10.3.05
Yet another Links comment.
A Links comment from Chui Tey. Pat Hanrahan of Stanford gave an invited talk at POPL 2005 on a similar theme, arguing that game processors ose a worthy challenge for language designers. He observes that game processors have a steeper "Moore's law" curve than traditional processors, and hence will outperform them by an exponentially increasing factor.
----
Another area worthwhile looking at is in designing languages for the current generation of game machines, where parallelization is important.
It is worthwhile because:
1) Computer game development do not suffer the constraints of corporate programming, and are more adventurous with tools. A lot of games run with embedded scripting languages. Why not have highly parallelizable functional languages for computer generated graphics?
2) Existing languages are inadequate, and it has taken several years for developers to learn how to squeeze performance from the Sony PS2 architecture.
3) Computer gaming is a big industry, so there will be sufficient commercial interests in the project
4) Specfic target constituency. It's not as important to have broad library support compared to a language targeted for general purpose use. Our baby language needs a niche to survive and prosper. Most of the goals mentioned do seem to fit this market.
5) Avoiding hard things. Programming for the business processes in the real world is hard. It's full of non-strictly typed interactions and full of side-effects. LISP picked up a bad name because AI was too nebulous an aim.
6) Cell processors - we need new languages for them. Once proven, there is a general future for this language.
7) The coolness factor associated with computer games can accelerate the adoption of a language
# posted by Chui Tey : 1:43 AM
Another area worthwhile looking at is in designing languages for the current generation of game machines, where parallelization is important.
It is worthwhile because:
1) Computer game development do not suffer the constraints of corporate programming, and are more adventurous with tools. A lot of games run with embedded scripting languages. Why not have highly parallelizable functional languages for computer generated graphics?
2) Existing languages are inadequate, and it has taken several years for developers to learn how to squeeze performance from the Sony PS2 architecture.
3) Computer gaming is a big industry, so there will be sufficient commercial interests in the project
4) Specfic target constituency. It's not as important to have broad library support compared to a language targeted for general purpose use. Our baby language needs a niche to survive and prosper. Most of the goals mentioned do seem to fit this market.
5) Avoiding hard things. Programming for the business processes in the real world is hard. It's full of non-strictly typed interactions and full of side-effects. LISP picked up a bad name because AI was too nebulous an aim.
6) Cell processors - we need new languages for them. Once proven, there is a general future for this language.
7) The coolness factor associated with computer games can accelerate the adoption of a language
----
Another area worthwhile looking at is in designing languages for the current generation of game machines, where parallelization is important.
It is worthwhile because:
1) Computer game development do not suffer the constraints of corporate programming, and are more adventurous with tools. A lot of games run with embedded scripting languages. Why not have highly parallelizable functional languages for computer generated graphics?
2) Existing languages are inadequate, and it has taken several years for developers to learn how to squeeze performance from the Sony PS2 architecture.
3) Computer gaming is a big industry, so there will be sufficient commercial interests in the project
4) Specfic target constituency. It's not as important to have broad library support compared to a language targeted for general purpose use. Our baby language needs a niche to survive and prosper. Most of the goals mentioned do seem to fit this market.
5) Avoiding hard things. Programming for the business processes in the real world is hard. It's full of non-strictly typed interactions and full of side-effects. LISP picked up a bad name because AI was too nebulous an aim.
6) Cell processors - we need new languages for them. Once proven, there is a general future for this language.
7) The coolness factor associated with computer games can accelerate the adoption of a language
# posted by Chui Tey : 1:43 AM
Another area worthwhile looking at is in designing languages for the current generation of game machines, where parallelization is important.
It is worthwhile because:
1) Computer game development do not suffer the constraints of corporate programming, and are more adventurous with tools. A lot of games run with embedded scripting languages. Why not have highly parallelizable functional languages for computer generated graphics?
2) Existing languages are inadequate, and it has taken several years for developers to learn how to squeeze performance from the Sony PS2 architecture.
3) Computer gaming is a big industry, so there will be sufficient commercial interests in the project
4) Specfic target constituency. It's not as important to have broad library support compared to a language targeted for general purpose use. Our baby language needs a niche to survive and prosper. Most of the goals mentioned do seem to fit this market.
5) Avoiding hard things. Programming for the business processes in the real world is hard. It's full of non-strictly typed interactions and full of side-effects. LISP picked up a bad name because AI was too nebulous an aim.
6) Cell processors - we need new languages for them. Once proven, there is a general future for this language.
7) The coolness factor associated with computer games can accelerate the adoption of a language
