The Future of Lisp

[vityok]

Whilst watching his talk, it seems to me that the LINQ implementation is described in detail in the famous book.

[findinglisp]

Whilst watching his talk, it seems to me that the LINQ implementation is described in detail in the famous book.

In general, I kept thinking to myself all the way through the talk, "Yea, Lisp can do that... and that... and that..." Ironically, even during the discussions of metaprogramming and DSLs, Lisp never really comes up. That really surprised me, particularly since Lisp is starting to be mentioned a lot more.

So, in general, from a Lisp-specific perspective, the talk is a bust. What I thought was most interesting, however, was the enumeration of the various programming problems that Anders is wrestling with because I think the problems are universal, whatever the language. Interestingly, I think that Lisp is well-positioned to help deliver solutions, but we need to move past 1990s Common Lisp in order to do so.

[dlweinreb]

Actually he does mention Lisp by name, but what he says is that it was the first "functional programming language". It depends what you mean by "functional", but if you mean no side effects or something like that, well, you could sure do a lot more without side effects than you could in FORTRAN, but clearly modern Lisp isn't "functional" in the sense that Haskell is.

However, I agree that not mentioning Lisp in the contest of "internal domain specific languages" was odd. But, few people acknowledge Lisp as a source of ideas, even when it clearly is. Java is the one that really bothers me; Gosling never acknowledges how much Java owes to Lisp.

Does anyone here know enough about C# and LINQ to know whether the syntactic sugar that they added was something that anyone could add in their own back yard (a la using Lisp macros), or whether the C# designers had to add it as a specific new feature in the C# compiler?

[dmitry_vk]

Does anyone here know enough about C# and LINQ to know whether the syntactic sugar that they added was something that anyone could add in their own back yard (a la using Lisp macros), or whether the C# designers had to add it as a specific new feature in the C# compiler?

LINQ is a compiler feature. In order to add it, both compiler and Base Class Library were modified.

[findinglisp]

However, I agree that not mentioning Lisp in the contest of "internal domain specific languages" was odd. But, few people acknowledge Lisp as a source of ideas, even when it clearly is. Java is the one that really bothers me; Gosling never acknowledges how much Java owes to Lisp.

Interestingly, Steele does acknowledge this, however. I always found his statement about "Java dragging all the C programmers halfway to Lisp" (paraphrased) to be quite funny.

[Jasper]

About parallelism, the pcall library makes it look(at least) as if it is easy. As others noted, you don't necessarily have to analyze the code to determine when to slip plets/join in, you can just run it and time which bits are worth paralellizing. So it is possible to do it automatically with existing common lisp code.