.NETDJ: Are there any other features out there that might be added to C# in the post-2.0 timeframe? Anything you can tell us?

AH: Well, I mentioned a little bit about the space we want to explore more, which is the tighter integration between programming languages and databases. We are now including nullable types, which will allow you to talk to your databases a little bit more effectively. There are many other things I'd like to look at, like the mismatch between relational types and language objects, and the lack of query and set objects in C#. You can query in the database - but once the data comes down, you can't further subquery without going back to the database.

The other thing inhibiting proper integration between data and languages is the general lack of type checking in the present generation of database APIs. The compiler won't know anything about whether or not the data in your database will work with your code until runtime. This is certainly a direction that we are keenly interested in exploring and we already have some good ideas.

.NETDJ: When last we spoke, you said that generics and other new language features for C# might not make it into the Compact Framework. Correct me if I'm wrong, but they actually all made it in, didn't they?

AH: I do believe that they have, but you might want to check. All of the other features are just language features, so they are - by definition - in the code.

.NETDJ: What originally drew you to Pascal? Are there any implementations of Pascal for .NET of which you are aware?

AH: I wanted to write an ALGOL compiler because I liked ALGOL, but my partner said there was a new thing called Pascal, so we should go and check it out. It is a simpler language both to teach and to implement, so we went with that!

As far as Pascal for .NET, Delphi is backwards-compatible all the way back to Turbo Pascal, and it can create .NET code quite nicely! Delphi is what Turbo Pascal became. In fact, the 16-bit Delphi was the actual Turbo Pascal code base!

.NETDJ: How did you originally get involved in computer technology?

AH: I started originally doing computers in High School, and that's where I learned ALGOL as my first programming language. When I started at the technical university in Denmark in 1979, they had this event for all the freshmen where we'd go away to a kind of camp with the seniors. So, there I was playing cards with a senior and he lost a bunch of money, so I had to get to know him. It turned out that he was interested in starting a company to sell kit computers.

We wound up founding a company to sell Z-80-based computer kits and I wrote a bunch of different software for this thing. One of the things I wrote after writing assemblers and games and such was what essentially amounted to Turbo Pascal. It was a little 12k Pascal compiler that was burnt into ROM. Ironically, we were recommending that folks yank out the Microsoft ROM for BASIC that came with the kit and put in our Pascal ROM instead!

.NETDJ: Did that company become Borland, then?

AH: No. My Pascal implementation grew and grew and eventually became a full implementation, and got moved to CP/M. We hooked up with the guys who founded Borland, because they were also based in Denmark (though they incorporated in the United States) and they were writing their software using a different implementation of Pascal. We told them about our software and why it was better and they said "Hey, why don't we do a joint venture where we will do the marketing and you do the development?"

This turned out to be a very good thing! I remember that, at that time, software development tools were generally about $500 a copy. So when they said "why don't we sell it for $50," we thoughts they were nuts!

It was, however, a great success. Pascal wound up being the only thing that I did at Borland, which soon moved their headquarters to the United States. It quickly became tedious to commute across the Atlantic, so I moved to the U.S. in 1987.

.NETDJ: Last time we spoke, I asked you for your favorite programming language and you said you didn't have one. So, this time I will ask you - of all the programming languages you've seen -- does any appear to you to just be a complete abomination? (My personal choice would be anything of the TCL/LISP family, for example).

AH: It's hard to say. I don't know. Batch files and the programming that is going on in those definitely need some help! Another language that is difficult to learn, but very powerful, is XSLT.

I see lots of little languages go by in various projects and I tend to stress internally at Microsoft that we all need to get on the .NET Framework so that we all share the same power of the API, regardless of our language choices.

One of the trends I think is interesting is the integration that is happening with programming models like ASP and XAML, which are mixtures of declarative and programming code that give you an amalgam of two different programming disciplines.

What I'm shooting for with the next generation of Microsoft platforms is to use XML for declarative tasks and C# for programming tasks. I'm less of a believer in using XML to "new up" objects in C#.

(continued on page 4)

• The Father of C# Improves His Brainchild
• From the .NET Developer's Journal Archives: A Talk with the Father of C#
• C#, Turbo Pascal, C++, PHP...and the LEGO Brick: Denmark's Leading Exports