Red Hat co-founder Marc Ewing files to become a millionaire. Finally.
June 28, 2000 7:40 PM Subscribe
Red Hat co-founder Marc Ewing files to become a millionaire. Finally. More power to him, I say.
I feel that a guy who can sell free software for US$2500, or bundle an app with the OS ready to go for US$8995 deserves whatever he can get. I mean, the guy knows how to market.
But letting one distro call the shots on packaging and file structure is going down a road we've seen before. There is an effort to remind the commercial software developers that linux is a standard, not an RPM schema to integrate into. Redhat is the station wagon of distros... it does a little of everything, and it does it pretty well. But trying to get RedHat to excel at any one thing is a nightmarish frustration.
To their credit, they have been conscious of what they're doing, and have been conscientious of causing a splinter. They are respecting the GPL, and they release good software that's easy to get up and running for a new user.
I went to a job interview a few months ago, and the network admin asked me, "Do you know Linux version 6.0?", to which I replied, "Which 6.0? Linux is at 2.2.4, but a couple of the distros are at 6.0 or newer."
That's problematic. Linux does not move at the speed of the numbers on the boxes. That's just packaging around a core set of tools that are incredibly powerful, reliable, and customizable if you take the time to build them right.
posted by katchomko at 8:31 PM on June 28, 2000
But letting one distro call the shots on packaging and file structure is going down a road we've seen before. There is an effort to remind the commercial software developers that linux is a standard, not an RPM schema to integrate into. Redhat is the station wagon of distros... it does a little of everything, and it does it pretty well. But trying to get RedHat to excel at any one thing is a nightmarish frustration.
To their credit, they have been conscious of what they're doing, and have been conscientious of causing a splinter. They are respecting the GPL, and they release good software that's easy to get up and running for a new user.
I went to a job interview a few months ago, and the network admin asked me, "Do you know Linux version 6.0?", to which I replied, "Which 6.0? Linux is at 2.2.4, but a couple of the distros are at 6.0 or newer."
That's problematic. Linux does not move at the speed of the numbers on the boxes. That's just packaging around a core set of tools that are incredibly powerful, reliable, and customizable if you take the time to build them right.
posted by katchomko at 8:31 PM on June 28, 2000
Well, if anyone's interested in my severely negative opinion about Red Hat and all the other Linux companies, they can read it here.
I think Linux is a wonderful thing. I think Linux is going to survive and spread and prosper. I think every single one of those companies trying to base their business on Linux is going to go down in flames. (Corel's gonna die first, now that their plan to loot Inprise has fallen through, but they'll have plenty of company.)
posted by Steven Den Beste at 12:30 AM on June 29, 2000
I think Linux is a wonderful thing. I think Linux is going to survive and spread and prosper. I think every single one of those companies trying to base their business on Linux is going to go down in flames. (Corel's gonna die first, now that their plan to loot Inprise has fallen through, but they'll have plenty of company.)
posted by Steven Den Beste at 12:30 AM on June 29, 2000
Steven-
in response to your article, 2 quick points:
your footnote seems to imply that the only way to protect a commercial product is to only release binaries. Witness Samba/Microsoft, Blackdown/Sun, Gnutella/Napster, Gimp/Adobe, etc.
Linux is a kernel that acts remarkably like a traditional unix. Unices that were closed and proprietary, or at least proprietary and prohibitively expensive. It's built to be stable first, then fast and compatible, and usability doesn't really factor in anywhere. the tcp/ip stack, for example, looks a whole lot different from the original bsd stack. But they all work together, and they both have their strengths and weaknesses.
And then a more detailed response:
The value that RedHat is trying to make profitable is exactly that which I dislike most about their product. Making the OS usable by J. Random User. Sure, they package a lot of server software, and market it as such. But being able to click through an install and knowing all the key sequences in linuxconf --text does not make an admin, it makes an operator work in 80x25 vga mode.
The real money, for the long term, is in training. It's beginning now with the LPIC, but that's too operator centric. As I understand, RedHat's certification training spends less than half a day on email. It takes longer than that just to read a .cf ;)
The tools are cheap and given enough layers of scripting on top them, accessible to anyone. The price of all that abstraction is customizability, which is what they're trying to sell their product on. "In this box is an enterprise scale DNS, FTP, HTTP, SMTP and NetBIOS domain controller". yeah, that's all there... but trying to actually scale RedHat into anything requiring knowledge of how those daemons and protocols work is a nightmare. They sacrificed speed, power and flexibility for ease and convenience.
Which is fine, because now everyone who's vaguely related to IT knows what it is. It's expanding at an excelerated pace; look how long it took to get SMP into the kernel, and look how long it took to make SMP work fast. If we ran the NetCraft tests again today, you'd see that linux has improved 60-70% in the last year, while 2k's improvement is around 10%. But getting those kinds of benchmarks is not easy in the best circumstances, and RedHat makes it almost impossible.
So shops move into (or start out with) linux, and then realize that making it work right is a chore, so they pay an admin to come in and do it right. RedHat's making lots of money... not for themselves, but that's ok with me =)
My hat is still off to anyone who can sell $300 gui libraries and a free OS for 3 grand, that's just amazing.
posted by katchomko at 4:49 AM on June 29, 2000
in response to your article, 2 quick points:
your footnote seems to imply that the only way to protect a commercial product is to only release binaries. Witness Samba/Microsoft, Blackdown/Sun, Gnutella/Napster, Gimp/Adobe, etc.
Linux is a kernel that acts remarkably like a traditional unix. Unices that were closed and proprietary, or at least proprietary and prohibitively expensive. It's built to be stable first, then fast and compatible, and usability doesn't really factor in anywhere. the tcp/ip stack, for example, looks a whole lot different from the original bsd stack. But they all work together, and they both have their strengths and weaknesses.
And then a more detailed response:
The value that RedHat is trying to make profitable is exactly that which I dislike most about their product. Making the OS usable by J. Random User. Sure, they package a lot of server software, and market it as such. But being able to click through an install and knowing all the key sequences in linuxconf --text does not make an admin, it makes an operator work in 80x25 vga mode.
The real money, for the long term, is in training. It's beginning now with the LPIC, but that's too operator centric. As I understand, RedHat's certification training spends less than half a day on email. It takes longer than that just to read a .cf ;)
The tools are cheap and given enough layers of scripting on top them, accessible to anyone. The price of all that abstraction is customizability, which is what they're trying to sell their product on. "In this box is an enterprise scale DNS, FTP, HTTP, SMTP and NetBIOS domain controller". yeah, that's all there... but trying to actually scale RedHat into anything requiring knowledge of how those daemons and protocols work is a nightmare. They sacrificed speed, power and flexibility for ease and convenience.
Which is fine, because now everyone who's vaguely related to IT knows what it is. It's expanding at an excelerated pace; look how long it took to get SMP into the kernel, and look how long it took to make SMP work fast. If we ran the NetCraft tests again today, you'd see that linux has improved 60-70% in the last year, while 2k's improvement is around 10%. But getting those kinds of benchmarks is not easy in the best circumstances, and RedHat makes it almost impossible.
So shops move into (or start out with) linux, and then realize that making it work right is a chore, so they pay an admin to come in and do it right. RedHat's making lots of money... not for themselves, but that's ok with me =)
My hat is still off to anyone who can sell $300 gui libraries and a free OS for 3 grand, that's just amazing.
posted by katchomko at 4:49 AM on June 29, 2000
> "In this box is an enterprise scale DNS, FTP, HTTP, SMTP and NetBIOS domain controller". yeah, that's all there... but trying to actually scale RedHat into anything requiring knowledge of how those daemons and protocols work is a nightmare.
Um, yeah?
There is a *reason* we of the priesthood (:-) took so long to get to know how to do what we do. (And I speak as a system and network administrator and designer for about 25 small-mid size clients with about 12 years experience.)
Microsoft and their idiotic "Wizards" notwithstanding, *it* *is* *not* *easy* to build things that big without lots of knowledge and *experience*. That's why SAGE is the System Administrator's *Guild*. That's why I hate VB and Access.
"Now, *anyone* can build the systems to run your business".
Shyeah, right. Into the ground, maybe.
Folks, I got a hot flash for ya: This shit ain't easy. If it was that damned easy, corporate IS people wouldn't make $70-100K a year.
Course, we wouldn't work so many hours, either, but... :-)
posted by baylink at 12:10 PM on June 29, 2000
Um, yeah?
There is a *reason* we of the priesthood (:-) took so long to get to know how to do what we do. (And I speak as a system and network administrator and designer for about 25 small-mid size clients with about 12 years experience.)
Microsoft and their idiotic "Wizards" notwithstanding, *it* *is* *not* *easy* to build things that big without lots of knowledge and *experience*. That's why SAGE is the System Administrator's *Guild*. That's why I hate VB and Access.
"Now, *anyone* can build the systems to run your business".
Shyeah, right. Into the ground, maybe.
Folks, I got a hot flash for ya: This shit ain't easy. If it was that damned easy, corporate IS people wouldn't make $70-100K a year.
Course, we wouldn't work so many hours, either, but... :-)
posted by baylink at 12:10 PM on June 29, 2000
« Older The New-Boy Network | Craig's List Newer »
This thread has been archived and is closed to new comments
posted by ghostrocket at 8:14 PM on June 28, 2000