3 ms·

https://en.wikipedia.org/wiki/Xerox_Alto https://en.wikipedia.org/wiki/Xerox_Alto Looks like Alan Kay's baby is loaded with 96-512kb of memory. So it's not a

https://en.wikipedia.org/wiki/Xerox_Alto https://en.wikipedia.org/wiki/Xerox_Alto Looks like Alan Kay's baby is loaded with 96-512kb of memory. So it's not a challenge - but seems like an appropriate target. Mind you the Alto was primo - Kay was prototyping the future using a wonderful resource called cash.

radiowave 8y ago
As J. C. R. Licklider put it: "If it is (or were) true that hardware economics double every two years, and if it's true that it takes eight or ten years to get something from the idea stage out into use, it says the researcher who begins working on something has got to be a factor of 16 or 32 ahead of the time in his lavish expenditure of computer resources to do his work, and it takes really enlightened sponsors and supporters to make that possible. And usually it isn't."
aidenn0 8y ago
I've been told that MIT managed to convince DARPA to buy them a full address space (~1MB) of core for a PDP-10 in the 70s. I have no idea how much that cost, but it had to be absurdly expensive. Granted this was a timeshare system, not a workstation, so that makes a huge difference. Also I was talking with someone who wanted to implement a lisp for a microcontroller and we came to the conclusion that it had slightly less storage than the IBM that the original LISP was implemented on in the late 50s, which had a combination of drum and core.
- pjmlp 8y ago
  Which is why when people nowadays say that languages with runtime don't fit, that is mostly urban myths as many micro-controllers (naturally not all of them) are like supercomputers in regard to those systems at Xerox and IBM.
- eesmith 8y ago
  https://hblok.net/blog/storage/ https://hblok.net/blog/storage/ says it would have been about $100,000. I don't know if those are in constant dollars, but I suspect they are 1975 dollars. If so, then $100K(1975) is $500K(2018). Edit: I found a reference to the 1MB PDP-10 at MIT in 1979! https://www.filfre.net/2012/01/zil-and-the-z-machine/ https://www.filfre.net/2012/01/zil-and-the-z-machine/ mentions "Zork, a massive game that consumed 1 MB of memory on the PDP-10".

radiowave 8y ago

As J. C. R. Licklider put it: "If it is (or were) true that hardware economics double every two years, and if it's true that it takes eight or ten years to get something from the idea stage out into use, it says the researcher who begins working on something has got to be a factor of 16 or 32 ahead of the time in his lavish expenditure of computer resources to do his work, and it takes really enlightened sponsors and supporters to make that possible. And usually it isn't."

aidenn0 8y ago

I've been told that MIT managed to convince DARPA to buy them a full address space (~1MB) of core for a PDP-10 in the 70s. I have no idea how much that cost, but it had to be absurdly expensive. Granted this was a timeshare system, not a workstation, so that makes a huge difference. Also I was talking with someone who wanted to implement a lisp for a microcontroller and we came to the conclusion that it had slightly less storage than the IBM that the original LISP was implemented on in the late 50s, which had a combination of drum and core.

pjmlp 8y ago
Which is why when people nowadays say that languages with runtime don't fit, that is mostly urban myths as many micro-controllers (naturally not all of them) are like supercomputers in regard to those systems at Xerox and IBM.
eesmith 8y ago
https://hblok.net/blog/storage/ https://hblok.net/blog/storage/ says it would have been about $100,000. I don't know if those are in constant dollars, but I suspect they are 1975 dollars. If so, then $100K(1975) is $500K(2018). Edit: I found a reference to the 1MB PDP-10 at MIT in 1979! https://www.filfre.net/2012/01/zil-and-the-z-machine/ https://www.filfre.net/2012/01/zil-and-the-z-machine/ mentions "Zork, a massive game that consumed 1 MB of memory on the PDP-10".