Pure Paul Graham:
"Do you have any idea how long the future is? Do you really
think people in 1000 years want to be constrained by hacks
that got put into the foundations of Common Lisp because
a lot of code at Symbolics depended on it in 1988?"
In my CS degree we never learned about how language will evolve. We learned how they work, how to write new ones, how to translate between them, but never how to future-proof them.
But right now it looks like C will remain the language of the future :)
In a "Long Now" talk [1], science fiction author Vernor Vinge mentions the "software midden heap", the layers of software standing on the shoulders of and papering over the bugs of earlier software layers. Will anyone be able to excise rotting middle layers without breaking software compatibility?
Another example is from Vinge's novel A DEEPNESS IN THE SKY. He describes a software system thousands of years in the future that still uses the time_t epoch. None of the system's space-faring users know the significance of the date 01970, though a few believe it is the date when computers were invented. :)
20 years after the "discovery" ;) of UNIX, Linus began developing a Linux from scratch. Linux is now 20 years old. Is some college student writing a new operating system that will cut the cruft from Linux? Even Linux is far from a clean slate design because it embraces (and extends) POSIX APIs and UNIX conventions.
Will we still be using Linux (or Linux derivative) in 2050? 2010?
[1] Vernor Vinge: “What If the Singularity Does NOT Happen?”
None of the system's space-faring users know the significance of the date 01970, though a few believe it is the date when computers were invented. :)
And some thought we measured time from the date of the first moon landing.
More relevantly in the book is the class of people known as programmer-archaeologists who specialise in digging through libraries of software looking for code that already does what they need.
The future is also too long to assume our current generation languages will define it. While Lisp is a major step forward, I think it is by no means the end of language progress in the current couple of decades, let alone 1000 years.
It might win hands down now, but probably not for a 1000 years.
Lisp certainly includes certain 'timeless' aspects, such as lambdas and homoiconicity, but other features (like conses everywhere, IMO) are simply implementation details of the time Lisp was first created.
I'm concerned that Lisp advocates, by repeatedly saying that Lisp is the final step in language design, the ultimate language,...etc might distract some from looking for additional, newer, insightful concepts in computation and language design.
Cons is an implementation detail that no longer exists in reality. Most Lisps implement conses as arrays, and some Lisps, like Dylan, don't even use them to represent structure.
No programming language will survive a 1000 years. Not many works of science or art, or even human language have.
Programming language research has already grown past Lisp itself, or any other programming language, really. What we have now are logical models, their derivations and their proofs. If any of them can be materialized in an implementation, made practical, or even commercialized, great. Otherwise PL research is happiest as .. logic.
Edit:
اخ محمد، عجبتني رؤية لغتك البرمجية العربية. قمت ايضا بأنشاء بضعة لغات برمجبيات عربيات، معظمهن كلغات شبه لسبية :-)
In my CS degree we never learned about how language will evolve. We learned how they work, how to write new ones, how to translate between them, but never how to future-proof them.
But right now it looks like C will remain the language of the future :)