There's a lot to computing! I think I will discuss the field as split into two distinct parts: achedemic studies that further the field and create new and exciting directions; and engineering where building for robustness and predictability are paramount.
In achedemia, the individual does a lot alone. New depths are probed and exciting issues are researched. There is rarely the scope for putting several people together to work on one research point. The preparation for this is the achedemic study you do alone at university. Learning the basics of search/sort algorithms, data structures, language constructs, IO and communications, image or sound processing, etc. These things are best done alone and can be tested individually to give results. No matter where you go with computing, you'll need these basics.
In engineering, you work in a large team to create a robust and predictable system. Each member brings something to the team, even if it's "just" key-tapping programming. Essentially, nothing is particularly new and everything ought to be tried and at the very least tested. Time is important too, so using well known techniques is mandatory and research almost out of the question. If a University is preparing someone for 'real' work the group projects are the order of the day. Assessment becomes more like a workplace review, peer-reviews are useful too.
In my experience, these two distinct items are often mixed up and learning/research tasks are hindered by group politics. Groupwork tasks are blown out of the water because the research aspects overrun. Perhaps I'm saying that my experience is that I was often taught by stealth, not really knowing what I was supposed to be learning in a particular excersise/module/etc.
Mebbe the system needs to change, mebbe I should have taken more notice of what was going on.. but I hope Uni-starters get something from these ramblings to help deal with the apparent mixup I felt:)
I left 3D stuff a while ago, utterly bemused with the various complexities of D3D and OGL. There was a lot of legacy stuff in there and little fiddly bits that were in 'just because'. Now I presume there will be a refactor in the light of various progresses in the field. I believe refactoring to be a GOOD THING (tm).
OGL2 will be 100% backwards compatible. Well, that could be either good or bad.. If we end up with so much legacy kak the new version will just be a bigger more combersome beast than before. Remember the DOS/windows incarnations.. eugh.. But it can still be done well. Refactoring need not mean throwing away + starting from scratch.
I've done some Java3D.. that's fairly nice, fairly fast and I hope some of those lessons are fed into OGL2, along with things that other lib-developers have learned. Afterall, OGL ought to be faster and more widely used, given it's foothold (however weak some might percieve it to be).
I really hope that OGL2 will become ther real OPEN standard for 3d programming. Various little-libs have surfaced to hide the old OGL/Dx, but it's time for those to be set aside as the core tech becomes the thing to use.
Quick note on 'extras', should OGL include libs for keyboard/mouse/sound/controller access? Well, I honestly don't know the answer, but my guess is that there should be hooks for plugging such things in, but I also think that OGL should be focussed on the job in hand: rendering fast, amazing 3D graphics.
The FAT file system is interesting. On a FAT file system this actually did perform compression. The FAT (table) is fixed size and takes up the same disk space however many entries are in it (up to it's limit). Thus the hiding mechanism takes advantage of an area of the file system that isn't normally used.
Still, I don't think he ever expected the $5000, and expected to loose the $100 too. Fair enough I guess.. some ppl are willing to pay a price to prove a point (look at OSS:)
Drivers and applications are a major problem to all 'new' OS's. Apple seem to have done a lot of work for the applications. However, OSX is basically BSD, so porting the GUI of other apps is the main problem. Also there's quite a few drivers for BSD, I wonder if these can be re-used? I would presume so.
Releasing for x86 would add one extra software problem: you can't release a single binary, they would have to release:
VRML is Virtual Reality Modelling Language. It fitted in with other web-based technologies when last I used it. It's not a replacement for anything.
What it is good for is the description of a 3D virtual world, with various APIs to interface into that. It was fairly regular and fairly complete in the version I used, and additions may or may not improve things, but please don't get confused as to what it is.
As to ppl re-writing their sites, they already do that with flash. As to computers not being capable, they are (see Quake). As to flexibility, it is (hey, I wrote a simple 3D 1pp multi-user environment using it in '97/8).
It's not as great at defining detailed 3D worlds as the pov-ray file format, it's not as wazzy as quake, but it is very handy at it's intended, general purpose market. Hopefully we'll see better and better browsers coming along.
from a conversation with a friend of mine: Hmm, was having a look at that earlier. My first reaction was "Bullshit!", but thought I'd give the guy the benefit of the doubt long enough to read the page. Now I've done that my considered opinion it that it's bullshit. The basic premise (cos I don't know where the equations come from) seems to be that you can beam photons out from your ship, and close to the ship, where the photon density is above a certain level, gravitons won't be able to get through. Hence you won't be affected by gravity. Now I can see a few problem, both theoretical and practical, with this idea. I wouldn't stand up in court with these, but I think they're right.
a graviton is still a theoretical particle and hence it's properties are unknown. Some have been theorised but I'd be willing to bet there are 2 theories for every property.
leading on from 1, this person assumes that photons will interact with gravitons. If they don't then the two sail straight through/passed one another and your ship stops.
assuming the two do interact, it seems reasonable (from my woeful understanding of wave theory) that you would have to frequency match your photons to gravitons in some way. This means you'd need complete detailed information of the gravitational field you're travelling through.
point sources of photons (or anything else) don't exist.
if you don't have your 'gravity maps' (from 3) then you'd need blanket coverage. So you'd be broadcasting photons across a huge spectrum.
stars kick out huge numbers of photons all the time, but somehow, they're not shielded from the gravational forces of their planets?
of course, the photon density of a star may not be high enough. This doesn't bode well for your passengers who are being bombarded by photons across the spectrum (see 5), including infra-red and microwave. They'd be cooked in an instant.
another practical problem is getting the photons out through the skin of the ship in the first place....
and on the superluminary travel, there's one small problem there. As soon as you pass light speed, you leave your shield behind:)
Points 3, 5 and to some extent 7 are a little iffy, but the rest are fairly conclusive. As someone once said, you can prove anything with equations...
isn't the power of the Dreamcast (and the others) in the fact that they:
Don't use the naff x86 archetecture (still full of 60's technology)
have multi-processing (a-la Amiga) where there's a processor for wazzy grafix, one for sound spatialisation, one for AI, etc etc etc..
I really don't think that an PC-based computer will be worth the investment as a games machine. This said, there's a lot of games for PC, so the only hope I have is that this will force the big-3 to up the stakes + make their machines/games even better.. (mebbe even open them to us programmers:)
In a lot of trials, juries are a waste of space. They do not descide whether someone is guilty or not. Juries descide issues of FACT! e.g. would the REASONABLE MAN have whalloped him over the head? the jury is the reasonable man + has to descide whether they would have done the whalloping. Many criminals plead not-guilty, get to trial (waste a lot of cash) and then change their plea. Further to this, in big fraud trials + stuff, a jury is a major problem. Big trials are complex, and the average man isn't informed or clever enough to cope with it. Don't all go american and say "There's no jury, it's not a fair trial!" 'cos it simply ain't true finally, we'll be having a Human Rights Act soon which will make the sensationalist interpretation of this illegal (meaning, the interpretation given will not be valid).
If you have nothing to hide, then what have you to be scared of? The only time this is a problem is when you ARE guilty of something. Should ppls have right to lie or "plead the 5th" just to get away with a crime? This is not as simple (nor as sensationalist) as the article declares.
Slashdot Mirror
There's a lot to computing! I think I will discuss the field as split into two distinct parts: achedemic studies that further the field and create new and exciting directions; and engineering where building for robustness and predictability are paramount.
In achedemia, the individual does a lot alone. New depths are probed and exciting issues are researched. There is rarely the scope for putting several people together to work on one research point. The preparation for this is the achedemic study you do alone at university. Learning the basics of search/sort algorithms, data structures, language constructs, IO and communications, image or sound processing, etc. These things are best done alone and can be tested individually to give results. No matter where you go with computing, you'll need these basics.
In engineering, you work in a large team to create a robust and predictable system. Each member brings something to the team, even if it's "just" key-tapping programming. Essentially, nothing is particularly new and everything ought to be tried and at the very least tested. Time is important too, so using well known techniques is mandatory and research almost out of the question. If a University is preparing someone for 'real' work the group projects are the order of the day. Assessment becomes more like a workplace review, peer-reviews are useful too.
In my experience, these two distinct items are often mixed up and learning/research tasks are hindered by group politics. Groupwork tasks are blown out of the water because the research aspects overrun. Perhaps I'm saying that my experience is that I was often taught by stealth, not really knowing what I was supposed to be learning in a particular excersise/module/etc.
Mebbe the system needs to change, mebbe I should have taken more notice of what was going on.. but I hope Uni-starters get something from these ramblings to help deal with the apparent mixup I felt :)
I left 3D stuff a while ago, utterly bemused with the various complexities of D3D and OGL. There was a lot of legacy stuff in there and little fiddly bits that were in 'just because'. Now I presume there will be a refactor in the light of various progresses in the field. I believe refactoring to be a GOOD THING (tm).
OGL2 will be 100% backwards compatible. Well, that could be either good or bad.. If we end up with so much legacy kak the new version will just be a bigger more combersome beast than before. Remember the DOS/windows incarnations.. eugh.. But it can still be done well. Refactoring need not mean throwing away + starting from scratch.
I've done some Java3D.. that's fairly nice, fairly fast and I hope some of those lessons are fed into OGL2, along with things that other lib-developers have learned. Afterall, OGL ought to be faster and more widely used, given it's foothold (however weak some might percieve it to be).
I really hope that OGL2 will become ther real OPEN standard for 3d programming. Various little-libs have surfaced to hide the old OGL/Dx, but it's time for those to be set aside as the core tech becomes the thing to use.
Quick note on 'extras', should OGL include libs for keyboard/mouse/sound/controller access? Well, I honestly don't know the answer, but my guess is that there should be hooks for plugging such things in, but I also think that OGL should be focussed on the job in hand: rendering fast, amazing 3D graphics.
The FAT file system is interesting. On a FAT file system this actually did perform compression. The FAT (table) is fixed size and takes up the same disk space however many entries are in it (up to it's limit). Thus the hiding mechanism takes advantage of an area of the file system that isn't normally used. :)
Still, I don't think he ever expected the $5000, and expected to loose the $100 too. Fair enough I guess.. some ppl are willing to pay a price to prove a point (look at OSS
Drivers and applications are a major problem to all 'new' OS's. Apple seem to have done a lot of work for the applications. However, OSX is basically BSD, so porting the GUI of other apps is the main problem. Also there's quite a few drivers for BSD, I wonder if these can be re-used? I would presume so.
Releasing for x86 would add one extra software problem: you can't release a single binary, they would have to release:
- separate x86 and PPC binaries
- a big combined binary
- release as source.
Ooooh, cat amongst pigeons!VRML is Virtual Reality Modelling Language. It fitted in with other web-based technologies when last I used it. It's not a replacement for anything. What it is good for is the description of a 3D virtual world, with various APIs to interface into that. It was fairly regular and fairly complete in the version I used, and additions may or may not improve things, but please don't get confused as to what it is. As to ppl re-writing their sites, they already do that with flash. As to computers not being capable, they are (see Quake). As to flexibility, it is (hey, I wrote a simple 3D 1pp multi-user environment using it in '97/8). It's not as great at defining detailed 3D worlds as the pov-ray file format, it's not as wazzy as quake, but it is very handy at it's intended, general purpose market. Hopefully we'll see better and better browsers coming along.
- a graviton is still a theoretical particle and hence it's properties are unknown. Some have been theorised but I'd be willing to bet there are 2 theories for every property.
- leading on from 1, this person assumes that photons will interact with gravitons. If they don't then the two sail straight through/passed one another and your ship stops.
- assuming the two do interact, it seems reasonable (from my woeful understanding of wave theory) that you would have to frequency match your photons to gravitons in some way. This means you'd need complete detailed information of the gravitational field you're travelling through.
- point sources of photons (or anything else) don't exist.
- if you don't have your 'gravity maps' (from 3) then you'd need blanket coverage. So you'd be broadcasting photons across a huge spectrum.
- stars kick out huge numbers of photons all the time, but somehow, they're not shielded from the gravational forces of their planets?
- of course, the photon density of a star may not be high enough. This doesn't bode well for your passengers who are being bombarded by photons across the spectrum (see 5), including infra-red and microwave. They'd be cooked in an instant.
- another practical problem is getting the photons out through the skin of the ship in the first place....
- and on the superluminary travel, there's one small problem there. As soon as you pass light speed, you leave your shield behind
:)
Points 3, 5 and to some extent 7 are a little iffy, but the rest are fairly conclusive. As someone once said, you can prove anything with equations...- Don't use the naff x86 archetecture (still full of 60's technology)
- have multi-processing (a-la Amiga) where there's a processor for wazzy grafix, one for sound spatialisation, one for AI, etc etc etc..
I really don't think that an PC-based computer will be worth the investment as a games machine.This said, there's a lot of games for PC, so the only hope I have is that this will force the big-3 to up the stakes + make their machines/games even better.. (mebbe even open them to us programmers
In a lot of trials, juries are a waste of space. They do not descide whether someone is guilty or not.
Juries descide issues of FACT! e.g. would the REASONABLE MAN have whalloped him over the head? the jury is the reasonable man + has to descide whether they would have done the whalloping.
Many criminals plead not-guilty, get to trial (waste a lot of cash) and then change their plea.
Further to this, in big fraud trials + stuff, a jury is a major problem. Big trials are complex, and the average man isn't informed or clever enough to cope with it.
Don't all go american and say "There's no jury, it's not a fair trial!" 'cos it simply ain't true
finally, we'll be having a Human Rights Act soon which will make the sensationalist interpretation of this illegal (meaning, the interpretation given will not be valid).
If you have nothing to hide, then what have you to be scared of?
The only time this is a problem is when you ARE guilty of something. Should ppls have right to lie or "plead the 5th" just to get away with a crime?
This is not as simple (nor as sensationalist) as the article declares.