamic (dynObject) { int GetSquare(int value); } int i = dynObject.GetSquare(100);
Anonymous
January 30, 2008
This is kind of off-topic, but since we're talking about C# future I would like to suggest an implementation for optional parameters. I understand why C# doesn't allow them since this transfers control from the user of a method to the implementor, and the implementor can choose to change default values at any time, possibly breaking the code using it. But what about declaring default values and letting intellisense auto-insert them for you? Here's a code example: public void PlaceOrder(int orderID, int price = 100); Then, when you write PlaceOrder intellisense will autofill the price param to be 100 if you want it to. This doesn't make a lot of sense when there's one parameter defining a default value, but with 10 it could really boost productivity.
Anonymous
January 31, 2008
I hope this just uses reflection under the hood and doesn't alter the .NET framework to accomodate late binding in anyway. Much like var in C# is just a compiler trick
Anonymous
January 31, 2008
Maybe this syntax instead: var myDynamicObject = p as dynamic; myDynamicObject.SomeMethod(); // call a method myDynamicObject.someString = "value"; // Set a field myDynamicObject[0] = 25; // Access an indexer
Anonymous
January 31, 2008
The comment has been removed
Anonymous
February 01, 2008
Kudos to Brian for editing everything last night, I think he's sleeping on my couch right now, but our
Anonymous
February 01, 2008
Kudos to Brian for editing everything last night, I think he's sleeping on my couch right now, but
Anonymous
February 01, 2008
This is Episode #1 of <insert.name.here>, a weekly recap show of our favorite things for developers
Anonymous
February 01, 2008
Personally, I have to side with the guys that like C# to stay static. LINQ is about the extent of "dynamic-ness" that I can tolerate. Class Libraries written in IronPython callable by C#? I honestly can't see many truly indispensible class libraries written in IronPython that couldn't otherwise be rewritten in a much more performant way in C#. Granted, usability is key here, but are we in the community asking for this? My assertion is "USE VB" since VB frankly is so close to C#, yet still has the late binding techniques built in.
Anonymous
February 01, 2008
Eric, there's a language for people like you (the instinctively conservative). It's called Java. ;-)
Anonymous
February 01, 2008
Eric, I understand your thinking about keeping C# statically typed. I sometimes wonder if the current duck typing pendulum has just swung to the far end of its period as it's known to do every few years. But I don't think so. This time, I think there is enough maturity in the marketplace that people are beginning to wonder why they can't have the best of both worlds. My response above about my experiences embedding Python in C# is an expression of the fact that I love the type safety of C# most of the time. But when it's time to do something dynamic, I regret having to step out of C# (or into deep reflection and lots of CodeDom trickery) to do it. Speaking of performance: for the dynamically-generated ClientBase<T> invocation of a WCF service in a tight loop versus using a statically-bound, SvcUtil-generated proxy called directly from C# that I described above, there is no difference in performance. The only real difference is that I never had to use SvcUtil in the former case and I really like that. So I can envision a world where static, early-binding and dynamic, late-binding get along just fine.
Anonymous
February 02, 2008
C# Futures - Dynamic Code Blocks
Anonymous
February 02, 2008
This is an exciting feature. I have several comments.

I really like the .. operator idea either instead of or in addition to a dynamic block. This is even more important in a maintainence scenerio than in new code. If one is adding 2 dynamic calls to some existing code the choices are a) 2 dynamic blocks or b) subtly changing the semantic of all the method calls between them. I really want more control over when I use dynamic invocation.
I hope you will expose the the name lookup logic through system.Reflection as well. I have been trying to write a generic "accessor" class that uses reflection to allow unit tests to access private and protected members of production classes. Doing the name resolution, based on a string member name, is easy to do wrong, and proving very difficult to do right. I would enjoy being able to give reflection a name and say "find me the member, public / private ... etc, that this name would map to."
WHat is going to happen when you do a Dynamic call on a RealProxy descendent. I hope that it just calls Invoke with the method name and arguements, and lets me work it out. (This would let me define some test objects with a very flexible interface.)
Add another vote for the : operator proposed by Dejan Stanic. I see this pattern all the time, especially when using Linq. John Melville

Anonymous
February 02, 2008
I've had a policy against posting on big news that's likely to be common knowledge in the Microsoft development
Anonymous
February 03, 2008
The comment has been removed
Anonymous
February 03, 2008
The idea of adding late binding to C# is a good one, but am not sure that going down the path of using a block marked by the dynamic keyword is a good idea. The rationale seems to be that "dynamic" would clearly flag a block of code that will have special rules, but the problem as I see it, is that there will be no way for the programmer to go back and forth inside this block to get strong typing (without doing plenty of gymnastics). By using the 'dynamic' block, everything inside this block that happens to be an object will be treated as a dynamic path and some errors that could have been caught will not be (accidental uses of it, return values that should really have been strongly typed). My preference would be to additionally declare variables with a keyword, an attribute or a new type, like this: dynamic p = GetDynamic (); p.Hello (d.World.DoSomething ()); This has the advantage that dynamic support will only be offered for "p" in this particular context, but still get full type checking with d, d.World and d.World.Something. Miguel.
Anonymous
February 03, 2008
The comment has been removed
Anonymous
February 04, 2008
Miguel, what about scoping? Consider this: ShoppingCart cart = new ShoppingCart(); dynamic jabberwocky = GetDynamicObject(); Product P = jabberwocky.SomeUnknownMethod( cart ); To do this today, I have a couple of options: (1) I can use reflection and find the SomeUnknownMethod signature for jabberwocky that best accepts a ShoppingCart, then call it, coercing the result into a Product reference. (2) I can instantiate a ScriptScope from the DLR, inject the ShoppingCart object using SetVariable, invoke some Python, for example, to do the work, then use GetVariable<T> to fetch the Product out of the script context. Personally, using the DLR seems like a better, long-term, more flexible option but it us certainly heavier than using simple reflection. I suppose the question I have for Charlie and others on the C# team is: how will late-binding in C# make it easy to write code like that above. I need to move data easily in and out of the dynamic scope from the surrounding static scope, whether it's a single object marked dynamic or a whole block. And it needs to be predictable with respect to the scoping rules in both forms. I can't help but think about how statement lambda expressions and anonymous methods reach out into the surrounding scope as needed. The problem with dynamism though is know which references should reach out and which ones should not. I suppose the more I think about this, the more I wish that the DLR were injectable into the statically typed scope, making (2) that I described above a bit easier to write. One last thought: assuming that it comes down to marking an object as dynamic (not a block), what would be the harm in enhancing the var "type" to do this, allowing: ShoppingCart cart = new ShoppingCart(); var jabberwocky = GetDynamicObject(); Product P = jabberwocky.SomeUnknownMethod( cart ); instead. -- Kevin
Anonymous
February 04, 2008
First impression: WOW Second impression: I am not entirely sure... i still go "wow" but I have maybe a few concerns. The dynamic{} code might look a bit awkward. Also what about totally prototype like code, like where objects are not tied to a particular class, but instead can be shaped lateron? And aside from this, what happens when objects are modified or extended at runtime, can these changes be reflected back upon the C# world? Anyway dont get me wrong, so far it looks great
Anonymous
February 04, 2008
Maybe I have an oversimplified view of this subject, but have you considered something like this? static void Main(string[] args) { interface IMyInterface { void SomeMethod(); string someString { get; set; } int this[int index] { get; set; } } dynamic IMyInterface myDynamicObject = GetDynamicObject(); if (myDynamicObject != null) { myDynamicObject.SomeMethod(); // call a method myDynamicObject.someString = "value"; // Set a field myDynamicObject[0] = 25; // Access an indexer } } where the dynamic keyword tells the compiler to emit code that at runtime compares the signature of the interface with the signature of the object. myDynamicObject would be set to an instance of the interface that is a proxy to the object if they are compatible otherwise it would be set to null. All the user of the object would compile and run as it does today.
Anonymous
February 04, 2008
The comment has been removed
Anonymous
February 04, 2008
Charlie Calvert blogged about dynamic support in C# 4.0 . I love this for two reasons. One it will enable
Anonymous
February 04, 2008
Charlie Calvert blogged about dynamic support in C# 4.0 . I love this for two reasons. One it will enable
Anonymous
February 04, 2008
Firstly, I apologize for my negative comment. I realize that things are early and that there are surely many other announcements in the near future. C#1 was a great start and C#2 was brilliant. Hands down the best thing about version 2 was the work done to support generics. However, I was extremely disappointed with C#3. Almost everything introduced was nothing more than syntactic sugar. It enabled nothing that I could not already do before. Now comes "dynamic lookup". Again, nothing that I can not already do with reflection. I'd like to see new features that enable new scenarios. Two in particular that I have been really missing are support for 1) generic variance and a way to do 2) generic operator overloading (can't constrain T to a static method). With that said, I'm looking forward to hearing more about the future of C#. Hopefully the next version will be more to my liking.
Anonymous
February 04, 2008
I agree with David. I use C# because it is statically typed. That is also why I stay away from VB and languages like IronPython etc. I do not want dynamic code in my C# code. Carlie Calvert wrote: "Please note that the proposed C# features discussed in this post are not designed to convert C# into a dyanmic language, only to make it possible to call code written in a dynamic language and to make COM interop simpler." Understood. However, it should be put in a library, not built into the language. I think that reflection is sufficient, but now that we have the DLR, just extend it as needed.
Anonymous
February 04, 2008
The comment has been removed
Anonymous
February 04, 2008
Any chance of having a similar keyword in Visual Basic? It's neat that I can already make dynamic calls in VB, but I believe it currently requires one to turn Option Strict Off. Something more fine grained would be awesome.
Anonymous
February 05, 2008
C#小组CharlieCalvert在其博客发了一篇有关C#语言未来方向的文章，这片文章介绍了一个叫做动态查找的特性，它为.NET语言（包括建立在DLR上的语言）能有一个统一的动态运行时名称绑定方案...
Anonymous
February 05, 2008
The comment has been removed
Anonymous
February 05, 2008
I think it will be very cool but probably unnecessary if the .Net frameworks let using multiple planguages in a class. A special project type class extensions may be defined or may be a syntax like below can be used: @Use Language=VB .Net ... Some VB .Net Code @End But by the way this may reduce readability and could make the code look uglier or messier.
Anonymous
February 05, 2008
The comment has been removed
Anonymous
February 05, 2008
Commenter, You asked: "How about using -> instead of . for dynamic lookup" This type of thing was discussed in depth. Early drafts of the team's plans tended to include it, but now the team is leaning away from it. No decisions have been made at this point in time, of course. And there were various different characters proposed, not necessarily focusing on the C like syntax you suggest. But still in the same general vein.

Charlie

Anonymous
February 05, 2008
I also strongly agree with Thomas. I'll also throw one other option into the mix. What about something like this object myDynamicObject = GetDynamicObject(); myDyamicObject."SomeMethod"(); The quotes around the method name make it clear that, as far as the compiler is concerned, the name of this method is just a string, to be looked up at runtime. To me, the main thing is to find something that doesn't require a block to be wrapped around the code, and which lets me (any my IDE's syntax highlighing) tell exactly which calls are dynamic - either by the variable they are made against (as in Thomas's suggestion) or by the way the call is written.
Anonymous
February 05, 2008
This belongs in a library, not baked into the language. We have reflection and we have DLR. You said that the DLR will be the "infrastructure on which the C# team implements dynamic lookup." If the existing DLR infrastructure is not sufficient, then why not just extend it? I fail to see what advantages special syntax will provide over a pure library implementation. It serves no purpose but to further pollute the language when a library is the appropriate place to be doing this. I am strongly opposed to this.
Anonymous
February 06, 2008
Based on my experience with VB (which really already has this), I agree with Thomas & Miguel. Blocks are the wrong granularity. Either variable or call level would be great.
Anonymous
February 06, 2008
L'�quipe de C# d�voile un peu plus sur les principaux prochains ajout du langage avec l'aide de rubriques nomm�s Future Focus. La premi�re de la s�rie concerne le Dynamic Lookup....
Anonymous
February 06, 2008
The comment has been removed
Anonymous
February 06, 2008
The comment has been removed
Anonymous
February 06, 2008
I'd much rather see something like this (as others have already stated) dynamic myDynamicObject = GetDynamicObject(); myDynamicObject.SomeMethod(); myDynamicObject.someString = "value"; myDynamicObject[0] = 25; or perhaps [dynamic] IMyInterface myDynamicObject = GetDynamicObject(); The block idea is a bad idea in my opinion. I'm sure it's easier to implement, but is that good for the language in the long run?
Anonymous
February 06, 2008
The idea of the interface posted by Hans gave me another similiar idea. I like the fact that c# is 100% type safe and I like that i can rely on that. But even so I do see where th Dynamic typing could be nice to have. but instead of dynamic { object MyObject = GetDynamicObject } how about: internal dynamic class MyDynamicBasedClass : DynamicClassIdentifier { public dynamic void SomeMethodOnDynamicObject(string someargument); public void MyWonMethod(){ //do nothing and return } } That would make the coding style the same as we have it today. It would be kinda type safe. It would meet the wishes of Thomas and others and it would be very readable and intellisense enabled (with a new icon for dynamic methods/properties o.c.)
Anonymous
February 06, 2008
The comment has been removed
Anonymous
February 07, 2008
what I'd really like similar to this is an infered duck typing. That way I get type safeness and intelisense but don't need adapters. In my example the 2 methods (f1 & f2) have return types which do not implement IHasText, however they both honour the interface. A new infer keyword could behave like the as keyword and return null if a type cannot be coerced into the supplied interface. Under the hood the compiler can follow the adapter pattern or what ever clever stuff is required :-) e.g. IHasText{string Text {get;}} Label f1(){} TreeNode f2(){} void func() { var ht = f2() infer IHasText; // use ht.Text }
Anonymous
February 07, 2008
Microsoft haven't committed to anything in C# 4 yet. However, there have been hints about what they've
Anonymous
February 07, 2008
C# 4 slicing index notation ala python. var l = Enumerator.Range(1,100).ToArray(); for (var i in l[25:75]) Console.WriteLine(i); //prints ints from 25 to 75 including pythons start/end defaults l[:5];// the first 5 l[96:];//the last 5
Anonymous
February 08, 2008
Heres an off topic question. Recently the C++ team has released a beta of the upcoming MFC changes. Included in this are some nice UI changes that incorrorate alot of the Office/Visual Studio UI enchancements, such as docking, theming, Ribbon Bar, and more... Is there any plans to update WInForms to include these enchancements?
Anonymous
February 09, 2008
I proposed the idea on this topic a while back of using the type inferencing available through the "var" keyword to support reflective, late-bound lookup. Now, we all know that "var" isn't a type but the compiler could allow references obtained that way to bind late. For example, to reflect and find SomeUnknownMethod() at runtime, this would NOT be allowed: MyClass m = GetMyClass(); // illegal b/c MyClass is not known // to implement SomeUnknownMethod m.SomeUnknownMethod(); However, this would compile: var x = GetMyClass(); // legal b/c var allows for // late binding on references x.SomeUnknownMethod(); Being marked as "var", the reference x gets special dynamic treatment. I like this better than anything else I've seen here because:

It is a simple extension of the type inference concept that already exists.
It would force certain C# developers who might use "var" a little too often (if you known what I mean) to avoid the possible dynamic side effects introduced by this enhancement. --Kevin

Anonymous
February 10, 2008
What a terrible idea. It seems like we keep taking this circular path with languages that takes us back to exactly where we started, but with new toys. So much for type safety and code that can be analyzed by external tools for issues
Anonymous
February 11, 2008
@John Rusk, I can certainly see your point about dynamic calls being easier if the syntax is available in the language. However, I am VERY concerned about the long term effect this feature will have on the language. Much has been made in the past about new language features starting out with "minus 100 points" (http://blogs.msdn.com/ericgu/archive/2004/01/12/57985.aspx), and that they must prove that they have significant enough value that they can overcome the inherent negative impact that adding any new feature has on a language. I am far from convinced that this feature meets that criteria. I would like to hear from the language team what they think about that issue, as well as other people in the community.
Anonymous
February 11, 2008
The comment has been removed
Anonymous
February 11, 2008
I think the best approach is using a special operator like .. And how about constructing objects dynamically? Like: Type t = typeof(SomeClass); object dynamicObj = new t(10); -or- Type t = typeof(SomeClass); object dynamicObj = dynamic t(10); And then the runtime call the correct constructor and create the object.
Anonymous
February 11, 2008
So, the proposed block syntax might have two motivations that I can think of:
It's for implementation/performance reasons. Some context will be built up and held within the block in order to perform the dynamic lookup. This can then be discarded when the block is exited. If this is the motivation then I think the syntax gets in the way of usage and any performance considerations should be 'hidden' and not emphasised in a block syntax.
It's to deliberately demarcate dynamic lookup code so that programmers can see it being used. If this is the case then I can see the point. But at the same time I think it might make the facility too annoying to use by introducing variable scoping issues and mixing up static and dynamic lookup. My opinion is that dynamic lookup happens when you call a method/property/indexer on an object, so the syntax should be at that level, not at a block level. Thanks for publishing this blog post by the way, even if it has resulted in us all annoying you with our impractical suggestions. :)
Anonymous
February 11, 2008
I welcome the prospect of being able to (easily) make late bound calls using C# and anything that makes COM/Office interop easier, without going the whole hog of introducing optional parameters into the languages, is particularly welcome in my opinion. Although I do have some sympathy with those who would prefer C# to remain wholly static, I think the time has now come where the language can afford to relax a bit on this and that, if it doesn't, it may find itself losing popularity to languages which do. Personally, I'd like to see the day where I can do all my programming in C# - it may be asking too much but that would be my ideal. As this is going to be a .NET wide facility, I suppose the implementation which the C# team eventually settles on will need to be consistent with the DLR infrastructure but, if there's any choice in the matter, I agree with most of the other posters that having a dynamic block is not the best solution. In fact if, as Charlie said, there is no reason to consider a dynamic variable as anything special, I see no reason to introduce a new 'dynamic' keyword or attribute at all! All you need is a special operator for dynamic lookup and a rule which says that this can be applied to any variable or expression of type System.Object but not any other type. The compiler won't then check the call for correctness (except as potentially valid C#) and the onus will be on the programmer to get it right. As for the choice of special operator, I'm not keen on .. which looks like a range operator and I don't like -> either because that's currently used only for unsafe code. I'd prefer something different such as: objectExp~SomeMethod(); which would stand out more. This will still make Office interop code very easy to write (with the help of Intellisense)and you won't have to decorate all your COM object variables with 'dynamic' or include them in a dynamic block. It would in any case be difficult to use a common word such as dynamic as a new keyword (except perhaps as a contextual one) because it's sure to break some existing code which hitherto the C# team have been almost paranoid to avoid doing.
Anonymous
February 11, 2008
That's pretty cool, Alan. I wish I had thought of that. ;) Now, what happens when I do this? =================================== string X = "kilroy"; objRef~SomeProp.SomeMethod( X ); =================================== What about the reference to string X? Should the dynamic call be able to reach into the local, statically defined context like this? What about? =================================== Customer Z; objRef~SomeProp.SomeMethod( out Z ); =================================== -- Kevin
Anonymous
February 12, 2008
That's certainly what I'd like Kevin and I can't really see why it shouldn't be possible - the call, in it's entirety, will either succeed or fail at runtime. However, given that the C# team have come up with this idea of a dynamic block rather than the (more obvious) ones of dynamic variables or a dynamic lookup operator, it makes me wonder whether we may be missing something here and that there are problems mixing dynamic and static contexts. You can't really tell anything from the current implementation of late bound calls in VB.Net. If you set Option Strict to Off, the two examples you gave work fine but it appears that the whole context of the program is treated as dynamic. For example, even the following statement is allowed by the compiler: Dim i As Integer = "Hello" which is certainly not something we'd want to see in C#, however dynamic lookup is implemented.
Anonymous
February 12, 2008
If this is implemented, I am highly in favor of: dynamic object myObject = ... instead of the dynamic block. Having said, that I'm not really sure that we really need this at all. Static binding is your friend. As for the Java suggestion earlier... the reason I like C# is that it IS like Java, except that it works correctly and isn't horribly slow (which is largely due to the fact that it is NOT dynamic.)
Anonymous
February 13, 2008
The comment has been removed
Anonymous
February 13, 2008
Maybe require that all uses of the "dynamic call" operator (whether it is written as .., -> or ~) appear only WITHIN UNSAFE BLOCKS? After all, these are unsafe operations....
Anonymous
February 14, 2008
@Octavio Your suggestion makes a certain amount of sense, however, in C# unsafe blocks are used to indicate code that might produce unverifiable IL (as opposed to normal C# code which is guaranteed by the compiler to produce verifiable IL). So there are two different meanings for unsafe here, and they don't really mesh well with each other.
Anonymous
February 14, 2008
Σε συνέχεια του post περί late binding , ας δούμε πως μπορούμε να πετύχουμε σωστό late binding, χωρίς
Anonymous
February 14, 2008
The syntax dynamic object myDynamicObject = GetDynamicObject(); is the best suggestion so far.

The dynamic { ... } code block is too coarse as many people have already said.
A dynamic object presumably would not have both, static and dynamic methods so why introduce a special dynamic call operator such as "..", "~", etc.? Can anyone ever envision calling myDynamicObject.GetName() (a static method with one dot invocation) and then calling a myDynamicObject..SomeUnknownMethod() (the two dots notation) on the same object? Not likely, dynamic objects will almost certainly be composed entirely of dynamic methods. If that is the case, then why introduce the extra calling syntax clutter? There is a certain aesthetic beauty and elegance to consistency that ought to be maintained. Ask C++ developers how they liked having two different calling operators in their language. This ex-C++ developer says NO.
Those who think they can pre-declare interfaces on dynamic objects are forgetting that the objects are dynamic... ie. some languages allow methods to be added or removed from such objects at runtime. Indeed some languages allow objects to be fleshed out with method sets determined completely by runtime conditions. Take a look at look at the vast codebase of oo Javascript in use in the wild. Take a lot at the latest Ajax or JSON libraries. The whole point of Dynamic objects is that you can't predict or know ahead of time what methods or properties they will possess at any point before you acquire a runtime reference to them. As well, who in the heck would want to pre-declare some of the Office Automation objects before they use them? Have you seen some of the interfaces on these things? There are hundreds if not thousands of methods whose signatures you'd ahve to predefine. Besides, why would you even do this instead of using the COM type library that ships with the interop assembly? And this is describing a scenario of using COM objects who actually HAVE a statically defined interface to begin with. Forget about all that if you are actually dealing with objects coming from a dynamic language. Where pre-declared interfaces may be of use if one is using only a limited set of functionality on an incoming dynamic objects and wants to impose a statically bound interface on the interesting subset of methods. In this case, the interface acts as an intellisense-friendly facade to the underlying object which may in fact possess a much larger set of uninteresting or undesired methods. In this case, the runtime would have to enforce that there exists a way to coerce the underlying object to map onto the pre-declared interface. ie. All the pre-declared method signatures would have to have matching methods in the underlying object.
For those who are concerned about C# straying from its statically typed roots, no need to worry. No one here is saying that C# will be equipped to MAKE dynamic objects, but rather CONSUME them... and given that, then what is the objection? Why not open the door of possibilities to using all that code in the wild from within C#? Doesn't seem to detract from the attractiveness and utility of the language one bit. If you are still ideologically opposed to this, then the answer is simple: just make sure you never use libraries that are created in dynamic .Net languages (like Python or Ruby, etc.) Simple. We can all live in peace and harmony. In summary, my vote is for: dynamic object myDynamicObject = GetDynamicObject(); NO to dynamic { ... } code blocks, and NO to special dynamic call operators.

Anonymous
February 14, 2008
I would like the .. operator :-)
Anonymous
February 15, 2008
The comment has been removed
Anonymous
February 15, 2008
C# nelle future versioni di Visual Studio
Anonymous
February 16, 2008
Good information. Thank you. http://blog.Laksha.net
Anonymous
February 17, 2008
I would like to request the ability to coerce anonymous types into interfaces. So you can use LINQ for more than just databinding, so you can actually return strongly typed objects from any method without having to declare all of your types. Something like: public IExample[] GetExamples() { return (from e in examples select new IExample { P1 = e.Text }).ToArray(); } If you had methods on the interface than either create an empty method that returns the default value or allow delegates to be attached to the interface at creation time. maybe like: public IExample GetExample() { return new IExample { P1=this.Text, Run=this.RunExample }; } private void RunExample() { //do some stuff here. }
Anonymous
February 19, 2008
I think they need dynamic types in C# because the other dynamic languages dont use the C style syntax(at least not popular/useable ones) Second- It sure would be nice to have some more dynamic behavior with Generic types even polymorphism. And while I am at it throw in interface property attributes recognizeable as [Column] in Linq so you can use linq more dynamically/genericly as well.
Anonymous
February 19, 2008
The comment has been removed
Anonymous
February 19, 2008
I feel like I haven't had a chance to respond to comments in the last few days. As a result, I wanted to take a moment to assure everyone that we have been reading these comments, and that the design team has seen them, and has taken them seriously. It's perhaps worth reminding everyone that our goals here are twofold:

We went to be open and share our plans with the community.
We want to absorb your feedback and make good use of it. Looking at those two yardsticks as a way of measuring our progress, I would say that this has been a very successful process for us. It is great to be able to share ideas with the community, and your input has been very useful to us. We are not yet ready to say exactly how these comments have influenced the team, but I can assure you that they have influenced us, and they have had a positive impact on our planning process. Thank you so much for the comments you have submitted so far.

Anonymous
February 19, 2008
@Charlie, Why do I get the feeling that you are trying to prepare us for the fact that even after all of the feedback you have gotten from this blog, you still decided to go with your original dynamic block idea?
Anonymous
February 19, 2008
@Simon, "A dynamic object presumably would not have both, static and dynamic methods" For objects instantiated by a dynamic library, you are probably right. But imagine if you want to use dynamic method invocation to invoke a method declared as private in a statically defined class using reflection? In that case, it is certainly possible to imagine cases where you might want to make both static and dynamic calls on the same object; and at that point, a dynamic call operator might make sense. Not necessarily voting in favor of it, just pointing out that it is not as ridiculous as you made it sound.
Anonymous
February 20, 2008
I've read all comments, and I vote for this: dynamic object a = GetDynamicObject(); a..SomeMethod(x); a.ToString(); a..SomeProperty = "123";
Anonymous
February 20, 2008
The comment has been removed
Anonymous
February 20, 2008
Love the idea, but the dynamic keyword is indeed a bit limiting - since it's most logically tied to an instance, the "this is dynamic" magic should be tied to an instance/variable instead of a scope. What Andrew Davey said about IQuackFu from Boo is good, but IMHO the easier and more powerful approach is also from Boo - the "duck" type (http://boo.codehaus.org/Duck+Typing), which you declare a variable as and then the compiler knows that it's a dynamic type. Then you can pass "duck" objects around without having to deal with the dynamic keyword. Then tools (like the debugger) also know the instance is "special" and interact with it as such. Instead of duck, though, perhaps System.Dynamic or similar, since it probably makes sense to have it represented in the BCL since you may want to string some static methods off of the type (System.Dynamic) that helps runtime interaction with such objects, like easier-than-reflection inspection of methods/properties/etc, or Ruby's "reopen this instance/the entire class" support, or whatever. Back to the IQuackFu/IDynamicDispatch idea - as long as the compiler implements it as a fallback so "regular" method calls that are implemented by the type keep working as normal, then that's fine with me - it's effectively Ruby's method_missing at that point, which would be useful to have in some situations. The combination of the two gives you both fully-dynamic instances (with the duck typing aka System.Dynamic) and partially-static, partially-dynamic (with IQuackFu aka IDynamicDispatcher), so you can choose what works best for you. The idea is great, but I really fear the dispatch keyword+scope approach would make it less useful (read: more cumbersome) to interact with.
Anonymous
February 20, 2008
Ok, so it's just in the planning stages, but you may have run across Charlie Calvert's recent blog post
Anonymous
February 20, 2008
VB6 and VB.Net developers know this feature: "Late Binding" and C# developers have often ridiculed
Anonymous
February 21, 2008
Ok, so it's just in the planning stages, but you may have run across Charlie Calvert's recent
Anonymous
February 21, 2008
The comment has been removed
Anonymous
February 21, 2008
The comment has been removed
Anonymous
February 21, 2008
Charlie, I can certainly appreciate that you cannot promise any specific outcome. I can also appreciate that sharing these plans with the community does not mean that this is our chance to design the feature; the inmates cannot be allowed to run the asylum, as it were. I suppose I am simply wary because I have seen this too many times before. A team decides on a new feature it wants to implement, debates it internally, and comes up with the approach they want to use, and then exposes that approach to the community. At which point, there is a very strong community response toward another approach, or away from the feature entirely; but at that point it is too late, because the team has already convinced itself of "the right way" to do it. I can only hope that will not be the case here; what prompted my comment was that, although the tone of your comment was positive, I felt it ringing with an undertone of "but don't get your hopes up." Which, as you say, you had already stated earlier. Please don't let my pessimism or cynicism dissuade you from continuing the series. I would love to see what else the team is coming up with, even if I don't get to do it the way I want :)
Anonymous
February 22, 2008
This will become very handy for mine MVC RouteProviders in SQL en XML. Now I have to use dynamic compiling to create a structure to pass the parameters which are normally build as new { Controller="HomeController", Action="View", ID="ObjectID" } and passed to the MvcRouteHandler as an 'Object' The compiler creates a anonymous typed when compiling, but you can't yet create really dynamic structures. Sounds very promising..
Anonymous
February 22, 2008
Regarding type safety, I believe it's good to limit the dynamic keyword to a variable declaration. But also, it shouldn't be limited to 'dynamic object' we should be able to declare dynamic variables of any interface type. Even on dynamic objects programmers know pretty well what to expect, but this way the compiler catches typos. Example: interface IMyCOMAPI { void MyMethod(); int CalculateSomething(int someParam); } dynamic IMyCOMAPI myO = GetDynamicObjectFromSomewhere(); myO.MyMethod(); //just fine! myO.NyMethod(); // hey!! a typo! I think that resolves many issues elegantly.
Anonymous
February 22, 2008
Perhaps I should add ... 'GetDynamicObjectFromSomewhere()' would be the only place where a binding related runtime excpetion would occur. Optimizations, like caching reflection objects, would be done in an anonymous class that implements the interface and actually takes care of the binding. This way optimal use of a dynamic object can be planned by developers. Like initializing dynamic objects once, automatically binding upon construction, sotring it in a field, and invoke methods efficiently throughout the field's scope thanks to this. Well, my 2 cents ... Sebastian
Anonymous
February 22, 2008
Sorry, on a roll ... Another benefit I just noticed (although I can see some criticism) of this method is that by using itnerfaces, we can pass dynamic objects to all consumers of the interface, even those that were not initially devised for dynamic object consumption (assuming we use a predefined and already popular interface). This is were the possible criticisms arise, because although everything except binding can can be statically checked, the potential remains for misuse. I however think compilers are there to protect us from typos and similar mistakes, not from not understanding and API. Sebastian
Anonymous
February 22, 2008
Hey, me again heh, It's later in the day and I realized the following: There doesn't need to be a 'dynamic' keyword at all. All that's really needed is a way to wrap dynamic objects in anonymous interface implementations. For COM: A dynamically bound COM object could be created by a generic (as in generically typed) COM object factory. Example: interface IMyCOMAPI { void MyMethod(); int CalculateSomething(int someParam); } IMyCOMAPI myO = new DynamicCOMObject<IMyCOMAPI>(guid /or pointer to object or whatever/); myO.MyMethod(); //just fine! myO.NyMethod(); // hey!! a typo! compiler detected as always For the DLR and its dynamic objects: The DLR can return dynamic objects, like it normally does, which will be dynamically bound to an interface at the developer's will. Something like perhaps: object o = dlrObject.GetDynamicObject(); IMyCOMAPI myO = DLR.BindObject<IMyCOMAPI>(o); // no special operator ever needed. Perhaps I'm missing something? some feature? It doesn't seem like an operator is needed at all. Sebastian
Anonymous
February 23, 2008
The comment has been removed
Anonymous
February 25, 2008
This Dynamic are direct runaway from strong typification, but the last concept is foundation of C#. We need other higher-order approach for this problem. I think, we must to proceed from strong typification as basic. Simple Example (WSH inline): // There is begin of example-------------------- using System.Scripting.WSH; ... string StringBox = string.Empty; // in the next scope provides other (read: dynamic) rules script (WScript ws) { ... exl = ws.CreateObject("Excel.Application"); StringBox = exl.SomeMethod(); ... } // There is end of example---------------------- P.S. Sorry for my English. I am russian developer.
Anonymous
March 05, 2008
It would be nice to have more flexibel operator overloading, like arbitrary operators instead of a fixed set.
Anonymous
March 05, 2008
Aren't you basically just adding VB-style late binding to C#?
Anonymous
March 05, 2008
I wanted to thank everyone again for their great suggestions. The positive and thoughtful nature of your feedback is greatly appreciated. We are now certain enough of our plans to confirm that your comments will lead to significant changes in the way we implement this feature. We are looking forward to the time when we can get back to you with concrete information on the exact nature of our new plans.
Anonymous
March 06, 2008
The success of the first Future Focus post has made it easy for us to continue sharing our plans for
Anonymous
March 07, 2008
Well since we're throwing random ideas about I'd like to join: There's already extension methods but they do seem a bit limited in some ways so what about 'extended scope': !dynamic,with(GetDynamicObject()) { .SomeMethod(); .someString = "value"; [0] = 25; } The only new keyword here is "!" for marking beginning of extended scope that brings in different entension delimited by ",". Now of course bringing a ton of weird extensions would likely always mean throwing away intellisense atleast in v1 but maybe if it proves popular some support could be added to add some intellisense features even in extended scopes.
Anonymous
March 07, 2008
"So much for type safety and code that can be analyzed by external tools for issues" That's a very good point. The "infered duck typing" suggested above seems like a good idea without going overboard like the extendable scope thing which practically would mean a language in language.
Anonymous
March 07, 2008
The comment has been removed
Anonymous
March 08, 2008
Will C# have syntax extension macros? It is a well known practice in Lisp world. In .NET there are some successful implementations like ones in BOO ans Nemerle. Of course what BOO and Nemerle provides are different than Lisp ones; yet they are very usefull tools. Those syntax extensions are type-safe. Is there any plan to implement that in C# 3.5 or 4.0? (There are many other features that I like to have in my C# toolbox. Yet this one is the most important one to me and many of other features that I like to have; can be implemented easily via a stable well-designed macro system.)
Anonymous
March 08, 2008
Kudos to Charlie Calvert for opening up the discussion of C# future features to the public (sure would
Anonymous
March 10, 2008
Hopefully the dynamics includes also the following syntax: string method = "MyMethod"; object.method(); or directly: object."MyMethod"(); or some other new syntax to accomplish the above. Otherwise it's not really dynamic, i.e. you have to know the method names when you write the code. Runtime resolution of the method name should be possible like with reflection.
Anonymous
March 10, 2008
The comment has been removed
Anonymous
March 11, 2008
The comment has been removed
Anonymous
March 12, 2008
I certainly hope I will be able to turn this "feature" off for our in-house developers.
Anonymous
March 12, 2008
Sorry for my English! I think that one of ways for creating dynamic classes is like this: using System; class A { static void Main() { string s="class B { " +" public void DisplayMessage() { " +" Console.WriteLine("Hello from dynamic C#.");" +" }" +"}"; dynamic B obj = new B(s); obj.DisplayMessage(); } } Reason is: 1)Possibilities to create varible s dynamicaly and then create class dynamicaly. 2)Posibilities to read existing cs files and after modification create new classes dynamicaly.
Anonymous
March 15, 2008
Its not big deal. Well its good to hear that "dynamic" would be a built-in feature for C# developers (too). I have already done this sort of work by writing my own classes using reflection. And really it worked some how the same way.
Anonymous
March 16, 2008
Please don't turn C# into a dynamically typed language...
Anonymous
March 16, 2008
Let me add my name to long list above opposing this. I use C# primarily because I can depend on it being type safe by default. Dynamic lookup will destroy this. Can I disable the feature? In a team environment, I would hate to think about manually checking for this abusive code. Further, as already mentioned, this type of functionality belongs in a library, not at the language level... Please reconsider this. Whether you realize it or not, you're destroying C#.
Anonymous
March 17, 2008
The comment has been removed
Anonymous
March 18, 2008
I really, really like Halo_Four's solution. It keeps C# entirely in the static typing world, doesn't require any new language features, but still gives us everything we are looking for in terms of making dynamic calls, both reflecting to the internals of an object and calls on dynamic objects. The only downside is that it requires an additional interface to be defined; but that can be a good thing, because that interface now exists in the metadata, and can be used like other type definitions. Charlie, what do you think?
Anonymous
March 21, 2008
That's a very interesting article, and I can easily see places where dynamic typing might be useful. Have you guys looked into how ActionScript 3.0 deals with dynamic typing? I think they came up with a really elegant solution for dynamic typing.
Anonymous
March 26, 2008
The comment has been removed
Anonymous
March 30, 2008
That's a very interesting article, and I can easily see places where dynamic typing might be useful. Have you guys looked into how ActionScript 3.0 deals with dynamic typing?
Anonymous
April 04, 2008
Hi, thank you for sharing you current plans. I've read through the comments, and saw some pretty good ones, yet also very bad ones. I don't think it is bad to get feature from other languages into the existing one, unless it influences anybody who does not want to use it. So:

Do not mix it with var keyword. It is strongly typed, safe type. Any attempt to use it for dynamics would make all the people against var having actually truth.
Do not mix it with interafaces. Interfaces are what they are for and taking them into dynamics makes the interface concept more complex, and turns dynamics into generics. Dynamics should just turn your calls into reflection ones. That's it, no interfaces. If anybody wants to wrap the dynamic object into class, why not.
Do not restrict the base dynamic object to the type of object. Dynamics should not force you to cast to object if you know the base type.
Make sure you can get anything you would need to do using reflection. Access static members, generic ones, etc. I agree with the suggestion of calling constructor dynamically as well, instead of kind of GetDynamicObject() method.
As far as the scope is mentioned... it seems you could fulfill wishes of peoply only if you allowed both object and block scope. I don't think any keyword is needed at all, if you make the member accessor different. From what I've seen I like most the .. one, though I still do not feel 100% comfortable with this due to the .ctor and ..ctor methods. If you use special accessor, then no scope is needed, the behavior would be absolutely predictable (no attemps to early bind if the accessor is used) and the code would be much more clear.
Do not allow dynamics to access non-public members. This is not the purpose of dynamics and developers should still have to use reflection for this. If you make indexed properties also accessible, I would welcome them in static C# as well. ;) I believe you will get this as right as you want and wish you to ship it as good as you wanted. Jan

Anonymous
April 07, 2008
Here are some thoughts I have regarding the ideas of dynamic method calls in C#: http://www.justnbusiness.com/Blogs/CSharp_Optionally_Dynamic_Calls.aspx
Anonymous
April 13, 2008
The comment has been removed
Anonymous
April 16, 2008
That's pretty cool, however, have you though about making it so that the "dynamic" isn't a scope keyword but instead applies to a variable? Like this: dynamic object myDynamicObject = GetDynamicObject(); myDynamicObject.SomeMethod(); // call a method myDynamicObject.someString = "value"; // Set a field myDynamicObject[0] = 25; // Access an indexer -tony V.
Anonymous
April 23, 2008
exellent,plz include me in the program. iwill be glade if you sent me more regarding C#. nice time
Anonymous
April 23, 2008
Allow me to refine my suggestion a tad. I wanted to clarify what I had intended as well as correct a minor problem with my sample. As mentioned I don't think the semantics of C# should be changed to support dynamic dispatch. I would prefer a more intentional method of interacting with dynamic types. My suggestion is to not modify the language at all. Instead, provide the facility through the DLR or through remoting to construct a proxy for an object given a defined interface. That proxy would handle the details of how the actual members are called and would handle .NET objects as well as COM objects to the best of it's ability. public interface ICalculator { int Add(int x, int y); } static void Main() { object o = GetCalculatorFromSomewhere(); ICalculator calc = o.TryDynamicCast<ICalculator>(); if(calc != null) { int value = calc.Add(2, 3); } } Extension methods would be provided through the DLR that would construct the proxy. Suggestion methods as as follows: // Try to create proxy for interface T, throw exception if object cannot conform to interface public static T DynamicCast<T>(this object o); // Try to create proxy for interface T, return default(T) if object cannot conform to interface public static T TryDynamicCast<T>(this object o); // Return whether or not object can conform to interface public static bool CanDynamicCast<T>(this object o); Or the methods could just be normal static methods. What I like about this suggestion is that:

It does not change any semantics of C# at all. Only normal run-of-the-mill interfaces are used and C# remains completely type-safe.
No language enhancements are required, so this functionality could be built right now on the existing C# 2.0/3.0 languages, depending on whether or not extension methods are used.
You gain the requested functionality of allowing "duck-typing" in C# automatically.
You only need to do this work in one place and any CLI-compatible language can immediately benefit since you are only using the existing facilities within the framework. This will increase the pervasiveness of support for dynamic dispatch. The disadvantage is that this method is more verbose since you have to declare the interface, but it is my opinion that requiring the intent to be declared keeps the language appropriately type-safe. Also, it would prevent a dynamic object from being partially conformant to the requested interface. I can envision problems with dynamic dispatch using a "dynamic" keyword where the consumer needs to make several calls and some of those calls exist and others do. The code necessary to compensate could be complicated. Anyway, I hope you consider my recommendation. In fact, I hope you can see it at all buried in the slew of comments on this article. Thanks, Justin

Anonymous
May 06, 2008
A number of people have suggested that for the dynamic case, the member access syntax should be made deliberately different to the usual dot syntax. There is a lot to be said for this. I was thinking that maybe to underline the fact that the method name is really no more solid than a string, you could have the syntax mimic accessing items in a dictionary. So instead of: d.LetThereBeLight(); It would be: d"LetThereBeLight"; It's not a lot more typing but it properly flags up the dynamic nature of what the code is doing. But then I realised we can already do that today. Just make a class that has an indexer, which returns a delegate that takes a variable number of parameters: http://incrediblejourneysintotheknown.blogspot.com/2008/05/dynamicobject-wrapper.html
Anonymous
May 07, 2008
The comment has been removed
Anonymous
May 08, 2008
As a follow up to my previous comment. Adding duck typing would be also trivial and natural having a class like class MyString {int Length{get;}} and method int GetLength(string s){return s.Length;} We might duck-type MyString to string and pass it to GetLength method like this: int length = GetLength(new MyString().dynamic);
Anonymous
May 11, 2008
Microsoft: "C# is strongly typed ... Just kidding!"
Anonymous
May 18, 2008
I choose to program in C# because I prefer statically typed languages to dynamic ones, so I'm really unsure about this new feature. If it was to be included in C# 4.0, I'd add my vote for the object."method"() syntax. I think that the double-dot operator could easily be confused with a single dot and dynamic blocks are a bit overkill. The object."method"() syntax makes it obvious that, like someone said, the method name is only a string (and will as such be resolved dynamically, unless it can be resolved statically). Operators might be accessed using a syntax such as object."+"(param1, param2) and generics like object."method"<Type>(param1). Although this is, as I see it, the best syntax for this feature, I'm still unsure about if it should be added as I fear it will be abused.
Anonymous
May 22, 2008
Please do not implement this in C#. You have already alienated me enough with LINQ and anonymous types. (I like LINQ to SQL, I just don't like the C# query syntax because I don't believe it belongs in a language like C#.) There are reasons I switched from VB/VBA/VBScript/JavaScript to C# many years ago and this is one of them. I don't want a language that promotes loose typing and laziness. If you really wanted to fix the problem with Office Interop, you'd make real managed classes instead of relying on tlbimp-generated code as a base. It can't be that hard cause I have already done it for Outlook and I'm just one guy.
Anonymous
May 30, 2008
The comment has been removed
Anonymous
June 27, 2008
Lexapro 20mg. Lexapro. Lexapro and side effects.
Anonymous
July 07, 2008
That's great,I am glad to heard that. I am sure the next version of Visual Studio will be perfect.
Anonymous
July 07, 2008
That's great,I am glad to heard that. I am sure the next version of Visual Studio will be perfect.
Anonymous
July 10, 2008
Just to add another idea: couldn't this perhaps be done partly at the library level instead of adding another keyword to the language, something along the lines of Nullable<T>? I mean something like: Dynamic d = new Dynamic(GetDynamicObject()); d.SomeMethod(); where the compiler translates d.SomeMethod() to something like d.Call("SomeMethod").
Anonymous
July 10, 2008
I’ve not watched many Channel 9 videos over the last few months, but every now and then some really good
Anonymous
July 11, 2008
The syntax outlined can become very clunky and dangerous because usually such block of code will contain mix of late binding as well as early binding (assuming not all programmers are perfect). I would suggest you invent a different calling notation other than "." to make dynamic call. For example, may be we can define ".." as replacement for "." for dynamic calls. So whenever ".." follows identifier it would specify that next operator/method should be invoked dynamically. object myDynamicObject = GetDynamicObject(); myDynamicObject..SomeMethod(); myDynamicObject..someString = "value"; myDynamicObject..[0] = 25;
Anonymous
July 23, 2008
I don't normally make these kinds of posts but I'm taken aback by this. We had this feature for years in VB. We called it Option Strict Off. We could set it at the Project level, the File level, and through use of partial classes even at the method level. And for years we caught nothing but grief from the development community over it. And now that you want to "shoehorn" it into C# it's a "cool new feature" that all the jocks and cheerleaders use while drinking Pepsi. I don't get it. This reversal of prejudices lobbed at VB for generations all because it's applied to the trendy languages confuses and insults me. This was a horrible, not-best-practice, bad-coding-practice feature 10 years ago and the C# community should have to live with and accept this as a vile black mark on its perfect language for all time like I've had to (no matter how useful it is). You can't make this "good" now that you want it. First Edit & Continue and now this. Can't the C# universe stick to its elitism and stop ripping off VB.NET? Nevermind that it encroaches on the VB market niche. The end result of this language bloat will be the same fate that befell VB and C++ before it: A monstrous all inclusive language anchored to legacy code with a subcontinent of diverse development groups that make it impossible to identify a target audience.
Anonymous
July 23, 2008
One of the common feature requests for IronPython is to support static compilation. While the feature
Anonymous
July 26, 2008
The comment has been removed
Anonymous
July 26, 2008
The comment has been removed
Anonymous
August 02, 2008
I want to cast my vote for the ! calling syntax instead of blocks: myObject!Method(); myObject!"Method"(); myObject!Property = 0; myObject!"Property" = 0; It fits nicely with anyone who used vb6 back in the day!
Anonymous
August 19, 2008
The comment has been removed
Anonymous
August 20, 2008
I agree with the comments made by Thomas Krause, Daniel Earwicker, and Simon Watfa: dynamic object MyOfficeObject (for syntax, instead of the dynamic code block), I dislike the idea of a separate member-access operator (C++ is a mess of operators), and that C# should only be a consumer of dynamic objects (and remain static). Charlie, you said declaring a whole object as dynamic is too broad. What you mean by that? What specific scenarios have you come up with that demonstrate this problem? If you always first try to do static lookups, and only perform dynamic lookups when that fails (and only on variables marked as dynamic), what problem remains? C# should remain static, but by using extension methods, it can take on the appearance of being somewhat dynamic, in the sense that we're "adding" members to types, at least for code that references the appropriate library and imports the right namespace. It's not flexible to the point of being able to add and remove members at runtime, but I can't think of a situation that more would ever provide significantly more value (which isn't to say there is none). My only request here is that C# incorporate general type extensions instead of limiting them to extension methods. Some kind of duck typing or structural typing would be very useful, and wouldn't require dynamic typing in the language, but as others have stated, this doesn't address the needs of accessing types whose members change during runtime. While I don't want to be able to modify classes during runtime by adding or removing members, it would be VERY useful (for aspect oriented programming) to allow dynamic dispatch for pervasive method interception. Keep the interface of a class the same, but allow calls to be rerouted to different bits of logic. .NET has a TypeForwardedTo attribute, but this would be a more powerful, fine-grained (member-level) approach. Perhaps extending the language, or providing new reflection-like classes for intercepting individual members (properties and methods), all members of a type, etc.
Anonymous
August 20, 2008
It should definitly try to resolve statically first, because we want to have code completion available inside dynamic blocks when it is available and just have nothing when its not... :)
Anonymous
August 20, 2008
The comment has been removed
Anonymous
August 20, 2008
The comment has been removed
Anonymous
August 20, 2008
Back to the issue at hand, i like the dynamic block idea, it makes it clear where the dynamic calling can occurr, which means its easy to spot, and easy to perform QA on to ensure its not being overused. The only problem i see is allowing a mix of dynamic and static calls, as far as im concerned if any calls can be statically linked and are in a dynamic block the compiler should fail. Sure this might mean a little more work because you may have to have more dynamic blocks than you would previous need, but this will stop overuse and keep C# to a static language as we all like it....
Anonymous
September 02, 2008
Hi, how about adding dynamic interfaces to the language? Interfaces are perfect for static typesafe dynamic coding and fit into the tool feature set quite well too. What do i mean? Given an object obj that may or may not be dynamic, anonymous, concrete whatever. I can declare a variable x such that x exposes a public method/property/field through a dynamic interface. for instance var x:{void Walk();}. With this type of feature, one can cast anything to anything as long as theor interfaces <dynamic or otherwise> match. At assignment time (ie- x = obj ;) the runtime or vs.net can check to ensure that obj exposes the interface as well. Its seems like an easy way to give us a TON of functionality; including finally able to interact with anonymous types outside of a method without having to have a known type to copy it into.
Anonymous
September 09, 2008
A post by Jeremy Miller caught my eye this morning in regards to extension methods in Javascript . While
Anonymous
October 18, 2008
The comment has been removed
Anonymous
October 22, 2008
The comment has been removed
Anonymous
October 22, 2008
And about the syntax, I think '..' is the best option, allowing of course the dynamic block also.
Anonymous
October 27, 2008
PDC 2008 arrives today, and that means that I am finally able to talk publicly about some of the things
Anonymous
October 27, 2008
Muitas novidades aconteceram e vem acontecendo no mundo de tecnologia e desenvolvimento. Principalmente
Anonymous
November 08, 2008
C# 4.0 Dynamic Lookup I really like the way the C# team tackled bring dynamic programming to the language
Anonymous
November 23, 2008
The comment has been removed
Anonymous
November 24, 2008
Many .NET users have no doubt run into issues using math functions with generics. Let's say you want
Anonymous
January 18, 2009
IronPython in action, embedding dell'IronPython engine
Anonymous
February 15, 2009
Lors des techdays 2009, je suis allé voir la session “ Programmation dynamique ” de Mitsu Furuta et Simon
Anonymous
May 19, 2009
An attentive reader pointed me at this long thread on a third-party forum where some people are musing
Anonymous
June 11, 2009
ヒマだょ…誰かかまってぉ…会って遊んだりできる人募集！とりあえずメール下さい☆　uau-love@docomo.ne.jp
Anonymous
June 12, 2009
話題の小向美奈子ストリップを隠し撮り！入念なボディチェックをすり抜けて超小型カメラで撮影した神動画がアップ中！期間限定配信の衝撃的映像を見逃すな
Anonymous
June 12, 2009
話題の小向美奈子ストリップを隠し撮り！入念なボディチェックをすり抜けて超小型カメラで撮影した神動画がアップ中！期間限定配信の衝撃的映像を見逃すな
Anonymous
June 13, 2009
カワイイ子ほど家出してみたくなるようです。家出掲示板でそのような子と出会ってみませんか？彼女たちは夕食をおごってあげるだけでお礼にHなご奉仕をしてくれちゃったりします
Anonymous
June 14, 2009
あなたは右脳派？もしくは左脳派？隠されたあなたの性格分析が3分で出来ちゃう診断サイトの決定版！合コンや話のネタにも使える右脳左脳チェッカーを試してみよう
Anonymous
June 15, 2009
The comment has been removed
Anonymous
June 16, 2009
セレブ達は一般の人達とは接する機会もなく、その出会う唯一の場所が「逆援助倶楽部」です。男性はお金、女性はSEXを要求する場合が多いようです。これは女性に圧倒的な財力があるから成り立つことの出来る関係ではないでしょうか？
Anonymous
June 17, 2009
貴方のオ○ニーライフのお手伝い、救援部でHな見せたがり女性からエロ写メ、ムービーをゲットしよう！近所の女の子なら実際に合ってHな事ができちゃうかも！？夏に向けて開放的になっている女の子と遊んじゃおう
Anonymous
June 18, 2009
まったぁ〜りしたデートがしたいです☆結構いつでもヒマしてます♪　m-g-j@docomo.ne.jp　年齢と名前くらいは入れてくれるとメール返信しやすいかも…

Partager via