(I would be delighted if this were upgraded to an "Ask Slashdot.")
(I think that) The fact that Linux has made it this far is just a mouth-watering *hint* of the progress/impact we could make with an entirely NEW operating system SPECIFICALLY designed for open-source contribution.
Let's face it, most of that "open" C code is butt-complex and only a few people can really get into it and contribute, either to the OS itself, or more importantly to the applications that run on top of it. I view this as a serious problem: traditional programming is just too damn hard. (Unix and C are *1960's* technology for Pete's sake! Who here has a slide-rule?)
I'm trying to imagine an operating system that offers a whole new class of services that makes program development an order of magnitude easier, but I need your help because I just can't put it all together in my own mind by myself. To get the pump primed, here are some dissociated ideas/clues that have crossed my mind:
o Now we have files and variables, both static and dynamic. Could some kind of database system take all these functions over somehow and make it all persistent and easier to program?
o C made Unix. Does this new operating system need a whole new language to enable/justify its existence? What would it be? What would it do? Let's face it: the only machine-code exectuable should be the OS, not the applications.
o Dammit there's just GOT to be a better way to program user interfaces, SOMEHOW. KDE and GNOME are great, but we'll still look back on them as kluges; (brilliantly executed) duct-tape wrapped around the hocked-together rattle-trap junk heap that is X. What's the next higher-level way of doing interfaces that's *way* easier on the programmer (and naturally more compute intensive)?
Look at it this way: we won't be writing X-windows programs in C 50 years from now. If we're smart, ambitious and a little lucky, what will we have by then?
Etc. etc. etc. Okay your turn. Go for it! --Craig Meyer meyer@aerovironment.com
Slashdot Mirror
(I would be delighted if this were upgraded to an "Ask Slashdot.")
(I think that) The fact that Linux has made it this far is just a mouth-watering *hint* of the progress/impact we could make with an entirely NEW operating system SPECIFICALLY designed for open-source contribution.
Let's face it, most of that "open" C code is butt-complex and only a few people can really get into it and contribute, either to the OS itself, or more importantly to the applications that run on top of it. I view this as a serious problem: traditional programming is just too damn hard. (Unix and C are *1960's* technology for Pete's sake! Who here has a slide-rule?)
I'm trying to imagine an operating system that offers a whole new class of services that makes program development an order of magnitude easier, but I need your help because I just can't put it all together in my own mind by myself. To get the pump primed, here are some dissociated ideas/clues that have crossed my mind:
o Now we have files and variables, both static and dynamic. Could some kind of database system take all these functions over somehow and make it all persistent and easier to program?
o C made Unix. Does this new operating system need a whole new language to enable/justify its existence? What would it be? What would it do? Let's face it: the only machine-code exectuable should be the OS, not the applications.
o Dammit there's just GOT to be a better way to program user interfaces, SOMEHOW. KDE and GNOME are great, but we'll still look back on them as kluges; (brilliantly executed) duct-tape wrapped around the hocked-together rattle-trap junk heap that is X. What's the next higher-level way of doing interfaces that's *way* easier on the programmer (and naturally more compute intensive)?
Look at it this way: we won't be writing X-windows programs in C 50 years from now. If we're smart, ambitious and a little lucky, what will we have by then?
Etc. etc. etc. Okay your turn. Go for it!
--Craig Meyer
meyer@aerovironment.com