It's all about processing power. In VS.NET with Visual Assist ( don't know which application has this feature, Visual Assist is a VS.NET PlugIn ) if you type your code it's compiled on the fly in the background and compiler errors are underlined like spelling errors in word. The tooltip shows the compiler error. To be more correct, it's only parsed, not compiled, but the result is the same. It even uses the office spellchecker for your comment and string lines.
It's a great tool, but I wouldn't want to use it on my old PII 400 that I ran VC6 on.
Yes, I have VS.NET, but I've been only using the Intellisense features. My main complaint is that it's difficult to hop between different places in the source code. Usually this isn't too bad unless if the code is badly designed but I would like an interface overall more like a browser. For example, I would like to view a class like Tile that looks something like this
Code:
----------------------
| Tile |
-------------------------------------------------------------
| Tile(const CDC* dcSrc, int x, int y, int w, int h) |
|........ |
| bool draw(CDC* dcDst, int x, int y) |
--------------------------------------------------------------
And then clicking on the appropriate method function will show a screen with filled with the code for the function and clicking on the back button will go back to the class definition. When editing any particular method function there's more information than is needed to edit the function, I think. A user editing a given function only needs to see the class data and maybe some global data. Ideally, a programmar would have two views, one showing the code he or she is editing and another showing the data structures.