This is a serious question:
Why do companies do this sort of thing with their products? It would seem to me that having the interface to a particular chip would not be particularly helpful to designing a competitor, ("Well, if I tell it to draw a blue square, it draws a blue square! I know how to copy that!") so what good does this do?
I've always been under the (possibly mistaken) impression that it made more sense to distribute specifications to everyone, so that others could use your hardware. If you have to write the drivers yourself for every operating system that you are going to allow to use your hardware then that would add quickly up to a rather large expense, wouldn't it?
Are drivers really that much of a proprietary, critical secret for hardware companies? Does having the source code for your drivers help anyone else create drivers for their products? What benefit is there in preventing others from having the drivers?
Sorry if these questions seem silly or unimportant, but I've never understood the other side of the secrecy of our drivers argument.
Based on programming for Motif and Qt (admittedly, this was two years ago), I liked the Motif widget sets and interfaces better.
Also, at least at that time, it seemed the Motif was a lot more polished, and had what I needed built in, in addition, it worked more seamlessly out of the box on my Solaris boxen, as compared to being quite as Linux oriented as Qt was (is?).
Is Qt a lot easier to use out of the box nowadays, on non-linux unix based platforms?
The problem with your example in the first paragraph is that you have the sign wrong.
20,000x$1,500=$30,000,000.00
20,150x$1,499=$30,204,850.00
Look at it this way: 150 customers who wouldn't have bought before, purchase machines at $1,499. This bring in $224,850 in revenue. The loss caused by lowering the price one dollar on the previous 20,000 people is... $20,000! Leaving a net *increase* in revenue of $204,850. This is the same principle that led to the industrial revolution... Selling one item at a high price returns a much lower profit than selling thousands (or millions) of items at asomewhat lower price (while still covering the costs).
I recently downloaded and installed OpenBSD on one of my machines. While I'll be the first to admit that I didn't work as hard as I could have at it, it still seemed to me to be completely paranoid. For instance, on my own machines, the first thing I do is create my user account, and then finish the setup by suing from there to root. (part of my setup on machines is to compile/install the necessary services, and I like having the source code owned by me, so that I can look at it without being a privelaged user). Straight off, I get the message that this user is not in the appropriate group to su to root. All in all, the machine seemed overly paranoid, and completely unworkable for a normal user. Reading the article, I discovered that I agree with most of their viewpoints (I think that limiting yourself to non-SMP because SMP hardware is more expensive is asinine - the power users are the ones most likely to need this kind of OS), but the hoops they make one jump to get a usable system are a pain. I mean, how difficult would it be for the installer to list the services, (with all of them off by default) and let you choose which ones to install? Or even to ask what normal user accounts should be in the admin group? Basically, I guess that I just want to say, I admire the idea behind their software, I just really don't like the way that they implemented it.
If you join the Oracle Technology Network (listed on their website as OTN), which is free, you can purchase a non-commercial use license to all of their products for a particular platform for $200 a year. I've done this myself. 4 updates (mailed to you on cd) during the course of the year, and access to online documentation for the products. All in all, since I work with Oracle for some projects as a consultant, I need to be able to hack on it at home to become familiar with it all. One thing to note: You can develop commercial applications with it, but you can't deploy them without purchasing a full license. So, if you want to try Oracle, it seems to make sense to join, since you can get the full product, work with it, and see if it works for your needs. If it does, you can purchase the full license, and be good to go.
Well, at the routers they can block all access to that port, they can set up a machine to check and see if any of the net connections they get are that protocol and kill the port if they are, and so forth. My school hasn't killed Napster yet, but the computer systems are run rather draconianly, ie, someone steals your ip address, and both your and their network connection are shut down until you talk to the Informatin services department. In any case, they can simply set up rules governing how much network traffic each machine is allowed to get over a non-reserved port. Ie, if port 37337 has 2 Gb transfer a day, you're probably doing something you shouldn't (according to the school's net use policy...)
Personally, I run solaris on my Ultra Sparc, and also on dual cpu x86 box, linux on my alpha, and also on my other x86 box, so I have to admit I really like both. I think that Solaris is much more polished, however, and I tend to think that it scales better than linux does - just from personal experience (On the same hardware, I can run more tasks simultaneously and have adequate response under Solaris than linux. Just my experience, with the standard configurations of both, ie, no tweaking/recompiling the kernel...)
My question is this: What are the chances that linux on Ultra Sparc will support cards which don't currently have support in Solaris, and what are the difficulties in porting those drivers to Solaris from there. In particular, I'm wondering about the 3com ethernet card (the 3c50x series). These cards are supported under Solaris for x86, but not on Ultra Sparc. I think that linux would be great at supporting popular hardware that might be *really* useful on Sparc hardware.
Another question: How different are the driver models for Solaris and linux? How much work would it take to make a driver for on OS to work under the other (Solaris/linux)? Also on a related note: whatever happened to the Unified Driver Model that Sun/HP/SGI were promoting about two/three years ago?
Looks interesting. No comment on what license, however, or whether or not any source is available. Anyone know what requirements they have about releasing the source?
Slashdot Mirror
This is a serious question:
Why do companies do this sort of thing with their products? It would seem to me that having the interface to a particular chip would not be particularly helpful to designing a competitor, ("Well, if I tell it to draw a blue square, it draws a blue square! I know how to copy that!") so what good does this do?
I've always been under the (possibly mistaken) impression that it made more sense to distribute specifications to everyone, so that others could use your hardware. If you have to write the drivers yourself for every operating system that you are going to allow to use your hardware then that would add quickly up to a rather large expense, wouldn't it?
Are drivers really that much of a proprietary, critical secret for hardware companies? Does having the source code for your drivers help anyone else create drivers for their products? What benefit is there in preventing others from having the drivers?
Sorry if these questions seem silly or unimportant, but I've never understood the other side of the secrecy of our drivers argument.
Based on programming for Motif and Qt (admittedly, this was two years ago), I liked the Motif widget sets and interfaces better.
Also, at least at that time, it seemed the Motif was a lot more polished, and had what I needed built in, in addition, it worked more seamlessly out of the box on my Solaris boxen, as compared to being quite as Linux oriented as Qt was (is?).
Is Qt a lot easier to use out of the box nowadays, on non-linux unix based platforms?
The problem with your example in the first paragraph is that you have the sign wrong.
... $20,000! Leaving a net *increase* in revenue of $204,850. This is the same principle that led to the industrial revolution... Selling one item at a high price returns a much lower profit than selling thousands (or millions) of items at asomewhat lower price (while still covering the costs).
20,000x$1,500=$30,000,000.00
20,150x$1,499=$30,204,850.00
Look at it this way: 150 customers who wouldn't have bought before, purchase machines at $1,499. This bring in $224,850 in revenue. The loss caused by lowering the price one dollar on the previous 20,000 people is
I recently downloaded and installed OpenBSD on one of my machines. While I'll be the first to admit that I didn't work as hard as I could have at it, it still seemed to me to be completely paranoid. For instance, on my own machines, the first thing I do is create my user account, and then finish the setup by suing from there to root. (part of my setup on machines is to compile/install the necessary services, and I like having the source code owned by me, so that I can look at it without being a privelaged user). Straight off, I get the message that this user is not in the appropriate group to su to root.
All in all, the machine seemed overly paranoid, and completely unworkable for a normal user.
Reading the article, I discovered that I agree with most of their viewpoints (I think that limiting yourself to non-SMP because SMP hardware is more expensive is asinine - the power users are the ones most likely to need this kind of OS), but the hoops they make one jump to get a usable system are a pain. I mean, how difficult would it be for the installer to list the services, (with all of them off by default) and let you choose which ones to install? Or even to ask what normal user accounts should be in the admin group?
Basically, I guess that I just want to say, I admire the idea behind their software, I just really don't like the way that they implemented it.
If you join the Oracle Technology Network (listed on their website as OTN), which is free, you can purchase a non-commercial use license to all of their products for a particular platform for $200 a year.
I've done this myself. 4 updates (mailed to you on cd) during the course of the year, and access to online documentation for the products. All in all, since I work with Oracle for some projects as a consultant, I need to be able to hack on it at home to become familiar with it all.
One thing to note: You can develop commercial applications with it, but you can't deploy them without purchasing a full license. So, if you want to try Oracle, it seems to make sense to join, since you can get the full product, work with it, and see if it works for your needs. If it does, you can purchase the full license, and be good to go.
Well, at the routers they can block all access to that port, they can set up a machine to check and see if any of the net connections they get are that protocol and kill the port if they are, and so forth.
My school hasn't killed Napster yet, but the computer systems are run rather draconianly, ie, someone steals your ip address, and both your and their network connection are shut down until you talk to the Informatin services department.
In any case, they can simply set up rules governing how much network traffic each machine is allowed to get over a non-reserved port. Ie, if port 37337 has 2 Gb transfer a day, you're probably doing something you shouldn't (according to the school's net use policy...)
Personally, I run solaris on my Ultra Sparc, and also on dual cpu x86 box, linux on my alpha, and also on my other x86 box, so I have to admit I really like both. I think that Solaris is much more polished, however, and I tend to think that it scales better than linux does - just from personal experience (On the same hardware, I can run more tasks simultaneously and have adequate response under Solaris than linux. Just my experience, with the standard configurations of both, ie, no tweaking/recompiling the kernel...)
My question is this:
What are the chances that linux on Ultra Sparc will support cards which don't currently have support in Solaris, and what are the difficulties in porting those drivers to Solaris from there. In particular, I'm wondering about the 3com ethernet card (the 3c50x series). These cards are supported under Solaris for x86, but not on Ultra Sparc. I think that linux would be great at supporting popular hardware that might be *really* useful on Sparc hardware.
Another question: How different are the driver models for Solaris and linux? How much work would it take to make a driver for on OS to work under the other (Solaris/linux)? Also on a related note: whatever happened to the Unified Driver Model that Sun/HP/SGI were promoting about two/three years ago?
Looks interesting. No comment on what license, however, or whether or not any source is available. Anyone know what requirements they have about releasing the source?