> Name me one useful invention / innovation done > by OSS in the past fifteen years (since 1987)
You are joking, right?
I'll only list very few:
1- Heaps of new, useful languages. To name a few:
- Perl (born in 12/87)
- Python
- Java (is open-source, if not Free).
2- Operating systems
- Linux (certainly an amazing innovation, if not an outright invention)
- Mach (concepts date back from 1985 though)
- all the free BSDs. Lots of innovation there
3- Compiler technology
- Gcc (1988). Gcc is the only truly portable C/C++ compiler (and now Java, Ada, etc) usable for production work. Nothing else touches it.
4- Parallel computation technology
- MPI
- OpenMP
These a minuscule portion of free software inventions and innovations. If you know anything about IT you will acknowledge the importance of the above software packages.
The parent post was just trying to say that *Microsoft* has a tendency not to invent. It has been cheaper for them to let other companies invent new concepts and technologies and buy them outright rather than finance the R&D. This has been true from the very beginning of MS. They did not invent DOS, Gates bought it from someone else.
For more recent examples think of internet explorer, hotmail. Many people contend that MS implemented their TCP stack from the BSD sources (as they were allowed to do).
There is nothing inherently wrong with that, except you have to worry being in a world where no one can challenged the huge behemoth MS has become. If you are a small innovative company with a successful and popular closed-source product, three things can happen to you:
1 - Microsoft ignores you for a while until you become a menace, 2 - Microsoft buys you outright, 3 - Microsoft releases a competing product and more or less destroys you.
For an example of (1) think Adobe. Apparently now Microsoft is interested in taking over PDF, I wonder what will happen. (2) Think hotmail. (3) Netscape, Borland.
Lead shielding is only useful against gamma/X-rays. Against Alpha is it completely overkill (a thin sheet of paper is enough) and with beta (at least high-energy beta), the slowing down of the Beta particles in lead actually *causes* X-ray radiation. This is called `brehmstrahlung' radiation. In fact this is how X-ray machines work.
I have a 1989 286 from Goupil, defunct European PC manufacturer. I used to consult for them. The motherboard is one of their development one, with all the bugs taken out with soldered wires and hand cuts on the PCB, there are dozens of them! It looks like an engineering nightmare, the ultimate Kludge. It runs fine though, DOS obviously.
That's good for you, but what if your job is to do 3D CAD, or animation, or simply scientific visualization that requires accelerated OpenGL?
Unfortunately under Linux you will have a *very* hard time with the open source accelerated 3D drivers. The best hardware you can run that way right now it the ATI Radeon 7500, and that's not great and that's aging rapidly, and there is little prospect of improvement or support for better hardware. I know because I've been there.
My life has become a whole lot simpler thanks to Nvidia, I hate to admit (even though their drivers aren't perfect, far from it).
BTW your argument about `convicing' somebody is the same whether it's open source software or not. In theory it's great that you can patch the source but in practice very few people have the time, skill and resources to do that for *every* application and framework that they use.
It's true that you can generate a seemingly unlimited amount of fuel using fast breeder reactors (that convert U238 to Pu239), but those are more dangerous to operate than conventional PWR reactors; and fast breeder reactors can easily be converted to generate weapons grade fuel. Can you say nuclear proliferation?
There is of course the vexing problem of radioactive wastes. All told nuclear energy is rather pricey.
No problem with that. Unlike fission, fusion is not self-maintaining. There is no foreseable way that a fusion reaction can be maintained continuously here on Earth, let alone get out of hand.
The way Tokamaks and laser fusion reactors work is by doing a series of miniature implosions that consume their fuel, harvest the heat, clean up and start again.
Patents don't forbid people to reproduce what they describe. They only forbid someone from exploiting them commercially. In other words you can freely form patented thoughts in your head but you cannot distribute them (via telepathy, say).
Actually the Earth's axis of rotation does wobble quite a bit, if not by huge amounts. It certainly does so in a measurable way.
The Moon helps stabilize the axis of rotation. If we didn't have the Moon, the Earth could have ended up like Uranus is now: with an axis of rotation almost parallel to the plane of orbit. Seasons would have a different meaning in this context...
- The combination of an AGP+PCI card is the most stable. Xinerama works fine, is well supported, and the display quality is great on both heads. At most you will get accelerated 3D on one of the heads though.
- Xinerama disables DRI, so you will not get accelerated OpenGL at the same time as having two working monitors. You need to revert to single-head for DRI to work.
- Nvidia is the only way (with their proprietary driver) if you want accelerated 3D on both heads. With the current version of the driver everything works fine on reasonably current hardware. I found that the driver is more stable if I enable Nvidia's version of the AGP driver (not agpgart, but nvAGP, it's an option in the XF86Config-4 file).
- often the display quality on the second head on hardware that support it will be inferior. This was certainly the case with my Radeon 7000 (second head display was *awful*) and with my GF4-MX (second head is fuzzier, bearable). With the high-end Quadro (at work) both displays are fine.
- My guess for high-quality on both heads is to buy one of these cards with both a DVI and a VGA output, not the kind with two VGA output. My reasoning is that since the manufacturer cannot assume which adaptor will be used for single-head display, the quality must be good on both. I made the error of buying a GF4-MX with two VGAs, one clearly labelled `second display'.
- suprisingly or unsurprisingly, I find the dual-head experience much simpler to deal with in general under win2k (autodetection of everything,neat little GUI tools for everything, etc), *HOWEVER*, once everything was working to my content under Linux, I find that things are much more flexible under Linux. There are fewer limitations in practice. For example under Linux with Nvidia's drivers I can decide which monitor is the primary one and how the desktop is organized independently. Under win2k there is only one combination that works.
DjVu is very good, but the document generation applications are not free; this is only fair as lizardtech is only a small company trying to survive. To their credit they have conversion tools available as web application (you upload your documents, they do the conversion automatically).
It is at least 2 orders of magnitude harder than reading the *commented* source.
Furthermore it may be `easy' technically, but the person doing the reverse-engineering may not be the one writing the competing piece of software or designing the piece of competing hardware. The reverse-engineer can only write a spec, from which a clean-room rewrite/redesign can be done by somebody else.
This is how Compaq reverse engineered the original IBM-PC. It took ages and the size of the ROM was minuscule compared to today's firmwares.
Here's a kind of vigilante guy in his helo taking pictures of properties down the coast looking for bad guys buidling walls along the beach (is this a real problem?, yeah, real criminals these guys are! scum of the earth!).
How many tons of CO2 is he releasing in the atmosphere doing that? But wait, it's alright! he's using a solar-powered PC and he is running Linux.
That makes it all OK. Guys, I'm leaving slashdot, goodbye.
At 50% of the speed of light, the time dilation factor is 1.145, i.e: 15%. The astronaut will look only a little younger than they could be when they come back.
On the other hand, unless they have really good radiation shielding, they'll probably be dead (see recent/. story on travelling to mars).
Nothing special happens to the travellers when they cross the horizon. The horizon is only a feature of time-space from the point of view of the distant observer.
Maybe IIS is no longer on by default with SP3, but up to SP2 it was.Given that SP3 is something that you apply after the installation unless you've put together an installation disk that does it for you, I think the `default on most' statement is correct.
On the sweatshops:
Not all western companies have a total absense of ethics when it comes to paying their workforce in the 3rd world. If you refuse to buy the products of a sweatshop-operating company but buy products from the more ethical ones, it's a win-win situation. Do your research about who is more ethical about the products you want to buy.
Sorry, your definition of a bug or design flaw is "whatever I don't like"?.
Compared to not being able to change the display resolution at all (eg: if you display at 1280x1024 that's it, if you want to change the resolution you need to re-configure and restart the server), which is the case on __absolutely *all* the commercial Unix distribution I know: Solaris, Tru64, Irix, etc__, the XFree86 zoom is 100% of a feature.
A feature is something that was purposely built into a piece of software. A bug is something that creeps in unintended, usually with bad effects, see the difference?
I also fail to see the design flaw here. A design flaw is something that prevents the user to accomplish a pre-determined task. Nothing of the sort here: your desktop is 100% usable without zoom or resize. Both are nice extras that can be handy sometimes, certainly not showstoppers if they aren't available.
At the worst the lack of desktop resize is a missing feature. No contest here. Desktop resize might be nice/useful but I can understand why it's not a high priority. Nothing is broken with the current situation. More hardware support would be good.
If that were the case then the frame rate would be limited to the refresh rate. Re-read what you just wrote.
What you describe is correct only if you tick the little box: `sync display with refresh' in the driver's preferences, which does limit your framerate.
Slashdot Mirror
That would work but just trying to speak a bit of French is enough, usually.
And that's a big one.
Thank you!
> Name me one useful invention / innovation done
> by OSS in the past fifteen years (since 1987)
You are joking, right?
I'll only list very few:
1- Heaps of new, useful languages. To name a few:
- Perl (born in 12/87)
- Python
- Java (is open-source, if not Free).
2- Operating systems
- Linux (certainly an amazing innovation, if not an outright invention)
- Mach (concepts date back from 1985 though)
- all the free BSDs. Lots of innovation there
3- Compiler technology
- Gcc (1988). Gcc is the only truly portable C/C++ compiler (and now Java, Ada, etc) usable for production work. Nothing else touches it.
4- Parallel computation technology
- MPI
- OpenMP
These a minuscule portion of free software inventions and innovations. If you know anything about IT you will acknowledge the importance of the above software packages.
The parent post was just trying to say that *Microsoft* has a tendency not to invent. It has been cheaper for them to let other companies invent new concepts and technologies and buy them outright rather than finance the R&D. This has been true from the very beginning of MS. They did not invent DOS, Gates bought it from someone else.
For more recent examples think of internet explorer, hotmail. Many people contend that MS implemented their TCP stack from the BSD sources (as they were allowed to do).
There is nothing inherently wrong with that, except you have to worry being in a world where no one can challenged the huge behemoth MS has become. If you are a small innovative company with a successful and popular closed-source product, three things can happen to you:
1 - Microsoft ignores you for a while until you become a menace,
2 - Microsoft buys you outright,
3 - Microsoft releases a competing product and more or less destroys you.
For an example of (1) think Adobe. Apparently now Microsoft is interested in taking over PDF, I wonder what will happen. (2) Think hotmail. (3) Netscape, Borland.
Lead shielding is only useful against gamma/X-rays. Against Alpha is it completely overkill (a thin sheet of paper is enough) and with beta (at least high-energy beta), the slowing down of the Beta particles in lead actually *causes* X-ray radiation.
This is called `brehmstrahlung' radiation. In fact this is how X-ray machines work.
I have a 1989 286 from Goupil, defunct European PC manufacturer. I used to consult for them. The motherboard is one of their development one, with all the bugs taken out with soldered wires and hand cuts on the PCB, there are dozens of them! It looks like an engineering nightmare, the ultimate Kludge. It runs fine though, DOS obviously.
I would love to put Linux on that thing.
That's good for you, but what if your job is to do 3D CAD, or animation, or simply scientific visualization that requires accelerated OpenGL?
Unfortunately under Linux you will have a *very* hard time with the open source accelerated 3D drivers. The best hardware you can run that way right now it the ATI Radeon 7500, and that's not great and that's aging rapidly, and there is little prospect of improvement or support for better hardware. I know because I've been there.
My life has become a whole lot simpler thanks to Nvidia, I hate to admit (even though their drivers aren't perfect, far from it).
BTW your argument about `convicing' somebody is the same whether it's open source software or not. In theory it's great that you can patch the source but in practice very few people have the time, skill and resources to do that for *every* application and framework that they use.
You have to pick your battles.
It's true that you can generate a seemingly unlimited amount of fuel using fast breeder reactors (that convert U238 to Pu239), but those are more dangerous to operate than conventional PWR reactors;
and fast breeder reactors can easily be converted to generate weapons grade fuel. Can you say nuclear proliferation?
There is of course the vexing problem of radioactive wastes. All told nuclear energy is rather pricey.
No problem with that. Unlike fission, fusion is not self-maintaining. There is no foreseable way that a fusion reaction can be maintained continuously here on Earth, let alone get out of hand.
The way Tokamaks and laser fusion reactors work is by doing a series of miniature implosions that consume their fuel, harvest the heat, clean up and start again.
Patents don't forbid people to reproduce what they describe. They only forbid someone from exploiting them commercially. In other words you can freely form patented thoughts in your head but you cannot distribute them (via telepathy, say).
Actually the Earth's axis of rotation does wobble quite a bit, if not by huge amounts. It certainly does so in a measurable way.
The Moon helps stabilize the axis of rotation. If we didn't have the Moon, the Earth could have ended up like Uranus is now: with an axis of rotation almost parallel to the plane of orbit. Seasons would have a different meaning in this context...
I've been running 99% linux dual head (except for the occasional game) for about 3 years now. I have used several different setups:
Nvidia TNT2 (AGP) + Matrox Mill-II(PCI)
ATI Radeon 7000 (AGP) + Mill-II
Nvidia Quadro (AGP) dual-heal
Nvidia GF4-MX dual head.
Here are a few comments:
- The combination of an AGP+PCI card is the most stable. Xinerama works fine, is well supported, and the display quality is great on both heads. At most you will get accelerated 3D on one of the heads though.
- Xinerama disables DRI, so you will not get accelerated OpenGL at the same time as having two working monitors. You need to revert to single-head for DRI to work.
- Nvidia is the only way (with their proprietary driver) if you want accelerated 3D on both heads.
With the current version of the driver everything works fine on reasonably current hardware. I found that the driver is more stable if I enable Nvidia's version of the AGP driver (not agpgart, but nvAGP, it's an option in the XF86Config-4 file).
- often the display quality on the second head on hardware that support it will be inferior. This was certainly the case with my Radeon 7000 (second head display was *awful*) and with my GF4-MX (second head is fuzzier, bearable). With the high-end Quadro (at work) both displays are fine.
- My guess for high-quality on both heads is to buy one of these cards with both a DVI and a VGA output, not the kind with two VGA output. My reasoning is that since the manufacturer cannot assume which adaptor will be used for single-head display, the quality must be good on both. I made the error of buying a GF4-MX with two VGAs, one clearly labelled `second display'.
- suprisingly or unsurprisingly, I find the dual-head experience much simpler to deal with in general under win2k (autodetection of everything,neat little GUI tools for everything, etc), *HOWEVER*, once everything was working to my content under Linux, I find that things are much more flexible under Linux. There are fewer limitations in practice. For example under Linux with Nvidia's drivers I can decide which monitor is the primary one and how the desktop is organized independently. Under win2k there is only one combination that works.
Cheers.
> Israel would be a tougher nut to crack than the EU.
To my knowledge Israel does not have nuclear submarines.
DjVu is very good, but the document generation applications are not free; this is only fair as lizardtech is only a small company trying to survive. To their credit they have conversion tools available as web application (you upload your documents, they do the conversion automatically).
It is at least 2 orders of magnitude harder than reading the *commented* source.
Furthermore it may be `easy' technically, but the person doing the reverse-engineering may not be the one writing the competing piece of software or designing the piece of competing hardware. The reverse-engineer can only write a spec, from which a clean-room rewrite/redesign can be done by somebody else.
This is how Compaq reverse engineered the original IBM-PC. It took ages and the size of the ROM was minuscule compared to today's firmwares.
Jeez man I can't believe this story.
Here's a kind of vigilante guy in his helo taking pictures of properties down the coast looking for bad guys buidling walls along the beach (is this a real problem?, yeah, real criminals these guys are! scum of the earth!).
How many tons of CO2 is he releasing in the atmosphere doing that? But wait, it's alright! he's using a solar-powered PC and he is running Linux.
That makes it all OK. Guys, I'm leaving slashdot, goodbye.
Good question, depends on the components.
CPU : 10-75W
Disk: 5-15W
Mobo: 5-25W
RAM : 1-5W/stick
GPU : 5-30W
PSU : 20-40W (lost efficiency)
Fans: 5W
Monitor: up to 200W.
All up an efficient PC will drain ~150W and an
inefficient or fully loaded one up to ~400W.
If you live in hot climes and use air con to
cool your house/office, you can easily double these numbers...
The time dilation formula is
/. story on travelling to mars).
delta = 1/sqrt(1-(v^2/c^2))
At 50% of the speed of light, the time dilation factor is 1.145, i.e: 15%. The astronaut will look only a little younger than they could be when they come back.
On the other hand, unless they have really good radiation shielding, they'll probably be dead (see recent
Sorry I was unclear,
Nothing special happens to the travellers when they cross the horizon. The horizon is only a feature of time-space from the point of view of the distant observer.
Reply to the minor point:
>And how is it enabled "by default on most"?
Maybe IIS is no longer on by default with SP3, but up to SP2 it was.Given that SP3 is something that you apply after the installation unless you've put together an installation disk that does it for you, I think the `default on most' statement is correct.
On the sweatshops:
Not all western companies have a total absense of ethics when it comes to paying their workforce in the 3rd world. If you refuse to buy the products of a sweatshop-operating company but buy products from the more ethical ones, it's a win-win situation. Do your research about who is more ethical about the products you want to buy.
OK, thanks for that.
I think you'll find that your templates are less flexible than the STL ones (List ?) but the speed and size might make up for the difference.
All the best
Sorry, your definition of a bug or design flaw is "whatever I don't like"?.
Compared to not being able to change the display resolution at all (eg: if you display at 1280x1024 that's it, if you want to change the resolution you need to re-configure and restart the server), which is the case on __absolutely *all* the commercial Unix distribution I know: Solaris, Tru64, Irix, etc__, the XFree86 zoom is 100% of a feature.
A feature is something that was purposely built into a piece of software. A bug is something that creeps in unintended, usually with bad effects, see the difference?
I also fail to see the design flaw here. A design flaw is something that prevents the user to accomplish a pre-determined task. Nothing of the sort here: your desktop is 100% usable without zoom or resize. Both are nice extras that can be handy sometimes, certainly not showstoppers if they aren't available.
At the worst the lack of desktop resize is a missing feature. No contest here. Desktop resize might be nice/useful but I can understand why it's not a high priority. Nothing is broken with the current situation. More hardware support would be good.
If that were the case then the frame rate would be limited to the refresh rate. Re-read what you just wrote.
What you describe is correct only if you tick the little box: `sync display with refresh' in the driver's preferences, which does limit your framerate.
Yeah, many people have described Microsoft as a great mice company with a software division.
These cards are not maintained by Donald anymore: /usr/src/linux/driver/net/8139too.c
> 8139too.c: A RealTek RTL-8139 Fast Ethernet driver > for Linux.
>
> Maintained by Jeff Garzik >
> Copyright 2000-2002 Jeff Garzik
I've got one of those cards, they work 100% fine.