It was very refreshing to see your letter on Slashdot today. I think you represented yourself and your company very well.
However, I do have a rebuttal for you.
While I applaud you in your strong stance against those who feel that Red Hat is the Microsoft of the Linux world, I do see some (possibly unintentional) similarities. The most recent one being the decision to include a development version of GCC in the Red Hat 7 product.
Many critics of Microsoft are familiar with it's embrace and extend (unofficial?) philosophy. It has been seen in their reworking of the Kerberos security protocol, and it's initial attempt to take over the direction of the Java Virtual Machine and Development environment. It is a simple process: take something that is open and readily available, use the desktop dominance of the Windows platform to force their own changes upon the customer, and allow the original product, protocol, or idea to die, as the new variant takes hold.
In itself, this process can be beneficial. It may add new features to a product that were sorely lacking in previous versions. Or, it may serve as a way to create incompatibilities, and with market share dominance, can lead to a split with current competitors, leaving the others on an uneven playing field. And if it's a reworking of an open source project, which is released as a proprietary, closed source product, it's definitely not beneficial to the continuation of open and compatible software.
In the case of the Red Hat 7/gcc 2.96 issue, I can't quite understand why Red Hat would do this. First off, the GCC Steering Committee has issued a statement, saying that this development version of GCC will not function correctly with previous versions, is not supported, and will not be supported in the future. In fact, they state that the next version of the development compiler will not support GCC 2.96. This leaves anyone with a Red Hat 7 system in a bit of a bind. Do they go with the latest version, and lock themselves into a compiler that is not, and will not, be supported? Do they switch to another distribution? If they've invested a lot of resources into rolling out Red Hat on their systems, they're probably not going to want to switch to SuSE, or Slackware, or TurboLinux. In short, they're stuck.
I believe Alan Cox defended the decision by saying that Red Hat had every right to innovate and to make decisions which further the fight to make a better Linux distribution. While I don't disagree with him, I think it can be done in a better way, something other than creating incompatibilities. And recent claims that Red Hat started the open source movement are a little disturbing. I can think of two particularly prominent people who might take issue with this statement. Aren't we all in the same camp here?
Now, in theory, Red Hat Software is not very Microsoft-like. Red Hat is committed to the open source movement. Microsoft is not. Red Hat is very forthcoming about bugs, security issues, and instability problems. Microsoft is not. Red Hat allows Linux users to operate in a very user friendly environment, while still maintaining the power of the Linux kernel, as it is developed and maintained by the excellent work of people like Alan Cox and Linus Torvalds. Microsoft will not.
Yet, I feel a little insecure about Red Hat's decision to release a product with as critical a component as a compiler that is not supported by it's creators, and does not permit interoperability with other distributions of the GNU/Linux operating system.
I must admit, I am not a Red Hat user currently. I started my Linux experience with Slackware, before there were any other distributions. I did run Red Hat on both x86 and SPARC platforms for a short period of time. In one of my work situations, I was exposed to the SuSE distribution, and since then, have only installed and managed Slackware and SuSE servers. However, I do give Red Hat an immense amount of credit for the visibility, and indeed, excellent additions to the free software/open source community. However, by including GCC 2.96, you have shaken my faith in your commitment to aiding the advancement of GNU/Linux as an open and GNU compatible operating system.
As I know your mail will be swamped, I have opted to post this letter on Slashdot, in the hopes that it might spark some further discussion, and that you may read it at your leisure.
Thank you for your time, and all the best to Red Hat for the future.
There's a Linux for Cisco. There's a Linux for Palm. There's Linux for S/390. There's always going to be a new Linux port for something. SuSE has Linux for SPARC. The current trend is to make Linux run on everything imaginable (if you had a toaster that had been hacked up, you could run Linux on that), and I doubt, with the amount of people owning SPARC-based machines that Linux on SPARC will die with Red Hat not supporting it.
Red Hat is a commercial entity. They make money off of Red Hat on x86. That's their plan. If SPARC ports don't make money, they cut it. However, I doubt that suddenly, because of this, the SPARC platform will be forgotten.
Time to move on, and try something different. My wife runs NetBSD on our SPARCstation 10. When we get a copy of SuSE for SPARC, we'll probably give it a go as well.
Agreed. NeXT is more than alive with Apple now. Everything they've done with MacOS and the design of the cube shows that Jobs still believes in what he did with NeXT. I think SGI should sue Cobalt over the colour of their RaQ, as the Indigos were an identical colour (if not as nite-glo[tm]).:)
Slashdot Mirror
Dear Mr Young.
It was very refreshing to see your letter on Slashdot today. I think you represented yourself and your company very well.
However, I do have a rebuttal for you.
While I applaud you in your strong stance against those who feel that Red Hat is the Microsoft of the Linux world, I do see some (possibly unintentional) similarities. The most recent one being the decision to include a development version of GCC in the Red Hat 7 product.
Many critics of Microsoft are familiar with it's embrace and extend (unofficial?) philosophy. It has been seen in their reworking of the Kerberos security protocol, and it's initial attempt to take over the direction of the Java Virtual Machine and Development environment. It is a simple process: take something that is open and readily available, use the desktop dominance of the Windows platform to force their own changes upon the customer, and allow the original product, protocol, or idea to die, as the new variant takes hold.
In itself, this process can be beneficial. It may add new features to a product that were sorely lacking in previous versions. Or, it may serve as a way to create incompatibilities, and with market share dominance, can lead to a split with current competitors, leaving the others on an uneven playing field. And if it's a reworking of an open source project, which is released as a proprietary, closed source product, it's definitely not beneficial to the continuation of open and compatible software.
In the case of the Red Hat 7/gcc 2.96 issue, I can't quite understand why Red Hat would do this. First off, the GCC Steering Committee has issued a statement, saying that this development version of GCC will not function correctly with previous versions, is not supported, and will not be supported in the future. In fact, they state that the next version of the development compiler will not support GCC 2.96. This leaves anyone with a Red Hat 7 system in a bit of a bind. Do they go with the latest version, and lock themselves into a compiler that is not, and will not, be supported? Do they switch to another distribution? If they've invested a lot of resources into rolling out Red Hat on their systems, they're probably not going to want to switch to SuSE, or Slackware, or TurboLinux. In short, they're stuck.
I believe Alan Cox defended the decision by saying that Red Hat had every right to innovate and to make decisions which further the fight to make a better Linux distribution. While I don't disagree with him, I think it can be done in a better way, something other than creating incompatibilities. And recent claims that Red Hat started the open source movement are a little disturbing. I can think of two particularly prominent people who might take issue with this statement. Aren't we all in the same camp here?
Now, in theory, Red Hat Software is not very Microsoft-like. Red Hat is committed to the open source movement. Microsoft is not. Red Hat is very forthcoming about bugs, security issues, and instability problems. Microsoft is not. Red Hat allows Linux users to operate in a very user friendly environment, while still maintaining the power of the Linux kernel, as it is developed and maintained by the excellent work of people like Alan Cox and Linus Torvalds. Microsoft will not.
Yet, I feel a little insecure about Red Hat's decision to release a product with as critical a component as a compiler that is not supported by it's creators, and does not permit interoperability with other distributions of the GNU/Linux operating system.
I must admit, I am not a Red Hat user currently. I started my Linux experience with Slackware, before there were any other distributions. I did run Red Hat on both x86 and SPARC platforms for a short period of time. In one of my work situations, I was exposed to the SuSE distribution, and since then, have only installed and managed Slackware and SuSE servers. However, I do give Red Hat an immense amount of credit for the visibility, and indeed, excellent additions to the free software/open source community. However, by including GCC 2.96, you have shaken my faith in your commitment to aiding the advancement of GNU/Linux as an open and GNU compatible operating system.
As I know your mail will be swamped, I have opted to post this letter on Slashdot, in the hopes that it might spark some further discussion, and that you may read it at your leisure.
Thank you for your time, and all the best to Red Hat for the future.
Look at it this way:
There's a Linux for Cisco. There's a Linux for Palm. There's Linux for S/390. There's always going to be a new Linux port for something. SuSE has Linux for SPARC. The current trend is to make Linux run on everything imaginable (if you had a toaster that had been hacked up, you could run Linux on that), and I doubt, with the amount of people owning SPARC-based machines that Linux on SPARC will die with Red Hat not supporting it.
Red Hat is a commercial entity. They make money off of Red Hat on x86. That's their plan. If SPARC ports don't make money, they cut it. However, I doubt that suddenly, because of this, the SPARC platform will be forgotten.
Time to move on, and try something different. My wife runs NetBSD on our SPARCstation 10. When we get a copy of SuSE for SPARC, we'll probably give it a go as well.
No harm done. Just a different path to take.
Long Live The Alternatives!
Agreed. NeXT is more than alive with Apple now. Everything they've done with MacOS and the design of the cube shows that Jobs still believes in what he did with NeXT. I think SGI should sue Cobalt over the colour of their RaQ, as the Indigos were an identical colour (if not as nite-glo[tm]). :)