Hi rick,

>You can do that today. .NET supports dynamic languages - JScript is one of them and you can code that way. But you can't mix it in a single project (unless it's ASP.NET) and you can't mix the features of the languages effectively inline as you would need to.

I know of the development of the DLR and dynamic languages, it is an interesting development.

>Microsoft has just announced the Dynamica Runtime which will open up more choices for dynamic languages but I doubt it will ever be a complete mix and match where you have a mix of both in the same block of code. Nor do I think should you have that option. Either you are explicit or you're not.

We'll see how this is going to develop.

>>I welcome strict typing for general algorithms, though there is some comment to make here as well. WHY NOT CHECK TYPE when typing or saving the source file ???? CAN ANYONE TELL ME PLEASE. I'm I just demanding too much, or does MS thinks developpers are stupid ?? Why do I have to wait through compile time ?

>And how would express that hmm? Is the compiler or the editor going to read your mind about what your intention is or whether that SSN number without dashes is a string or a numberic value?

Well, why not track the problems at the source: when typing. So yes, the editor already should catch 95% of the problems. I think we have to realise that programming through notepad or any other dumb text editor is something of the ancient past. We need full intelligent editors with intellisense and automatic formating AND compile checking. Hell we have that in VFP for a long time. One thing I asked JohnK a couple of years ago, is a new feature in VFP to check for undeclared variables in the editor which will make it a lot easier to avoid the problems of mistyping variable names.

Again it is soo much easier to track the problems at the root, rather than wait until designtime. Why wait for the compiler to catch it? This does not mean that the compiler should skip it, but only that I want to catch the error as I make them.

>Frankly I honestly don't get your point. What is the issue exactly? That you have to provide a type name or god fobid cast a value to a specific type?

It is very simple:
- If I want to use strict typing (obviously for the majority of straightforward algortithms), it should respect that.
- If for some reason I can't be sure of the type (untyped datasets, late bound com servers) I want to have the possiblity to work with variables that are dynamic without having to CAST use reflection or any other techniques. Of course, when using a dynamic variable in a stricttype expression we obviously need to cast, but the other way arround would not be neccesary. This would for example working with late bound com servers as easy as it is in VFP.

Again, it is stepping outside of that fixed mindset of what people think about the black and white world of strict typed and dynamic typed languages. The truth is somewhere in the middle, not one or the other.

Walter,