Right. In 1990ish I put 6 monitors on a macII 1 color and 5 19inch monocrome, just because we had them around and unused that week. AMost of the monitors were just unusable, even if all you wanted to do was open 27 pages of online documentation. The only way to do is was to open the document on the screen with the toolbar and drag it to the right monitor. But to try to do something other than scroll or enter text, forget it. Another case where optimizing for one case totally screws up other cases. Here we have one that worked well in 1984 with a tiny monitor, but as display technology evolved became less useful, then a real impediment.
In a way it surprises me because macs had multiple screen capabilities built into them from the macII onward.
What foolishness the commentary on this is. SInce the Flying Spaghetti Monster sees the fit of this Museum, there must be some humor in it. We must find that.
What could be funnier than asserting that the universe is less 6000 years old? All the rocks show the false evidence and thus must be refuted. But of course knowing better scientists have little reason to argue with them. I have tried to convince both sides of the truth, but meet with disbelief and nothing can be done to sway the faithful. This is the nature of religion. Revealed truth cannot be overturned by facts. If one can't smile at this, there is no hope for them. They will be condemned to a life of serious, humorless plodding negativity.
I of course believe in creation 100 years ago. Who can deny that our ancestors have caused the present? What happened before the creation is subject only to speculation, no one I know has a clear recollection of those events.
How can the orthogonality between these truths be reconciled. I laugh at the thought. What could be more intersting than the embodied delusions of an influential sect in the USA? Can there be some insight into the personality structure of believers in the exhibits there?
Doth complain too much. I'm at 731MB virtual 438MB resident, so what's your beef? Now admittedly I have 19 windows with a total of perhaps 250 tabs, but this is only about twice my normal usage.
How else do you keep 2 projects going, with all the required reference materials, and red slashdot at the same time?
Oh come on, my 50 year old paperbacks are a bit brown and brittle but
quite readable and look like they will last another 50 years. I'm sure they aren't acid free, and admittedly they are in a lot worse shape than 150 or 200 years old books that I store in the same conditions. It doesn't help to exaggerate.
>Figuring: 80 watt savings, 9 hours/day, 227 days/year, and 8 cents/kwh, I come up with $13/year savings. Of course >the price difference between the two is getting pretty small so it could pay for the additional cost (just not >total cost).
Try 24 * 356 * 24 cents/kwh = $168.19 This is much more characteristic of my usage. My complaint is that my 2X21inch CRT's run at 1824x1424 each and I can't find a replacement that good. Perhaps this calculation will make me look harder!
Reverse a linker list in place. Thats a hard problem because: Why in place? What are the limitations: 1) is there a limitation on memory? on stack? 2) what are the virtual memory constraintsIn fIn Fortran, traversing a linked list gratuiously can really screw things up. On a garbage collector I once debugged, the linked list included a lot of memory, you didn't wnat to traverse most of it at once, just a little (it was an incremental gc). 3) what are the relivant optimization priorities. This doesn't matter if the list is small, but there are tricks if the list is large.
In fact I don't remember a good way off the top of my head, but my solution would depend on the answers to these and other restrictions, threads, multi processor, language...
Discussing these would probably give me anough time to reinvent the solution in the background.
I don't resent this kind of questions, it's the language and library specific ones that annoy me. I've learned and forgotten so many languages and dialects that having to learn one for a job is no problem. For an interview however, it's barely worth remembering details, let alone a large set of libraries. Java for instance is a very simple language, the set of stuff that it scomes with is unending, and probably grows faster than a mortal can read, let alone be familar with.
Better yet, don't ask just rewrite it. I disclose my current project by generic type ie. rocket & rocket engine designs, advanced hypertext systems &algorithms including novel interfaces, robotics, etc. Any to these that may be conflicting, I can specify in more detail. I've never had a problem with this, although once my employer did. They were being acquired and need me to specify that none of mine confilcted with any of theirs. At that point arguing about specifics and not potentials, there was no problem. This is usually the case, if you have them, they are quite reasonable, and don't try to steal anything.
I've got a lot of apple label computers that I might want to modify to run OSX In particular a Mac II and an original 128k Mac. Just a little upgrade to the mb and the cpu and off I go!
I bought my first Sun 100 (serial #82 1M ram 80M disk) later upgraded to Sun 100U in Aug 1982. When upgraded to bsd 4.3 in 1983 it had a GUI, so when the Lisa came out, so what? Admittedly, it cost over $20,000 but Moore's was in effect so, the Lisa a year later was about the same at $10,000. In 1984 when the Mac came out at about $2,500, it was below the price curve by 50% but it didn't have a disk! The Lisa had great software, what there was of it. The Sun had a real Unix, what more could you ask for?
All of the above machines had some version of the Motorola 68000, so where other people saw a Mac, I saw a very crippled Sun, needless to say I wasn't impressed, and I'm still not impressed.
I'd say the Sun is much more the model of what Linux is today, for better or worse.
Stalin had set an invasion date for Japan, after this announcement.
London, Aug., 8, 1945 - Foreign Commissar Molotoff's (sic) announcement of the declaration of war, as broadcast by Moscow, follows:
On Aug. 8, People's Commissar for Foreign Affairs of the U.S.S.R. Molotoff received the Japanese Ambassador, Mr. Sato, and gave him, on behalf of the Soviet Government, the following for transmission to the Japanese Government:
"After the defeat and capitulation of Hitlerite Germany, Japan became the only great power that sill stood for the continuation of the war.
"The demand of the three powers, the United States, Great Britain and China, on July 26 for the unconditional surrender of the Japanese armed forces was rejected by Japan, and thus the proposal of the Japanese Government to the Soviet Union on mediation in the war in the Far East loses all basis.
"Taking into consideration the refusal of Japan to capitulate, the Allies submitted to the Soviet Government a proposal to join the war against Japanese aggression and thus shorten the duration of the war, reduce the number of victims and facilitate the speedy restoration of universal peace.
"Loyal to its Allied duty, the Soviet Government has accepted the proposals of the Allies and has joined in the declaration of the Allied powers of July 26.
"The Soviet Government considers that this policy is the only means able to bring peace nearer, free the people from further sacrifice and suffering and give the Japanese people the possibility of avoiding the dangers and destruction suffered by Germany after her refusal to capitulate unconditionally.
"In view of the above, the Soviet Government declares that from tomorrow, that is from Aug. 9, the Soviet Government will consider itself to be at war with Japan." Here is the cronology. Potsdam Conference--Truman, Churchill, Atlee (after July 28), Stalin establish council of foreign ministers to prepare peace treaties; plan German postwar government and reparations (July 17-Aug. 2). A-bomb dropped on Hiroshima by U.S. (Aug. 6). USSR declares war on Japan (Aug. 8). Nagasaki hit by A-bomb (Aug. 9). Japan agrees to surrender (Aug. 14). V-J Day--Japanese sign surrender terms aboard battleship Missouri (Sept. 2).
I haven't been able to find the reference to the planned Russion invasion date, in time to put into this post, but I've read it somewhere. There is no doubt that keeping Russia out of Japan was a if not the major factor in the timing of the decision to drop the bomb.
Short version: GNU needed some heavy lifting. Some enlightened members of corporate America stepped up to the plate.
I had to snort at this. The Cygnus founders, John Gilmore, Gumby and Michael Tieman as "enlightened members of corporate America" ? Enlightened, I'll grant, but members of corporate america? Well perhaps defacto they are, but only because they succeded.
Most search engine sites have a lot more load searching the net than answering queries. The net is still the same size but there are only a few queries a day. Google is a different kind of beast.
Google has datacenters full of machines spread over the world answering queries, but still needs to spider and archive the web once. Their query servers have a complete index. They are packaged about 40 to a rack, and the racks replicated as needed.
A search engine spends almost all of it's time pulling indexes off of the disk. That's how the factor of 200,000 comes in. I forget the size of the index I was quoted, but it was in the range of 40Gb within a factor of two. I expect that puting the index in RAM will shift the bottlenecks elsewhere from the index but the breakeven performance gain is probably ($40,000/server with 40Gb DRAM)/($2000/server) = 10 and a lot of other factors weighing on the side of fewer servers.
Note that transitory data failures can be refresed out of disk. And that with over 8000 servers (an old figure) they would still have lots of servers to distribute.
I once had a 386 that I cooled by having a sponge
on the cpu and adding water to that directly. Not too surprisingly it worked and caused no problems.
There is too much worry and sillyness about cooling these days. A quick search shows a bunch of freezers under $500, just plonk the whold damn machine in that and use it for your remote cruncher. No need to worry about condensation, the machine is the warmest thing in sight so all the water freezes out elsewhere. I'll be trying this soon, as I expect to be building a beowolf cluster and this may be the cheapest hack.
The next step is just to imerse the whold damn thing in LN2 (lowly cooling down to that level)
the major expected problem is that many plastics get brittle, an be sure to have good venting,
maybe I'll try LOX it's safer!
Ted Nelson's book Computer Lib/Dream Machines, was published in 1974. It fleshed out a lot of hypertext notions, the microsoft press edition just had a few updates about the current state of computing. However for real prior art look at his published papers http://xanadu.com.au/archive/bibliography.html
I started work with Ted in 1976, and have a box of printout labeled "Patent Fodder" from 1979. Ted of course has complete audio and video tapes and notes from all conversations related to hypertext (not an exageration) dating all the way back to 1960.
Doug Engelbart's 1965 hypertext presentation is clearly prior art, the video is widely available.
Slashdot Mirror
Right. In 1990ish I put 6 monitors on a macII 1 color and 5 19inch monocrome, just because we had them around and unused that week. AMost of the monitors were just unusable, even if all you wanted to do was open 27 pages of online documentation. The only way to do is was to open the document on the screen with the toolbar and drag it to the right monitor. But to try to do something other than scroll or enter text, forget it. Another case where optimizing for one case totally screws up other cases. Here we have one that worked well in 1984 with a tiny monitor, but as display technology evolved became less useful, then a real impediment.
In a way it surprises me because macs had multiple screen capabilities built into them from the macII onward.
What foolishness the commentary on this is. SInce the Flying Spaghetti Monster sees the fit of this Museum, there must be some humor in it. We must find that.
What could be funnier than asserting that the universe is less 6000 years old? All the rocks show the false evidence and thus must be refuted. But of course knowing better scientists have little reason to argue with them. I have tried to convince both sides of the truth, but meet with disbelief and nothing can be done to sway the faithful. This is the nature of religion. Revealed truth cannot be overturned by facts. If one can't smile at this, there is no hope for them. They will be condemned to a life of serious, humorless plodding negativity.
I of course believe in creation 100 years ago. Who can deny that our ancestors have caused the present? What happened before the creation is subject only to speculation, no one I know has a clear recollection of those events.
How can the orthogonality between these truths be reconciled. I laugh at the thought. What could be more intersting than the embodied delusions of an influential sect in the USA? Can there be some insight into the personality structure of believers in the exhibits there?
than a few moments you may see that too.
Doth complain too much. I'm at 731MB virtual 438MB resident, so what's your beef? Now admittedly I have 19 windows with a total of perhaps 250 tabs, but this is only about twice my normal usage.
How else do you keep 2 projects going, with all the required reference materials, and red slashdot at the same time?
Oh come on, my 50 year old paperbacks are a bit brown and brittle but
quite readable and look like they will last another 50 years. I'm sure they aren't acid free, and admittedly they are in a lot worse shape than 150 or 200 years old books that I store in the same conditions. It doesn't help to exaggerate.
>Figuring: 80 watt savings, 9 hours/day, 227 days/year, and 8 cents/kwh, I come up with $13/year savings. Of course >the price difference between the two is getting pretty small so it could pay for the additional cost (just not >total cost).
Try 24 * 356 * 24 cents/kwh = $168.19 This is much more characteristic of my usage. My complaint is that my 2X21inch CRT's run at 1824x1424 each and I can't find a replacement that good. Perhaps this calculation will make me look harder!
Reverse a linker list in place. Thats a hard problem because:
...
Why in place? What are the limitations:
1) is there a limitation on memory? on stack?
2) what are the virtual memory constraintsIn fIn Fortran, traversing a linked list gratuiously can really screw things up. On a garbage collector I once debugged, the linked list included a lot of memory, you didn't wnat to traverse most of it at once, just a little (it was an incremental gc).
3) what are the relivant optimization priorities. This doesn't matter if the list is small, but there are tricks if the list is large.
In fact I don't remember a good way off the top of my head, but my solution would depend on the answers to these and other restrictions, threads, multi processor, language
Discussing these would probably give me anough time to reinvent the solution in the background.
I don't resent this kind of questions, it's the language and library specific ones that annoy me. I've learned and forgotten so many languages and dialects that having to learn one for a job is no problem. For an interview however, it's barely worth remembering details, let alone a large set of libraries. Java for instance is a very simple language, the set of stuff that it scomes with is unending, and probably grows faster than a mortal can read, let alone be familar with.
Better yet, don't ask just rewrite it. I disclose my current project by generic type ie.
rocket & rocket engine designs, advanced hypertext systems &algorithms including novel interfaces, robotics, etc. Any to these that may be conflicting, I can specify in more detail.
I've never had a problem with this, although once my employer did. They were being acquired and need me to specify that none of mine confilcted with any of theirs. At that point arguing about specifics and not potentials, there was no problem. This is usually the case, if you have them, they are quite reasonable, and don't try to steal anything.
http://udanax.com/
http://halfwaytoanywhere.com/
I've got a lot of apple label computers that I might want to modify to run OSX
In particular a Mac II and an original 128k Mac. Just a little upgrade to the mb and the cpu and off I go!
I bought my first Sun 100 (serial #82 1M ram 80M disk) later upgraded to Sun 100U in Aug 1982. When upgraded to bsd 4.3 in 1983 it had a GUI, so when the Lisa came out, so what? Admittedly, it cost over $20,000 but Moore's was in effect so, the Lisa a year later was about the same at $10,000. In 1984 when the Mac came out at about $2,500, it was below the price curve by 50% but it didn't have a disk! The Lisa had great software, what there was of it. The Sun had a real Unix, what more could you ask for?
All of the above machines had some version of the Motorola 68000, so where other people saw a Mac, I saw a very crippled Sun, needless to say I wasn't impressed, and I'm still not impressed.
I'd say the Sun is much more the model of what Linux is today, for better or worse.
Stalin had set an invasion date for Japan, after this announcement.
London, Aug., 8, 1945 - Foreign Commissar Molotoff's (sic) announcement of the declaration of war, as broadcast by Moscow, follows:
On Aug. 8, People's Commissar for Foreign Affairs of the U.S.S.R. Molotoff received the Japanese Ambassador, Mr. Sato, and gave him, on behalf of the Soviet Government, the following for transmission to the Japanese Government:
"After the defeat and capitulation of Hitlerite Germany, Japan became the only great power that sill stood for the continuation of the war.
"The demand of the three powers, the United States, Great Britain and China, on July 26 for the unconditional surrender of the Japanese armed forces was rejected by Japan, and thus the proposal of the Japanese Government to the Soviet Union on mediation in the war in the Far East loses all basis.
"Taking into consideration the refusal of Japan to capitulate, the Allies submitted to the Soviet Government a proposal to join the war against Japanese aggression and thus shorten the duration of the war, reduce the number of victims and facilitate the speedy restoration of universal peace.
"Loyal to its Allied duty, the Soviet Government has accepted the proposals of the Allies and has joined in the declaration of the Allied powers of July 26.
"The Soviet Government considers that this policy is the only means able to bring peace nearer, free the people from further sacrifice and suffering and give the Japanese people the possibility of avoiding the dangers and destruction suffered by Germany after her refusal to capitulate unconditionally.
"In view of the above, the Soviet Government declares that from tomorrow, that is from Aug. 9, the Soviet Government will consider itself to be at war with Japan."
Here is the cronology.
Potsdam Conference--Truman, Churchill, Atlee (after July 28), Stalin establish council of foreign ministers to prepare peace treaties; plan German postwar government and reparations (July 17-Aug. 2). A-bomb dropped on Hiroshima by U.S. (Aug. 6). USSR declares war on Japan (Aug. 8). Nagasaki hit by A-bomb (Aug. 9). Japan agrees to surrender (Aug. 14). V-J Day--Japanese sign surrender terms aboard battleship Missouri (Sept. 2).
I haven't been able to find the reference to the planned Russion invasion date, in time to put into this post, but I've read it somewhere. There is no doubt that keeping Russia out of Japan was a if not the major factor in the timing of the decision to drop the bomb.
Short version: GNU needed some heavy lifting. Some enlightened members of corporate America stepped up to the plate.
I had to snort at this. The Cygnus founders, John Gilmore, Gumby and Michael Tieman as "enlightened members of corporate America" ? Enlightened, I'll grant, but members of corporate america? Well perhaps defacto they are, but only because they succeded.
This one has two levels deep of nonintuitveness.
Most search engine sites have a lot more load searching the net than answering queries. The net is still the same size but there are only a few queries a day. Google is a different kind of beast.
Google has datacenters full of machines spread over the world answering queries, but still needs to spider and archive the web once. Their query servers have a complete index. They are packaged about 40 to a rack, and the racks replicated as needed.
A search engine spends almost all of it's time pulling indexes off of the disk. That's how the factor of 200,000 comes in. I forget the size of the index I was quoted, but it was in the range of 40Gb within a factor of two. I expect that puting the index in RAM will shift the bottlenecks elsewhere from the index but the breakeven performance gain is probably ($40,000/server with 40Gb DRAM)/($2000/server) = 10 and a lot of other factors weighing on the side of fewer servers.
Note that transitory data failures can be refresed out of disk. And that with over 8000 servers (an old figure) they would still have lots of servers to distribute.
I once had a 386 that I cooled by having a sponge
on the cpu and adding water to that directly. Not too surprisingly it worked and caused no problems.
There is too much worry and sillyness about cooling these days. A quick search shows a bunch of freezers under $500, just plonk the whold damn machine in that and use it for your remote cruncher. No need to worry about condensation, the machine is the warmest thing in sight so all the water freezes out elsewhere. I'll be trying this soon, as I expect to be building a beowolf cluster and this may be the cheapest hack.
The next step is just to imerse the whold damn thing in LN2 (lowly cooling down to that level)
the major expected problem is that many plastics get brittle, an be sure to have good venting,
maybe I'll try LOX it's safer!
Ted Nelson's book Computer Lib/Dream Machines, was published in 1974. It fleshed out a lot of hypertext notions, the microsoft press edition just had a few updates about the current state of computing. However for real prior art look at his published papers http://xanadu.com.au/archive/bibliography.html
I started work with Ted in 1976, and have a box of printout labeled "Patent Fodder" from 1979. Ted of course has complete audio and video tapes and notes from all conversations related to hypertext (not an exageration) dating all the way back to 1960.
Doug Engelbart's 1965 hypertext presentation is clearly prior art, the video is widely available.