Visual C++ vs. All other Environments ??

hi,
I used Dev C++, then switched to Codeblocks, but I'm having nothing but problems with it, lots of bugs on Vista.
I just watched a tutorial on Microsoft Visual C++ and was blown away !
you can just click on tools and it adds all the code for you.... and actually SHOWS what you're making WHILE you make it....
do any other programs do that ? does codeblocks do that if I could get it to work ?

so far, all the ones I've tried are straight up code editors and don't " show " what your doing in real time.... and don't have these tools to just add features to your windows...
is this true ? is Visual Studio the only one that does that ? can codeblocks do that ?

thank you all very much in advance.

I'm not a fan of that sort of "drag and drop" GUI design. I mean, it's great for prototyping, but I like knowing what code is in my program.

Quote:

---

Originally Posted by Dr.ticklefish

I just watched a tutorial on Microsoft Visual C++ and was blown away ! you can just click on tools and it adds all the code for you....

---

"Point and click" programming may be OK if you're already experienced in Windows programming, but you can't learn to actually write programs using that method.

I could understand if you're experienced in Windows programming, and the automatic code generation makes your work easier. For example, the resource editor allows you to define menus and dialogs, and I use it all the time. But at the same time, I know what a resource file is, what the different components of one are, and if need be, I can edit one so that I can fix problems on the fly. Code generation is another one -- I usually use the project wizard to create MFC classes, but again, if the time comes I need to actually fix or add some code, I know exactly where to go, how to debug the code, and how to properly fix/enhance the code.

If you don't know what you're doing and your just "clicking away" to create your programs, you are in a big deal of hurt if you ever need to fix a bug, add something customized, or just in general, what the heck you've just produced (in terms of code).

As an indication, look at the thousands of posts from persons here who have gone the route of not learning proper programming, and instead used the "point and click" approach of creating a program. Most of them ask the most basic C++ questions as to how to fix their problem, fix a bug, make a variable accessible in a class, etc. All questions that are answered within the first chapter or two of any C++ book.

Regards,

Paul McKenzie

Quote:

---

Originally Posted by Paul McKenzie

"Point and click" programming may be OK if you're already experienced in Windows programming, but you can't learn to actually write programs using that method.

---

Hear, Hear!

I'm in agreement with Lindley, and as you might expect have another 2 cents worth on that point.

First, I'll admit I find the Visual Studio IDE and it's C++ compiler to be first rate. Nothing is perfect, but the VS debugger, the editor and it's tabbed interface make a combination that's hard to match in a single product.

I'm from the era before debuggers were widely available (well, just before). I've watched the tools advance, as well as the languages.

I have no idea which video you saw, but I can expect the focus was on .NET, even if not explicitly stated. VS 'wizards' for MFC, combined with the dialog editor, do essentially what you saw, but it's not magic. The benefit is quite limited. For someone not all that familiar with MFC it is like 'automated documentation/skeletal application starter' service. You want to add a button, so instead of looking that up in the manual and creating the response function in the appropriate class, the editor does that, connecting the control you've painted to the code responding to it and puts the cursor on the first empty line of the response function so you can begin coding.

It's nice at first, and some people do depend upon it, but it's a crutch IMO.

If you do anything outside the conventions of the wizards (and sometimes it seems to hiccup even if you don't), the whole thing breaks down like a house of cards. Perhaps it's better in VS2008, I've not tried, but since VC6 and all through VS2005, there was this annoying tendency for it to get a little confused, sometimes refusing to allow control/code linkage, not having appropriate response message options to add - it's actually quite restrictive.

It's also very MFC/Windows/.NET centric (depending on which targets you're making). The results aren't portable, and while it's easy to dismiss that requirement for a range of products, there are other ideas that may be of interest to you.


I did something long ago that Adobe did in their development strategy (I can't say which came earlier, Adobe was secretive about this for some time). Adobe created "Adam" and "Eve", two interface ideas that when combined allow one to develop cross platform dialog resources. It takes quite a change of mindset, because you basically give up some control (the system formats dialogs for you, not the other way around, basically). The results look native in the target OS (consider looking at Photoshop on Apple and then on Windows - neither are in MFC, nor QT, nor wxWidgets, to my knowledge).

Most of the other frameworks, QT, wxWidgets, etc. have some form of resource editing/scripting and a convention for connecting resource to code. The adobe approach is interesting (and I have to go from vague memory here, I might be projecting my own approach) in that the definition of the dialog resource is also the code mapping.

In my own work I created a dialog mechanism for GUI work. It's independent such that I've used in wxWidgets, QT and MFC (the backend is portable). The dialogs can be 'hosted' on standard popup/modeless/modal dialog windows, or as panes in dockable panels, or as main or portions of main windows in MDI or tabbed dialog host frames.

When I create a dialog it's via a dialog context (it's named, some data definitions might be provided from code as a start). The context is placed into a host (popup dialog, mdi child, tab, etc. The context is initially empty, like an empty dialog resource in VS.

I compile and run this 'empty' application and run the dialog, in a designer build (the designer portions are optional compilation units). When I launch that dialog, it's blank, and I enter a designer mode within the running application. I paint the dialog from there. This painting isn't done in an IDE or development tool, it's done in the skeletal application.

As I add controls, they're connected to representative data sources as appropriate. Sources of data include application supplied configuration data (user options dialogs), XML data, SQL data, other binary sources, etc. Within the dialog editor I can add "fields" which will become SQL columns, XML fields, etc.

If I don't have any other specialized work to do (say this dialog is just collecting a client record, for example) - that may be all the work there is to do. The context connected to a SQL table, for example - added columns to the SQL table's definition, if I requested that - allowed me to select form SQL columns from that table (and of course, this all involves related tables) - in other words I'm not just focused on tying controls to empty code.

The whole concept stores the dialog in XML data, which can be transported by SQL if required. This means that the UI of the application can be stored in SQL, transported with a user by identity to any workstation they use, and can be tailored for each identity or group. In other words, for a multi-user corporate application, each person might have a different view of the GUI/Data.

Long ago I realized about the 3rd dialog I made, I was repeating myself. Instead of letting that go on, I analyzed what I was doing, made that code part of a framework, and now they're selectable control/feature/data connection options in an editor that might as well be a stand alone dialog creating engine.

The controls can be attached to messages which allow for validation functions, message responses (commands, etc), other I/O, custom controls (which usually get added to the library, so custom becomes fewer and fewer over time) - etc.

In other words, I'll continue to develop in C++ as I design the dialog, and the connection between the two may be through response functions/functors/classes which apply only to those controls with genuinely unique requirements. I find it more productive than the wizards of VS. I haven't had the inclination to wrap it into a product, but it's not all that tough to do.

In the 90's, when I first put this together, it took a few weeks to get the basic idea going, but then it accumulated over the span of my career.

I got the idea by meeting one of the people who worked at PARC in the creation of the first Windowed OS/Computer system - the Alto, from which Jobs and Gates were inspired. The lessons from that Xerox system have yet to be fully realized - the GUI interface was about 1/3 of the genuinely important invention of that project. The notion of programmable controls and objects are among them.

Anyway, I've gone far off the end - VS is good stuff, but not the pinnacle. We actually should research the history of the Alto - some gems are still in there somewhere. It was based on object oriented LISP, and all the interface elements were objects (according to the person I had the fortune of working with briefly) - all in the mid 70's.

I have a bit different perspective and view knowing the MFC wizards as part of the learning curve.

Sure, you can't really learn the underlying code if you start out by only relying on the wizards, but once you've progressed past the initial learning stages, then there is much to be gained by using the wizards and following MFC convention.

I kind of view it like the same as following code style guidelines within a team. Too often I see coders go off and try to create MFC code from scratch only to miss details and cause problems later (then of course complain because MFC 'was broken').

Of course, I'm not talking about the coder that blindly uses the wizards without understanding what code the wizards generated and how the generated code works.

As far as portability, I have to respectfully disagree. If the developer has a requirement to code on a variety of platforms, then he/she may want to use frameworks that are platform agnostic.

However, there are many projects out there that only target a single platform [family] and aren't ever realistically going to be migrated to another platform. To me, using non-targeted frameworks is just extra overhead because the frameworks must code to the lowest common features of the target OS's or incur added complexity from special casing. Imo, it's not unlike the mistake of over-generalizing every class for reuse.

thank you all for the insight.
I am fairly new to C++ but I'm aware of what's going on in the code in most cases.
I just plan to use it for fun and make little programs for my own use.
I liked Dev C++ but was told on here to give it up because it's too old and outdated.
I tried Codeblocks next, and liked it.... but it just won't work right on my new computer with Vista, too many bugs.
I am willing to buy a program, but I was wondering before I spend my cash,

#1. is Microsoft VS is the ONLY program that has a GUI wizzard that allows the drag and drop style creating ?

#2. and can someone give me a link to download a trial version ? there seems to be tons of versions and types of builds and tools and the couple I've tried so far from Microsofts page took forever to download and run, and then don't work.

I have Vista Home Premium with a Centrino proccessor.

thanks

Quote:

---

Originally Posted by Dr.ticklefish

I am willing to buy a program, but I was wondering before I spend my cash,

---

Well you really can't go wrong using Visual C++.

Quote:

---

#1. is Microsoft VS is the ONLY program that has a GUI wizzard that allows the drag and drop style creating ?

---

Borland has (or had) the same type of interface. That is the only other one that I know of that had a "Visual" type of design (not using MFC but with their own proprietary classes -- VCL if I remember).

Regards,

Paul McKenzie

http://www.microsoft.com/express/download/default.aspx
You can download free VC++ 2008 Express Edition. MFC is not supported, but Win32 applications are supported, including the Dialog editor and drag-and-drop...
See more about Express Editions here: http://www.microsoft.com/express/support/faq/

Full Visual Studio 2008 90-Day Trial:
http://msdn.microsoft.com/en-us/visualc/bb655861.aspx

Hey Thanks Alex. I'll try them Both.
I also took Pauls tip and looked into Borland. They do have one very much like MVS ,
and it's completely free. it's called Turbo C++ and it's very similar. I'll try the MVS and see if it's that much better than Turbo C++ and worth buying. So far I'm liking Turbo C++ and it's FREE !
thanks for the tips. If someone else wants to try Turbo C++ here's a link
https://downloads.embarcadero.com/free/c_builder

( third one down the list )

cheers

hey Alex,
that's the same download page I tried a few days ago.
The Epress VS C++ 2008 download won't work for me. It downloads Ok, then during install, it hangs at the 52MB mark and eventually quits and says installation failed.
anyone else have this problem with this download ?
thanks

the full version doesn't work either. the download manager says " error " every time.

: (

Try this page:
http://msdn.microsoft.com/en-us/eval.../bb633753.aspx

According to file size, it is full installation and not WEB installer. This may be ISO image which can be burned to DVD or opened as virtual drive or archive.
I don't use Vista, so don't have such installation problems... Do you have Administrator account?
Turbo C++ is really free, but it is too old, this is not option, to my opinion.

This is direct download link to VC++ Express Edition ISO file:
http://go.microsoft.com/fwlink/?LinkId=104679
I found it making Google search, from some reason Microsoft doesn't give this link.

Quote:

---

Originally Posted by Alex F

Turbo C++ is really free, but it is too old, this is not option, to my opinion.

---

Borland has recently updated their compilers. This was about a year or so ago.

If you're refering to the old Turbo 'C' from the early 1990's, then yes, that is way too old (unfortunately, it seems that many schools still use this compiler).

Regards,

Paul McKenzie

hey,
well Alex the ISO file won't download for me either. It's not the links, it's the download manager.
it keeps saying error. is there a way to download it without using their download manager that they provide ?

and Turbo C++ seems to be pretty new according to their site. It seems pretty cool,
I would really be interested in hearing from anyone out there that's good with C++ that uses MVS on a regular basis. If any of you would be willing to try Turbo C++ for writing a program or two and letting me know if it's up to par or if MVS is worth shellin' out some bucks for,
I'd really appreciate the feedback,
and I think it would make a good topic here in the forum as well.

Thanks

To download the ISO, you need to install the ActiveX download mgr first.

I took the time to install Borland C++ and check it out. I went to Borland's site and did a search for 'Turbo C++' and it redirected me to an install.

I wasn't too impressed during the install process, since it failed and told me I needed to install Visual J# .Net v1.1 Redist package first.

Okay, I do that and then reinstall and when I'm done it shows up as "Borland Developer Suite 2006".

My first impression was that it hasn't changed much since I quit using Turbo C++ back in 1995.

The forms designer looks like VB, no intellisense.

I created a new SDI project and console project, and it looks like something that I could create with Visual Studio 4.2.

Not impressed.

So, Does it sounds.. MS VC++ is the one and only one best compiler with IDE and debugger available today? :p

For all its shortcomings in so many other software areas, Microsoft has produced a first-rate development suite and oddly enough (as far as I know) they did it all internally and didn't just buyout/re-badge someone else's hard-work as they've done with so many other products.

Quote:

---

Originally Posted by BigB

So, Does it sounds.. MS VC++ is the one and only one best compiler with IDE and debugger available today? :p

---

For Windows, true. But they wrote Windows after all...

Quote:

---

Originally Posted by hoxsiew

For all its shortcomings in so many other software areas, Microsoft has produced a first-rate development suite and oddly enough (as far as I know) they did it all internally and didn't just buyout/re-badge someone else's hard-work as they've done with so many other products.

---

Well, they have been developing it for a long time. I remember using the compiler manually from the command line. Turbo C for Windows (1990) got MS to come out with Visual C 1.0. Didn't work very well, so they quickly came out with VC++ 1.5. (1.52 was the last DOS based). Simultaneously they came out with 2.0 which was Win32 and required NT 3.0 to operate. Anyway, 1.5 was the first usable Visual C++ implementation (also MS lowered the price from $4500 to about $300). It used the Microsoft Word keys in the GUI and was fairly easy to use. It pretty much stole most of the market share back from Turbo C for Windows. While Borland is still around, it has never recovered the market share that it lost in 1991.

MS is getting dumb again in that it raised the price back to around $1000 (unless you are a student). Guess they feel secure in their share of the market.

Thanks egawtry, that was something interesting history of the compiler :)

Quote:

---

MS is getting dumb again in that it raised the price back to around $1000 (unless you are a student). Guess they feel secure in their share of the market.

---

wah! $1000 for MS VC++ ?:eek: oops.. Its double the price of Windows7! and, if i'm gonna purchase VC.. i'll surely get my pant's torn:D

How can Microsoft be this much greedy for money? gates had earned enough, and he still need bucks for a bath?:mad:

@egawtry

Excuse me for my off question.. i had a doubt.., since you've some knowledge of past, i think i'd ask you.. is Windows a copy of some other operating system? did they copied anything from other people or did they wrote that from scratch?

Did any other people copied from Microsoft Windows? i mean, linux or unix or any other?

Why the Open Source people hates Microsoft the most? As Microsoft is not the only company who sells their product for price... why Opensource is not hating Adobe, Autodesk and all? I have seen Opensource buddies crying at Windows and Microsoft like Billgates had stolen their underwears.. :D

i heard that Apple Jobs stoled some concepts from Windows like multitasking, taskbar and all.. and Apple Jobs stoled GUI from Xerox and all..

whhhhaaaa ???? $1,000 ?????

FORGET IT.

I'll stick with Turbo C++ for now. It might not be as good, but for a guy like me it's hard to justify $1,000 over FREE.

I still want to try MVS, but I can't get it to install. : (

thanks for all the insight.

I just clicked BUY NOW on microsofts site to see what the price is.

$ 5,469.00 !!!!!!!!


are they OUT OF THEIR FREAKIN' MINDS ?????????

Quote:

---

I'll stick with Turbo C++ for now. It might not be as good, but for a guy like me it's hard to justify $1,000 over FREE.

---

Visual C++ 2008 Express Edition is free... :D

I'm on it... :p

Turbo Crap++ sucks:sick: its a weird compiler:thumbd:

I feel VC the best. its easy, advanced and cool :)

Quote:

---

I just clicked BUY NOW on microsofts site to see what the price is.

$ 5,469.00 !!!!!!!!


are they OUT OF THEIR FREAKIN' MINDS ?????????

---

You might have checked the price of entire Visual Studio 2008 Suite.... i think VC++ individually might have $1000 bucks:confused:

Quote:

---

Originally Posted by BigB

...is Windows a copy of some other operating system? did they copied anything from other people or did they wrote that from scratch?

i heard that Apple Jobs stoled some concepts from Windows like multitasking, taskbar and all.. and Apple Jobs stoled GUI from Xerox and all..

---

Well, it is very complex. The original GUI concept was designed back in the late 60s by Xerox. From there it was a involuntary collabrative effort (everyone stole from everyone else). I personally came into the picture in the late 70s. There was text based windowing, but no mouse or light pen on the average home computer or UNIX terminal. When the power of the original PC arrived, the game developers demonstrated that graphical windowing was possible. X-Windows was around for big mainframe machines with very expensive graphics terminals (Sun stations - 198x). Apple created the Lisa (1984), which prompted MS to create Windows 1.x (1986). Both had interfaces mostly stolen from X-Windows. Apple had the taskbar stolen from the original Xerox interface. I only saw copies of it running, but never actually used it myself. The Lisa was a bomb, so Apple futzed for a few years and came out with the original Mac while MS came out with Windows 2.0 (1988). Also around was the Amiga, a multitasking computer about 10 years ahead of its time (it could do everything a Windows Multimedia machine from the late 90s could do). About that time the IBM/MS wars were taking off, so IBM came out with OS/2 which had the first real multitasking on a PC.

Take the driver ideas and multithreading from OS/2, combine with Windows 2.0/386, and you get Windows 3.0. After fixing the network stuff, then you get Windows 3.1x. Add in the taskbar from the Mac and Amiga, and you get Windows 95. WinNT 3.0 (I heard a rumor MS bought it from someone) was usable with WinNT 3.5 in 1995.

About this time (early 90s) Linux was adapted from Minux by Linus Torvolds. Full UNIX for the PC which allowed X-Windows on a PC. Since then the core X-Windows hasn't really changed, but the layers on top of it have improved to the point where they are actually usable. Mostly borrowing ideas from Macs and Windows.

=========================

OpenSource is an ideal that sounds great philosophically, but is really, really dumb business. To release Open Source code, you have to assume that you have more resources than anyone else out there, so you can maintain it better. Essentially what it does is give away all your trade secrets. Since you cannot copyright a method like a jpeg decoder or command line parser, you are just giving your competition a solution that they then don't have to pay programmers to figure out.

That is why OpenSource hates MS. MS does not give away its core code, so they have to do it themselves. Also, they have the basic Ideal that everyone should share everything and, of course, keeping code secret violates that Ideal.

Also, MS bashing is almost considered a sport. While I enjoy a MS joke as much as the next guy, I do think that it is mostly undeserved. The anti-MS lawsuits here and in Europe are a good example. MS did nothing, but got royally bashed. Especially with the ignorant courts in Europe who thought that IE was causing a monopoly.

Quote:

---

Originally Posted by egawtry

WinNT 3.0 (I heard a rumor MS bought it from someone) was usable with WinNT 3.5 in 1995.

---

My understanding is that NT was designed from the ground up by David Cutler and his team at Microsoft.

Quote:

---

Originally Posted by Arjay

My understanding is that NT was designed from the ground up by David Cutler and his team at Microsoft.

---

Could be. It was just a rumor. I know that I had to custom write software for it because so much was missing from the API. I heard that there was a big argument about NT3 at MS, that the designers wanted to go one direction and they were being told to go another; a big corporate shoe came down and forced them to conform to the Win32 API used by Win95 thus creating NT 3.5 and then NT 3.51. That is why NT POSIX died.

I worked in that group from just before the release of NT3.51 to just before XP shipped and hadn't heard anything about that. I can tell you, although the function signatures of NT may be similar to Win95, the internal code is far different.

As far as making the api set match on both operating systems, it was a good business decision. Remember at that time, it was difficult to get folks to write 32 bit programs. There was a Win32s translation layer that was suppose to allow Win32s app to run on a 16-bit OS so that developers could get an early start. Unfortunately, this didn't work too well because not many developers took advantage and later on it caused confusion because folks kept trying to write Win32s apps even years later. A big challenge of NT systems were trying to get applications to run on it. Back then, NT was the unpopular step child as compared to Win95. All the NT folks knew it was superior, but the public didn't know it yet.

The issues with the applications running on NT was that Win95 was way more laxed at what it would let run on the api level. If someone had got some gdi calls messed up, often times Win95 would let you do it; whereas NT would throw an error. Microsoft was in constant communication with 3rd party app developers to try to get them to fix their code. It was a struggle, because the 3rd party devs would say, "but it ran on Win95...". Eventually, MS just added a shim layer that would identify the apps and correctly fix up the bad api calls.

Even though the Win32 api set matched on both platforms, it was difficult to get folks to transition from Win9x to NT based systems. I believe MS had hoped to do that by Win2000, but it didn't really start until XP. A good part of this is because MS just had to wait until some of the older applications died off. Earlier Win9x and NT systems supported running 16-bit legacy applications; however, Win95 supported running most every 16-bit app including those that directly accessed the hardware (like video card access in games). NT didn't allow app's direct hardware access and Dave Cutler & Co. wasn't going to compromise and allow app's direct hardware access (for good reason). As time went on, access on various levels became locked down even more as security became more on everyone's radar. While security was always important in NT from its inception, it became even more so.

Quote:

---

Originally Posted by Arjay

I worked in that group from just before the release of NT3.51 to just before XP shipped and hadn't heard anything about that. I can tell you that, although the function signatures of NT may be similar to Win95, the internal code is far different.

---

Like I said, you probably know better. I am just going by what was discussed on forums/newsgroups at the time.

Of course the internal code is different, what I was referring to was the extra control libraries that made NT 3.51 actually able to run Win95 software where it couldn't before. I sure hope MS wasn't dumb enough to completely rewrite those controls (ListView, ToolBarCtrl, etc) !

I remember when MS finally came out with NT4. There was excitement! The Explorer Shell commands would now work on Windows NT! :-)

As for the security, I fought that a bit. I had a major application that in Windows 2.0/286 wrote to a second monitor by accessing B800. Windows 3 had a back door which still allowed this. I liked NT because the Networking was fixed and it multitasked better. It took a lot of work to get my second monitor runing again though. I eventually had to use a third party driver to do the trick.

Quote:

---

Originally Posted by egawtry

Of course the internal code is different, what I was referring to was the extra control libraries that made NT 3.51 actually able to run Win95 software where it couldn't before. I sure hope MS wasn't dumb enough to completely rewrite those controls (ListView, ToolBarCtrl, etc) !

I remember when MS finally came out with NT4. There was excitement! The Explorer Shell commands would now work on Windows NT! :-)

---

Actually much of the existing Win9x shell code had to be rewritten for NT4. It wasn't being 'dumb', it was out of necessity. Win9x was essentially a 16-bit OS with a 32 bit layer on top. Much of the shell code written for it wasn't going to work or was robust enough to work on NT4, so it was rewritten.

Quote:

---

Originally Posted by Arjay

Actually much of the existing Win9x shell code had to be rewritten for NT4. It wasn't being 'dumb', it was out of necessity. Win9x was essentially a 16-bit OS with a 32 bit layer on top. Much of the shell code written for it wasn't going to work or was robust enough to work on NT4, so it was rewritten.

---

And thus was born the infamous "rootbeer" commands.

@ Arjay and egawtry

lol, that was a great discussion... its really Interesting to hear from the history of MS and Windows....

so, from one of the above post, i think Arjay worked in MS Windows team...

Anyway, i am an yesterday comer.. the earliest operating system i've seen installed on PC was Windows 3.1 in my elder bro's PC... i was not big that time :D

Its great, MS had re-written entire code for WinNT... now, there is some rumour i've heard they have completely dumped the code of Windows Vista, which was a flop, and had Re-Written almost all the code for Windows7

Although impossible to believe, but i had to believe.. when i first used Windows7... because. its simply Awesome!!! Even i can rate it 5 out of 5 for its User Experience and Wow factor....

I think MS really had re-written the entire code... :)

Quote:

---

Originally Posted by BigB

@ Arjay and egawtry

lol, that was a great discussion... its really Interesting to hear from the history of MS and Windows....

so, from one of the above post, i think Arjay worked in MS Windows team...

Anyway, i am an yesterday comer.. the earliest operating system i've seen installed on PC was Windows 3.1 in my elder bro's PC... i was not big that time :D

Its great, MS had re-written entire code for WinNT... now, there is some rumour i've heard they have completely dumped the code of Windows Vista, which was a flop, and had Re-Written almost all the code for Windows7

Although impossible to believe, but i had to believe.. when i first used Windows7... because. its simply Awesome!!! Even i can rate it 5 out of 5 for its User Experience and Wow factor....

I think MS really had re-written the entire code... :)

---

Frankly I don't know where these rumors come from. I guess the MS haters will come up with anything. If you re-read what I wrote, I didn't say that MS re-wrote entire code for the NT4 shell - I said a portion of code was re-written (and when you look at how large the entire code base for NT was, it would be a relatively small portion).

I'm not involved with MS at present, but I can tell you there is no way that the entire code base for Vista was dumped. Vista is the same family this is NT starting with NT3.1, NT3.5, NT4, Win2K, XP, 2K3 Server, Vista and soon to be Windows 7.

As far as I understand, there were quite a few optimization made to Win7 and UI improvements, but it's primarly the same code base.