Re: C# 3.0

I'll see if I can do a bit better.

The way that the query below works is that the compiler writes a type
like:

public class ProductsGenerated0x333 
{
        public string ProductName;
        public string Category;                 // might be an enum, or
something else...
        public Decimal Price;
}

And what you get as the result of the query expression is:

Sequence<ProductsGenerated0x333>

Where sequence<T> is somewhat like IEnumerable<T> (I forget the
differences).

So you can easily write something like:

foreach (var info in productInfos)
{
        Console.WriteLine(info.ProductName, info.Category,
info.UnitPrice);
}

Hope that's a bit easier to understand. 

On the database side, that presumes some sort of mapping behind the
scenes - the compiler has to have a way of knowing that there's a table
named products and a column named Category, and if those things change,
then the mapping has to be updated, and at that point you would get a
compilation error.

OT: At one time there was access to some terminal server machines that
you could remote desktop into. Not sure if they are still around...

Eric

-----Original Message-----
From: Discussion relating to the specifics of the C# and Managed C++
languages [mailto:DOTNET-CX@xxxxxxxxxxxxxxxxxxx] On Behalf Of Peter
Ritchie
Sent: Friday, September 16, 2005 11:19 AM
To: DOTNET-CX@xxxxxxxxxxxxxxxxxxx
Subject: Re: [DOTNET-CX] C# 3.0

Thanks for responding, I know you're not "the C# 3.0 guy"--you just have
a blog entry on it :-).

I must be missing something.  I understand that, from the compiler/CLR
point of view, var is strongly-typed.  But, from a
compile-time/programmer point of view it is not.

if I write:
    var productInfos =
        from p in products
        select new {p.ProductName, p.Category, Price = p.UnitPrice};

What methods could I possibly use on collection entries in productInfos
that is guaranteed to be available, other then Object members--like
ToString().  Can I even use other-than-Object members at compile time?
If I can, and I expect Category to be a String type will this work:
foreach(var p in productInfos)
{
    Debug.WriteLine(p.Category.Length); // String.Length }

If that's valid; what happens if I change the database schema and
Category is now an int?  Do I get a run-time exception?

OT: Was I dreaming, or is there some on-line tool at microsoft.com
equivalent to a virtual Visual Studio session?

http://www.peterRitchie.com/

On Fri, 16 Sep 2005 10:38:57 -0700, Eric Gunnerson
<Eric.Gunnerson@xxxxxxxxxxxxx> wrote:

>As I noted in my blog entry I'm a bit out of touch on some of the 
>rationale, but I can give some insight (I hope...). All the additions 
>are there to make the Linq stuff work (there may be other 3.0 additions

>not related as well).
>
>Well, the problem with object instead of var is that you end up with a 
>weakly-typed implementation. For example:
>
>    List products = GetProductList();
>
>    object productInfos =
>        from p in products
>        select new {p.ProductName, p.Category, Price = p.UnitPrice};
>
>What can I do with productInfos? I don't have the name of a type to 
>cast it to, so the best I can do is access the values through
reflection.
>
>With the var syntax, I get something that's strongly typed, and I can 
>get IntelliSense on the instance when I want to access the values.
>
>It is possible (or at least it was going to be possible - not sure if 
>it
>is) to use a real type if you want to pass that type around (ie 
>something like EmployeeContactInfo with just an employee ID, a name, 
>and a phone), but in many cases you don't care about doing that.
>
>Extension methods are mostly there to support the query expressions. 
>The "select" part of the query expression is, IIRC, implemented using a

>query expression.
>
>Eric
>
>-----Original Message-----
>From: Discussion relating to the specifics of the C# and Managed C++ 
>languages [mailto:DOTNET-CX@xxxxxxxxxxxxxxxxxxx] On Behalf Of Peter 
>Ritchie
>Sent: Thursday, September 15, 2005 10:34 AM
>To: DOTNET-CX@xxxxxxxxxxxxxxxxxxx
>Subject: Re: [DOTNET-CX] C# 3.0
>
>Eric Gunnerson has some ramblings at
>http://blogs.msdn.com/ericgu/archive/2005/09/14/466510.aspx, on 
>implicitly types variables and extension methods.  I don't buy the 
>justification of these two additions.
>
>I don't see the need for implicitly typed variables.  The note in the 
>spec. that basically suggests avoiding implicitly typed variables 
>whenever possible says it all.  What's wrong with "Object"?  I have the

>same opinion on extensions that require "var": Anonymous Types and 
>Implicitly Types Arrays
>
>Extension methods seems to be a hack to recover from lack (or removal, 
>considering C++ roots) of nonmember functions or implicit conversion 
>constructors.  It also reaks of a means to circumvent Sealed classes.  
>I started a blog some months ago on extending C# classes without 
>touching the base; but never posted it.  You can accomplish what 
>extension methods give you very simply.
>
>The additional object initializers seem to be just syntactic sugar to 
>avoid having to write constructors.
>
>I don't know what I think about Query Expressions yet.
>
>The only compelling additions, in my opinion, seem to be Collection 
>Initializers and Lambda Expressions.  Lambda Expressions remind me of 
>Continuation Passing Style.
>
>http://www.peterRitchie.com/
>
>On Wed, 14 Sep 2005 15:52:36 -0500, Russell Collins 
><RCollins@xxxxxxxxxxxxxxxxxxx> wrote:
>
>>I couldn't agree more.
>>
>>-----Original Message-----
>>From: Stoyan Damov [mailto:stoyan.damov@xxxxxxxxx]
>>Sent: Wednesday, September 14, 2005 9:25 AM
>>To: DOTNET-CX@xxxxxxxxxxxxxxxxxxx
>>Subject: Re: [DOTNET-CX] C# 3.0
>>
>> On 9/14/05, Eyvind Axelsen <eyvind.axelsen@xxxxxxxxxx> wrote:
>>>
>>> Language specification, examples etc can be found here:
>>> http://msdn.microsoft.com/vcsharp/future/default.aspx
>>>
>>> Very cool!
>>>
>>> Eyvind.
>>
>>  I think I'm going to puke:
>> The implictly typed local variables and arrays extension is letting 
>> lazy
>bastards write JavaScript in C#.
>> The object initializers extension is something which I expected in C#
>1.0.
>> The collection intializers extension is useless, and can only serve 
>> book
>writers write contrived examples.
>> The anonymous types extension is a replacement of a worst practice,
>which is returning a quick-and-dirty tupple (tripplet, whatever). We 
>used to return object[] for doing that.
>> The lamda functions extension is for lazy bastards, which also like 
>> to
>write write-only code.
>> The query expressions extension is, well, no comment. Really, this is
>CSql, not C#.
>> I would really appreciate a slightly more powerful generics to all of
>the "extensions" above.
>> Cheers,
>>Stoyan
>> P.S. Haven't had much sleep recently

===================================
This list is hosted by DevelopMentor(r)  http://www.develop.com

View archives and manage your subscription(s) at
http://discuss.develop.com

===================================
This list is hosted by DevelopMentor®  http://www.develop.com

View archives and manage your subscription(s) at http://discuss.develop.com