i don't know the size of the RedHat team, how much they get paid, and how long it takes to develop the updates themselves. Figure a routine update takes 3 weeks to develop, at 70,000 a year for each developer, and 15 developers. This comes to (70,000 / 52 weeks) * 3weeks * 15people =~ $60576 per update, in salary. RedHat of course has many other expenses to consider as well, so figure an update is probably a $100,000 investment. That's probably on the low-end, really.
What I meant was, you need a product you can sell and bundle linux along with it. Like IBM sells server hardware, and ships Linux with it. They make their money with the server, and save the OS license fee by using linux.
You need a real product you can sell that makes up for the cost for development of linux. Whether it be some other software, or hardware. But just support and distribution medium is NOT ENOUGH.
I use linux for running servers at work (RedHat serves me fine), and win2000 for everything else. The linux GUI's in their current state remind me of all the elegance of windows 3.1, and win2000 is a very nice OS for general computing tasks. It doesn't lock up on me, and works well for my needs.
Linux is a better server OS. Less overhead for more long-term stability.
I don't care about distros. I'm not close minded or open minded about it really. I'd use Joe Bob Blow's Linux if they decided to load it on my work servers.
That said, I take a neutral, realistic viewpoint on Linux. That makes it unpopular around here, because if you're not a zealot, you're the enemy.
they are a business. Businesses must make money. The sad reality of the linux software industry is this:
Software costs thousands of dollars to produce, because you must pay the developers. While this is not true of the volunteer software, of course, it is true for companies such as RedHat.
With linux, they cannot charge for the software itself, because of the GPL. Thus, they come up with flaky plans such as selling "support" (unless your software is really really poor, you're not going to make a ton of money from support. Plus you also have to pay your support personnel).
Selling distribution mediums, such as CDs and Books. Well with the great internet, you can download the software, so theres no need for a CD, and the books are available online too. So why spend money buying the distribution mediums? Frankly, not many people do.
So, you have a very high input price - the developers, going into a product you can't legally sell, and try to make up the difference through product complements. Basic economics will tell you that this doesnt work.
So, back a couple years ago when the economy was really hot, and venture capitalists would throw money at anything marked "computer" despite the lack of a decent business plan, linux companies flourished. Today, the economy is suffering, investors are looking hard at business plans and seeing the fact that linux company's Rate of Return is lower than that of "traditional" software companies, and investors are pulling out.
Mark my words, the linux software corps days are numbered. The only way you can survive with linux, is if you have more products and you develop linux ALONG with them. IBM is the perfect example of this.
I don't know how Debian works or even if there is a company based around it. Nor do I really care, since I don't use it, never have, and never thought about it. So it simply does *not* fit in my "world view".
This thread is about why RedHat is charging for updates, not about some other linux distro.
Software developers are extremely expensive people. I should know, I'm one of them. You don't think the people that work at RedHat and write software work for free do you?
It costs thousands, nay, millions to develop big pieces of software. This kind of cost cannot be made up by selling "support" (unless you're software is so shitty it requires users to sit on the phone with you for hours everyday) nor distribution costs.
applications are growing in numbers rapidly. Not just "server-side" stuff, but the server software itself. Java has only been around a few years, and its only been lately that people have finally gotten a good grasp in writing large projects using it. There's also the fact that developers are among the last people to start using new different tools.
I'm a server developer. I use java. It's ease of creating an advanced multi-threaded application that runs on multiple platforms is amazing, and believe me, its far easier than developing with C++, which makes all the difference in reducing the amount of bugs to begin with, but in the bug hunt after one is discovered.
I do my development in win2000, and deployment usually on linux. It's wonderful. I don't have to change a single thing, just copy my JAR file over to the linux box and it works the same as it did in windows.
Any program not written properly in *any* programming language is slow and unstable.
Java is a *very* fast language when using today's modern tools and written properly (Resin webserver, written in java, is nearly as fast as apache for static pages, and its JSP/servlet processing matches the speed of mod_php, and simply blows away mod_perl).
Today, people have much more knowledge of how to code in java correctly, as opposed to several years ago when java first started appearing. And the tools today are quite a bit more advanced.
rather than using the napster protocol (which is very limited) i've developed a different one so all files can be accomodated, and with more advanced searching capabilities.
It uses the distributed server arrangement, rather than sharing indexes around (big bandwidth clog) it instead passes search requests around. It'll be up on sourceforge soon. Stay tuned.
In my newly founded religion, you are allowed multiple wives (up to as many as L. Rob Hubbard has shitty books) and our god is John Travolta. Please hang a picture of him in your house.
ITN (you'll be hearing about this soon) has stored file indexes on various 3rd party servers, however the servers are controlled by individuals rather than corporations. Whats the legal ramifications for this? Though no actual data passes through, and the sysadmin has no way to tell what is stored in his index unless he manually outputs the database.. there seems to be a gray area about whether this constitutes the "aiding and abetting" clause. Consider also the fact that these do not have financial goals either.
When you use a computer, are you more interested in writing the software for it, or how the hardware itself works? As in designing the cards and electrical architecture.
If you're good at math and physics, CE might be for you, because theres quite a bit more in that than CS (though CS still requires some math, perhaps physics depending on your university)
If theres anything gnutella and the stats of freenet have shown, its that pure peer-to-peer distribution DOES NOT WORK. The current design of the internet makes it exponentially slow above a few thousand users. Thus, it is not effective, too slow, too hard to use and too unreliable.
It's not difficult putting users together, or even starting a search. Its difficult knowing WHEN to stop a search to prevent enormous bandwidth consumption..something that the pure peer's dont consider.
Anyway, an ISP does not have to setup the server. Thats absurd. I can run a webserver serving thousands of hits a day off a DSL line. A well written protocol and a server allows individuals to setup their own server (its easy to do..current test software ive got installs and configures in 10 minutes)
Second, understand this is NOT irc. I use the irc as an example because its servers interconnected to each other.
Scalability. As the network grows, so will users running servers. Thus, it self-scales.
Flexibility. Networks can be independent of one another..one could be a music trading, another devoted to movies. You could even setup your own private network for whatever purpose you desire.
I lived there for years of my life before finally moving to Austin last summer. I'm so glad I did. Got tired of hugeness of the city (takes 1.5 hours going 80 mph to cross it), the heat and humidity and the fact that there really is not much of a nightlife there.
Taking a queue from the network layout of IRC, it works in a similar fashion. users connect to servers still, but they're not congregated together in one place, or by one entity. Servers can easily talk to one another across the network, passing search requests on, pairing up users. The protocol spec will be ready within a week or two, and before committing to much software, i'm going to open the protocol itself to public commentary (this is something that is simply too big for one person to claim to have the "perfect" plan). Some demo software will be available too.
because i have many friends who are less-than-technical who love the damned thing.
i don't know the size of the RedHat team, how much they get paid, and how long it takes to develop the updates themselves. Figure a routine update takes 3 weeks to develop, at 70,000 a year for each developer, and 15 developers. This comes to (70,000 / 52 weeks) * 3weeks * 15people =~ $60576 per update, in salary. RedHat of course has many other expenses to consider as well, so figure an update is probably a $100,000 investment. That's probably on the low-end, really.
You need a real product you can sell that makes up for the cost for development of linux. Whether it be some other software, or hardware. But just support and distribution medium is NOT ENOUGH.
Linux is a better server OS. Less overhead for more long-term stability.
I don't care about distros. I'm not close minded or open minded about it really. I'd use Joe Bob Blow's Linux if they decided to load it on my work servers.
That said, I take a neutral, realistic viewpoint on Linux. That makes it unpopular around here, because if you're not a zealot, you're the enemy.
Software costs thousands of dollars to produce, because you must pay the developers. While this is not true of the volunteer software, of course, it is true for companies such as RedHat.
With linux, they cannot charge for the software itself, because of the GPL. Thus, they come up with flaky plans such as selling "support" (unless your software is really really poor, you're not going to make a ton of money from support. Plus you also have to pay your support personnel).
Selling distribution mediums, such as CDs and Books. Well with the great internet, you can download the software, so theres no need for a CD, and the books are available online too. So why spend money buying the distribution mediums? Frankly, not many people do.
So, you have a very high input price - the developers, going into a product you can't legally sell, and try to make up the difference through product complements. Basic economics will tell you that this doesnt work.
So, back a couple years ago when the economy was really hot, and venture capitalists would throw money at anything marked "computer" despite the lack of a decent business plan, linux companies flourished. Today, the economy is suffering, investors are looking hard at business plans and seeing the fact that linux company's Rate of Return is lower than that of "traditional" software companies, and investors are pulling out.
Mark my words, the linux software corps days are numbered. The only way you can survive with linux, is if you have more products and you develop linux ALONG with them. IBM is the perfect example of this.
This thread is about why RedHat is charging for updates, not about some other linux distro.
It costs thousands, nay, millions to develop big pieces of software. This kind of cost cannot be made up by selling "support" (unless you're software is so shitty it requires users to sit on the phone with you for hours everyday) nor distribution costs.
you mean, after spending thousands and thousands of dollars for developing updates..they're going to charge?? How dare they follow economic reality!
Resin webserver. Makes tomcat seem lazy ;) Matches apache in static serving speed, too.
I'm a server developer. I use java. It's ease of creating an advanced multi-threaded application that runs on multiple platforms is amazing, and believe me, its far easier than developing with C++, which makes all the difference in reducing the amount of bugs to begin with, but in the bug hunt after one is discovered.
I do my development in win2000, and deployment usually on linux. It's wonderful. I don't have to change a single thing, just copy my JAR file over to the linux box and it works the same as it did in windows.
Java is a *very* fast language when using today's modern tools and written properly (Resin webserver, written in java, is nearly as fast as apache for static pages, and its JSP/servlet processing matches the speed of mod_php, and simply blows away mod_perl).
Today, people have much more knowledge of how to code in java correctly, as opposed to several years ago when java first started appearing. And the tools today are quite a bit more advanced.
It uses the distributed server arrangement, rather than sharing indexes around (big bandwidth clog) it instead passes search requests around. It'll be up on sourceforge soon. Stay tuned.
Thank you.
ITN (you'll be hearing about this soon) has stored file indexes on various 3rd party servers, however the servers are controlled by individuals rather than corporations. Whats the legal ramifications for this? Though no actual data passes through, and the sysadmin has no way to tell what is stored in his index unless he manually outputs the database.. there seems to be a gray area about whether this constitutes the "aiding and abetting" clause. Consider also the fact that these do not have financial goals either.
/. posts something that takes a common sense approach to linux as opposed to the zealot way. Good job, slashdot. You get a cookie for this.
maybe because lanl does more than nuclear weapons?
is like comparing a moped to an 18 wheeler. Sure, you can do delivery jobs with either. But how much do you need to deliver? And how fast?
It remains mine. I still own the copyright of it. It is absolutely essential that you get this in writing.
They paid handsomely, too. It was a win-win situation for all involved.
no one has said "GPS SUCKS ITS NOT OPEN SOURCE ENOUGH!" yet. *knocks on wood*
alot more people use windows than those who use unix, therefore it only seems natural to have more virii spread this way.
If you're good at math and physics, CE might be for you, because theres quite a bit more in that than CS (though CS still requires some math, perhaps physics depending on your university)
It's not difficult putting users together, or even starting a search. Its difficult knowing WHEN to stop a search to prevent enormous bandwidth consumption..something that the pure peer's dont consider.
Anyway, an ISP does not have to setup the server. Thats absurd. I can run a webserver serving thousands of hits a day off a DSL line. A well written protocol and a server allows individuals to setup their own server (its easy to do..current test software ive got installs and configures in 10 minutes)
Second, understand this is NOT irc. I use the irc as an example because its servers interconnected to each other.
Scalability. As the network grows, so will users running servers. Thus, it self-scales.
Flexibility. Networks can be independent of one another..one could be a music trading, another devoted to movies. You could even setup your own private network for whatever purpose you desire.
I lived there for years of my life before finally moving to Austin last summer. I'm so glad I did. Got tired of hugeness of the city (takes 1.5 hours going 80 mph to cross it), the heat and humidity and the fact that there really is not much of a nightlife there.
eh?
Taking a queue from the network layout of IRC, it works in a similar fashion. users connect to servers still, but they're not congregated together in one place, or by one entity. Servers can easily talk to one another across the network, passing search requests on, pairing up users. The protocol spec will be ready within a week or two, and before committing to much software, i'm going to open the protocol itself to public commentary (this is something that is simply too big for one person to claim to have the "perfect" plan). Some demo software will be available too.