which would you prefer?

a) Make something very useful, but it is never released because the firm you work for cannot use it for profit at the time being, and it is sealed away to prevent it from being used by competition, forever.

b) Make something that is free and available, but is useless and not used by anyone, forever.

c) Make something that is free and available, and very useful, but too different to interface with the technology people use, and is not used by anyone, forever.

everything

pururin~

>>1
a) lisp machines
b) losethos
c) lisp

but I'm not anyone! >:<

>>4
wow

http://en.wikipedia.org/wiki/Worse_is_better

>>1
I'd go with C. Although there are other important questions I would ask, like "What will I learn by completing this project?", which would affect my decision.

>>1
very useful, but too different to interface with the technology people use

I always wonder how people manage to convince themselves in this bullshit. "This is a thing that gives you godlike powers, but alas it can't make C calls. That's too hard for it. It's not that powerful". What the fucking fuck?

>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>4>>

i'd prefer dubs

>>9

I was thinking of things like advanced and revolutionary computer architectures that will never gain acceptance because no one will use anything other than a faster better version of what is currently in use.

>>11
nice dubs

>>12
An advanced and revolutionary general purpose computer architecture that can't emulate a faster better version of what is currently in use is not really advanced at all. That's not what "advanced" means. If a thing is better than other things, then it is better, if you call it better but it's not really better, then it is not better. That's a tricky concept to understand for an adolescent who is still ears deep in magical thinking, sure.

Also, things that really are better have no problems finding users, even when they are really different, like GPUs for example. That's another aspect of lispfags and other followers of technologies "too good for this shitty world" -- how they manage to ignore all the success stories.

>>12
Why would you want to use something that isn't faster and better than what is currently in use? Only a retard would do that.

>>13
OpenCL is becoming the real no-bullshit champion of the massively parallel age.

Too bad faggots on the Internet haven't realized it yet, they still think their Haskell, Scala, Clojure or Go is the wave of the future.

Once OpenCL 2.0 has the much promised standardized bytecode intermediate representation, there will no longer be an excuse for esoteric language connoisseurs to target OpenCL with their toy projects.

>>14
I meant...

here will no longer be an excuse for esoteric language connoisseurs to not target OpenCL with their toy projects.

>>13
Also, things that really are better have no problems finding users, even when they are really different, like GPUs for example. That's another aspect of lispfags and other followers of technologies "too good for this shitty world" -- how they manage to ignore all the success stories.
But how did GPUs gain acceptance? By providing graphics drivers for popular platforms such as Win32, object-oriented APIs and shading languages with assembly or C-like syntax. GPUs are scalable and integrate seamlessly with existing technologies.
This is why Haskell and Lisp will never take off and why bashing X11 has become the freedesktop.org national sport: they don't integrate at all within a stateful UNIX environment centered around dynamic libraries and small scripts.

>>14
Holy shit, I had no idea anything like OpenCL even existed

>>16
Seamlessly? LOL

>>17
OpenMP is cleaner.

>>18
OpenMP has zero support for heterogenous computing devices or grid parallelism and it lacks SIMD vector intrinsics for useful math instructions built into modern hardware.

OpenMP is built on 1990s ideas about concurrency.

>>13

things can eventually become too different to easily or efficiently emulate the other. There is no strict hierarchy of complexity or design. Nodes can share common ancestries, but not have much in common directly. The requirement of being able to efficiently run assembly for a legacy architecture may be too restrictive of cumbersome.

>>16

GPUs originally came about to satisfy demand for fast programmable graphics effects. Then people looking for resources for general parallel computing started using them because they were relatively cheap and already available, as they were then a regular component in a desktop computer. Using the GPU for regular programming is a creative adaptation of something that is already in use. But there wasn't really anything comparable to a GPU before its use, so I don't think that is a good example. There was no cost of moving to a GPU. With an alternative CPU architecture, millions of applications with machine dependent code will not be able to run on it without effort from developers. And the cost of the world porting their code to an architecture with a better design may be worse than using a suboptimal but popular and targeted architecture.

>>19

OpenMP is more useful than OpenCL in some situations.

a. because it includes a paycheck

Currently writing on a c). Maybe I can earn a bit of money with training and consultancy.