What sort of books have you used as a touchstone for shaping your web site?
There's a lot of material out there, and a lot of experts, but who do you think really provided you with good advice on how to put together your website?
Most of the political web sites I've seen out there are what I refer to in the web development industry as "brochure sites". They tend not to be terribly interactive, and mostly serve as an online brochure.
What can you do with a political site that really enhances your ability to get the message out there?
Can someone explain to me why stuff like this shouldn't be clear evidence to the FCC or trust-busters that there's a problem with this merger? This is the kind of behavior that should trigger an immediate response.
On the "betatest" newsgroups, there are literally thousands of posts with testers posting what they consider "bugs" which, in my opinion, are nothing but "niggly" things that doesn't deserve any attention.
I think this is a really good example of where open source has some advantages. Little "niggly" things get fixed by the person they're "niggling".;-)
Well, the fact that escapist fiction talks about experiences that are illegal for the general populace is no surprise.
I think the role of "counter-culture" on the Internet is misleading. Certainly, the counter-culture has benefitted from the anarchy that is the Internet. Similar phenomena have happened in the past. I believe Dyson is quoted as say, "The Internet is great for conspiracies and bad for propoganda." I think she's right.
Certainly a lot of the Unix gods from the 70's were using drugs, like a lot of people their age at that time. Unix and the Internet were part of University culture for a long time, and the counter-culture is/was also strong there. Indeed, to this day geeks culture is still closely tied to University. It's no surprise that it's also tied to concepts that are strong in University.
However, I disagree with the notion that the computers are linked to the modern counterculture. There is definitely a lot of computer people who weren't part of mainstream culture, particularly when they were growing up (I still think being a geek may not be too trendy in school;-). However, computers are bigger than that community now. More people are on the Internet than develop the Internet... by an overwhelming margin.
But let's face it, the bulk of the people on the Internet use it to send people e-mails, buy things online, look up information on companies and consumer goods, and plan vacations. This is DEFINITELY not counter-culture stuff.
Not only that, but a lot of the CONTENT on the Internet these days is mainstream media, adverstising, marketing material from companies, etc. This is also the segment that's going to grow as more non-tech people come online.
The Internet was a brave new counter-culture world before AOL was in on it, before Al Gore (counter-culture visionary???;-) started talking about it, before guys in suits built entire economies on it, before banner ads, before SPAM, etc.
Today, the Internet *is* the world's culture. Every person who's a part of it participates in it, and that's now enough people that it can't help but be largely representative of mainstream culture.
I also don't agree that the Internet is an overwhelming sensory experience. TV can be, movies can be, but the Internet isn't. That's what the media wants to think the Internet is all about: flash and bang, but they're basically missing the point.
Ultimately, the Internet is a very intricate and structured piece of reality, and I don't see pushing yourself farther from reality as preparing you for that experience.
I'm not panning the drug experience, just the notion that it is in anyway connected to the Internet.
I'm talking about fundemental weaknesses being found in the encryption algorithms. Pre-WWII algorithms were breakable by any cryptanalyst out there with relative ease, regardless of key size, for about 100 years. The only uncrackable system was Enigma. Enigma was busted in roughly 5-10 years, at which point, it didn't matter what was done.
It is conceivable (and indeed, it's sort of expected) that at some point each of the algorithms out there will suddenly have simple solutions, at which case decryption will be trivial.
Similarly, with the advent of things like quantum computing, key length might become irrelavent regardlesss.
That's why you have the equivalent of a digital notary. Different systems work different ways, but essentially it always boils down to another party "vouching" for your signature (saying it's really you), and signing your signature.
One thing that is fairly easy to prove is that the same signature was used for multiple documents. So, for example, one could say, "well, if it wasn't you, then who signed for the mortgage on your house???!?!"
Presuming the telephone network is secure, particularly if you're not using digital signatures on it.;-)
Even if it's not, all the dial-out thing would do would be to confirm that someone at that phone number picked up the phone and said, "sure". The only way to really be sure it was him/her would be if you used a digital signature.
Let's not even talk about how clunky this would be to implement.;-)
Actually, if a digital signature was used the process of authenticating you with Outpost.com would be significantly easier. Think about this: that password you used to login.. how long was it? With computers doubling in speed every 18 months, the necessary length of a password is increasing at the rate of about one alphanumeric every 18 months (based on the fact that people tend to use english passwords and they have limited entropy).
Additionally, it's hard for Outpost.com to prove that you actually ordered a specific item. You can easily claim it was a mistake in their software and it looked like you were ordering something else. With a digital signature, there can be little doubt (barring a bug in your video driver;-) that you saw exactly what you were signing for.
First of all, for those of you who are concerned that this opens up some huge problem in security because bits can be copied easily, please read up on digital signatures and how they work. Rest assured that provided you use them properly, it is VERY hard for someone to add your digital signature to another document (unless the contents of that document are bit-for-bit the same as one you've already signed --in which case, who cares?).
The concern I have is that this is based on what we CURRENTLY know about encryption technology. I've been reading up on the history of cryptography, and it really looks like a horse race between code makers and code breakers. The thing is, the code makers HAVEN'T been consistently winning. Indeed, if you look back in history prior to the 1970's, you'll find that there were very limited periods of time when code makers were winning, and frequently it was only for short periods of time.
What's going to happen when the inevitable happens? Particularly if cryptographers don't have a new discovery to replace the broken approach. Once the infrastructure of using digital cryptography is in place, it's going to be hard to undo it (case in point: how companies/governments/individuals elected to avoid Y2K problems by simply pulling the plug?).
I don't think this is a reason not to use digital signatures. I think it's a reason to start thinking NOW about how to handle the seemingly inevitable moment when someone figures out how to crack existing approaches... particularly if there is no replacement.
I grow tired of this debate. You're missing what I'm saying, and ignoring obvious facts.
First of all, Redhat for the most part DOESN'T write GPL software. For the most part, they PACKAGE GPL software. They do write GPL software to help broaden the market appeal, but 99% of the code on a RedHat CD was not developed by them.
As far as CD sales go, I'm not saying this is Red Hat's future. I am just demonstrating that the name brand has value. Red Hat sells the box set for $75 in stores, next to other boxes that sell for less. They still sell a lot of CD's. Red Hat sells (online) a guru-friendly version of Red Hat for $30. It comes wth no docs or add-ons, and yet it sells for 15x more than a cheapbytes CD. Amazingly, it sells very well. That my friend, is the power of a brand name, and in the OSS world, RedHat has the best one out there. Granted, it's a small market still, but that doesn't mean that Red Hat doesn't benefit from having that name brand out there.
My point about Deloitte & Touche and Anderson Consulting was that you CAN charge a premium in the consulting business without having proprietary knowledge, something which you said wasn't true.
Indeed, they have established impressive reputations, but to a large extent their reputation for computer systems work (to the extent that it's relevant in today's world) has been from perfromance in the past 15 years.
Cygnus Solutions, as I mentioned before, is a VERY well established consulting company with a very well respected reputation. They are THE guys that compay's go to for solutions built with GNU tools. This is a company that's been around for 11 years, with 180 employees, with offices around the world, all doing purely consulting. They are very well established in the embedded systems market, as well as various IT projects. They have an extensive client list, etc., etc. As the maintainers of gcc and glibc they have an unparalleled reputation for working with those tools (which you need to do if you want to play the Linux game). They are now owned by Red Hat. So Red Hat does indeed have a reputation. IT managers and managment HAVE ALREADY written checks based on this repuration. Small little companies you may have heard of like Nortel, SGI, Motorola, and I believe Nintendo.
I have never disagreed that the whole tech sector is overvalued, but I think that RedHat is proportionally valued to it's competition. I do think though, that the computer sector has proven that they have a significant difference with Bio Techs: initial investment is smaller, and results are more immediately available. The lack of both of these REALLY hurts the biotechs.
Other minor points: There's plenty of freely available documentation out there for Linux, and RedHat isn't the only distribution that comes with documentation. (Don't forget, when RedHat was establishing themselves, one of the most common ways to get Linux was to buy a "Learning Linux..." book with a Linux CD in the jacket.) Nonetheless, people pay a premium for RedHat. That's brand name strength. RedHat can use it to peddle more than CD's.
Coke and Pepsi are VERY similar formulas, and if you like one or another it wouldn't take you very long to make something that matches it. I don't know what you're getting your comment about carbonation from. As a kid, I worked at a fast food place and one of the things I did was work the fountains. I'm telling you, the only difference between Coke and Pepsi was the sack of syrop they came from. They both ran through the same carbonation machine. It is TERRIBLY relevant if it's people's imaginations that seperate the two, because if it is... that's a BRAND NAME that's really impacting them.
Finally, can you explain this to me: why would the market for virtually free Red Hat software be smaller than the market for $75 Red Hat software? (your point c). This makes no sense to me, unless you're saying that people will only buy Linux if they feel there's a...don't say it... BRAND NAME behind it!
You can't be called a true SF reader until you've read the classics. These include:
Asimov's Foundation Series Asimov's Robots Series Heinlein's Stranger in a Strange Land Clarke's 2001 (the rest of the series is a wash to me) Herbert's Dune Series Adam's Hitchiker's Guide To the Galaxy Series
I'd throw in Heinlein's "Moon is a Harsh Mistress" but not everybody agrees that's a classic.
Then there's a couple of modern books that make sense:
Gibson's Neuromancer Stephenson's books, read at least one of Snow Crash, The Diamond Age, and Cryptomnicron
After that, you've established your SF credentials and are free to persue other works of the genre.;-)
Do you really think there's much that's unique about Coke, Pepsi, or Heinz? PLEEAASE! The major differences between the first two is the amount of sugar put in to what they purchase. The profits of these companies also aren't relevant. In RedHat's sector, some of the most successful companies are not profitable, and still worth billions. (Case in point look at IPIXD... and they've never made money and a good portion of their valuation comes from bamboo.com, which has really only been in the this business model for 2 years --compared to RedHat which has been doing this for a while).
RedHat now owns Cygnus, who've essentially been THE support guys for GNU tools for the last decade. They're well established (and you'll notice that their executive team is taking over RedHat), and they are schrewd businessmen, and the history of that company demonstrates that. So RedHat's support is far from non-existent. Keep in mind they have several support contracts with OEM's as well.
If you think branding doesn't work with software, ten ask me how RedHat has sold a SINGLE CD-ROM! You can get the same thing from cheapbytes for 1/30th the price! You can download it for free off the Internet! Still, RedHat sells CD-ROMs. They also sell T-shirts, caps, etc. They actually make money doing it too. They're brand name is already worth a lot. That's why Dell signed them up, why companies pay to get the "runs on RedHat" write up, etc.
Consulting company which charges a premium for it's servicse on a large scale w/o any proprietary technology: Andersen Consulting. Indeed, if you talk to them they make money BECAUSE they don't have any proprietary technology. Deloitte & Touche have a pretty good business in this area as well. It happens. You do a good job, you get a reputation, and you charge for it.
Think of it this way, you're a company, you want to do something with Linux, who are you going to call to make a deal? If you call RedHat you know your stock will go up just that much more. RedHat can make a little bit of money off of every company that wants to do business in the Linux buxiness space if they play their cards right, and the best part is the more they do it, the more their brand name is even more valuable.
As a final point: I bet you that just about anyone can make better ketchup than Heinz if they try. They just simply don't.
Well, Redhat's valuation is crazy, but I don't think it's crazy when you compare it to it's competitors. The market is crazy right now, and so valuations reflect that. Redhat has significant first-mover advantages, and has the highest name recognition in it's space. It's also aquired Cygnus, which is a well proven business, and will be the source of a lot of their revenue.
Redhat's game isn't just to provide service. What they really provide is "Branding". "Branding" has justified the valuations of companies like Coke, Pepsi, and Heinz, so I don't think Redhat should be any different. Redhat doesn't just brand their support. They use their brand to charge a premium for services, and in the long run that can be VERY profitable.
Man, if ever I saw a reason why steganography, this is it.
I'd like to suggest a solution to this problem. Let's call it the "Redundant Distributed Network Steganographic File System", mostly because it's an acronym that can't be made into a cute name.;-)
Now, the idea is, everyone gives up some disk space, say 0.5gb, which they make acessible on the Internet. In exchange, they get like 0.1gb of space on the RDNSFS. The filesystem LOOKS like noise, but if you have the right key, you can extract a certain amount of data from it. If you have the wrong key, you get fake data.
Now, the big problem is how to allocate space for someone without giving away that they have data out there somewhere....
Ok. first of all, I *am* an entrepreneur. I also work with entrepreneurs on a regular basis. I still very much disagree with your premises:
a) A startup entrepreneur can definitely apply for a patent. A college student experimenting in his apartment cannot. Both are at a disadvantage to a major corp., because it's essentially out there trying to plug all the possible holes in it's IP, whereas they're going to apply for the one or two patents that really embody their idea. In the long run, the corp will have more leverage.
b) I think most people on Slashdot are aware of this. The point is that it's a tragedy. The idea of patents is SUPPOSED to be to encourage the exchange of ideas, not patent suits. What you're saying is that the Patent Office has no real credibility, and that needs to change.
c) The suggestion that we scrap old business models and go with new ones ISN'T because of RedHat's success. Indeed, it's more like the reverse. ESR has provided some very good examples of this. Let me ask you this, what is the market value of Office '97 the day after Office 2000 comes out? What was it the day before? What is it the day after MS officially desupports it? In all those scenarios, we're talking about the same product, with the same features and benefits, but the price fluctuates a lot. Why? Could it be because the real value to the consumer is the support? Now, instead of looking at the profits of RedHat, look at the costs for the industry as a whole. Is it mostly production costs or support costs? Always support costs! Either that or you give shoddy support, tell your customers they're stupid, and quickly desupport old products and sell customers "upgrades". 20 years ago, some companies tried to change the economic model for software, and succeeded. However, it's not a good model, and the problems with it are very apparent these days. THAT is why RedHat is valued so high, even though it has no track record.
d) To break down patent holders into "evil" and "not evil" is about as ludicrous as Teddy Roosevelt's comments about "good trusts" and "bad trusts". It doesn't matter whether a company is deemed evil or not, it matters what the impact of their specific actions with patents are.
Ok. So, now I'm your entreprenuer (success is questionable, but that's another story...). I'll tell you what I've learned: most VC's still won't invest in a company without IP. Every startup tries to patent what they are doing, even if six different companies are doing the same thing. The reality though is that ideas are a dime a dozen. (Remember, the whole thing with patents is that they don't protect any of the blood sweat and tears you put in to developing an idea, what they really protect is the idea itself, and you get this protection NOT because you came up with it yourself, but because you filed it first.) The real thing that seperates success from failure is all the hard work that goes in to implementing an idea.
I see tons of great ideas every week, and I talk to IP lawyers and VC's who see tons every day. None of them sign NDA's, and they back it up by saying exactly what I'm saying. You want to succeed as an entrepreneur? Do market research, write an excellent business plan, hire the best people, build an excellent organization, establish the right relationships/partnerships. But whatever you do, don't think that all you need is a good idea and the right to sue everyone else about it.
If you wanna look at the source of a good JVM you can always look at Sun's if you don't mind agreeing to SCSL. If you'd rather not be chained, then I'd suggest grabbing the VM from Transvirtual. It's not perfect, but it's not bad.
Other good VM's for other languages: Squeak and CMU Lisp (not really a VM in some ways). E-mail me if you want pointers.
Good JVM's do incremental GC, which should prevent you from experiencing any noticable delays from GC. Java2 also allows you to use references to influence the GC routine. Finally, good JVM's run the GC in a seperate thread. Most of the time GC's will happen when the system is otherwise idle.
The truth is it's all in the VM, and Java, because of it's youth, has largely been populated by poor VM's.
The biggest problem with Meta-Object-Programming is that it makes it hard to integrate 3rd party tools. This isn't necessarily a bad thing, but it's a trade off.
The biggest problem with frameworks is that you're imposing a certain flow of control on developers. Again, this isn't necessarily a bad thing, but it does limit their flexibility. J2EE is a really good example of this. Starting a thread in Java is the most natural thing in the world, but in J2EE you basically can't do it inside of an Enterprise Java Bean. Consequently, the work todo the equivalent is EXTENSIVE to say the least. Now, this isn't a problem for many J2EE projects, but you simply can't be all things to all people.
I'm surprised that nobody has pointed this out yet: abortive closes ARE supported on Linux. They just aren't enabled by default, and that's a good thing. Abortive closes are great if you're a client running a benchmark, but if you're a server, you could receive packets at a port from a previous connection that will now appera to be coming from a new connection! Not a good thing.
Slashdot Mirror
What sort of books have you used as a touchstone for shaping your web site?
There's a lot of material out there, and a lot of experts, but who do you think really provided you with good advice on how to put together your website?
I thought there was now a web front end for this....
--Chris
Most of the political web sites I've seen out there are what I refer to in the web development industry as "brochure sites". They tend not to be terribly interactive, and mostly serve as an online brochure.
What can you do with a political site that really enhances your ability to get the message out there?
Can someone explain to me why stuff like this shouldn't be clear evidence to the FCC or trust-busters that there's a problem with this merger? This is the kind of behavior that should trigger an immediate response.
What's the point if it doesn't?
On the "betatest" newsgroups, there are literally thousands of posts with testers posting what they consider "bugs" which, in my opinion, are nothing but "niggly" things that doesn't deserve any attention.
I think this is a really good example of where open source has some advantages. Little "niggly" things get fixed by the person they're "niggling". ;-)
Well, the fact that escapist fiction talks about experiences that are illegal for the general populace is no surprise.
;-). However, computers are bigger than that community now. More people are on the Internet than develop the Internet... by an overwhelming margin.
;-) started talking about it, before guys in suits built entire economies on it, before banner ads, before SPAM, etc.
I think the role of "counter-culture" on the Internet is misleading. Certainly, the counter-culture has benefitted from the anarchy that is the Internet. Similar phenomena have happened in the past. I believe Dyson is quoted as say, "The Internet is great for conspiracies and bad for propoganda." I think she's right.
Certainly a lot of the Unix gods from the 70's were using drugs, like a lot of people their age at that time. Unix and the Internet were part of University culture for a long time, and the counter-culture is/was also strong there. Indeed, to this day geeks culture is still closely tied to University. It's no surprise that it's also tied to concepts that are strong in University.
However, I disagree with the notion that the computers are linked to the modern counterculture. There is definitely a lot of computer people who weren't part of mainstream culture, particularly when they were growing up (I still think being a geek may not be too trendy in school
But let's face it, the bulk of the people on the Internet use it to send people e-mails, buy things online, look up information on companies and consumer goods, and plan vacations. This is DEFINITELY not counter-culture stuff.
Not only that, but a lot of the CONTENT on the Internet these days is mainstream media, adverstising, marketing material from companies, etc. This is also the segment that's going to grow as more non-tech people come online.
The Internet was a brave new counter-culture world before AOL was in on it, before Al Gore (counter-culture visionary???
Today, the Internet *is* the world's culture. Every person who's a part of it participates in it, and that's now enough people that it can't help but be largely representative of mainstream culture.
I also don't agree that the Internet is an overwhelming sensory experience. TV can be, movies can be, but the Internet isn't. That's what the media wants to think the Internet is all about: flash and bang, but they're basically missing the point.
Ultimately, the Internet is a very intricate and structured piece of reality, and I don't see pushing yourself farther from reality as preparing you for that experience.
I'm not panning the drug experience, just the notion that it is in anyway connected to the Internet.
I'm talking about fundemental weaknesses being found in the encryption algorithms. Pre-WWII algorithms were breakable by any cryptanalyst out there with relative ease, regardless of key size, for about 100 years. The only uncrackable system was Enigma. Enigma was busted in roughly 5-10 years, at which point, it didn't matter what was done.
It is conceivable (and indeed, it's sort of expected) that at some point each of the algorithms out there will suddenly have simple solutions, at which case decryption will be trivial.
Similarly, with the advent of things like quantum computing, key length might become irrelavent regardlesss.
That's why you have the equivalent of a digital notary. Different systems work different ways, but essentially it always boils down to another party "vouching" for your signature (saying it's really you), and signing your signature.
One thing that is fairly easy to prove is that the same signature was used for multiple documents. So, for example, one could say, "well, if it wasn't you, then who signed for the mortgage on your house???!?!"
Presuming the telephone network is secure, particularly if you're not using digital signatures on it. ;-)
;-)
Even if it's not, all the dial-out thing would do would be to confirm that someone at that phone number picked up the phone and said, "sure". The only way to really be sure it was him/her would be if you used a digital signature.
Let's not even talk about how clunky this would be to implement.
Actually, if a digital signature was used the process of authenticating you with Outpost.com would be significantly easier. Think about this: that password you used to login.. how long was it? With computers doubling in speed every 18 months, the necessary length of a password is increasing at the rate of about one alphanumeric every 18 months (based on the fact that people tend to use english passwords and they have limited entropy).
;-) that you saw exactly what you were signing for.
Additionally, it's hard for Outpost.com to prove that you actually ordered a specific item. You can easily claim it was a mistake in their software and it looked like you were ordering something else. With a digital signature, there can be little doubt (barring a bug in your video driver
First of all, for those of you who are concerned that this opens up some huge problem in security because bits can be copied easily, please read up on digital signatures and how they work. Rest assured that provided you use them properly, it is VERY hard for someone to add your digital signature to another document (unless the contents of that document are bit-for-bit the same as one you've already signed --in which case, who cares?).
The concern I have is that this is based on what we CURRENTLY know about encryption technology. I've been reading up on the history of cryptography, and it really looks like a horse race between code makers and code breakers. The thing is, the code makers HAVEN'T been consistently winning. Indeed, if you look back in history prior to the 1970's, you'll find that there were very limited periods of time when code makers were winning, and frequently it was only for short periods of time.
What's going to happen when the inevitable happens? Particularly if cryptographers don't have a new discovery to replace the broken approach. Once the infrastructure of using digital cryptography is in place, it's going to be hard to undo it (case in point: how companies/governments/individuals elected to avoid Y2K problems by simply pulling the plug?).
I don't think this is a reason not to use digital signatures. I think it's a reason to start thinking NOW about how to handle the seemingly inevitable moment when someone figures out how to crack existing approaches... particularly if there is no replacement.
I grow tired of this debate. You're missing what I'm saying, and ignoring obvious facts.
...don't say it... BRAND NAME behind it!
First of all, Redhat for the most part DOESN'T write GPL software. For the most part, they PACKAGE GPL software. They do write GPL software to help broaden the market appeal, but 99% of the code on a RedHat CD was not developed by them.
As far as CD sales go, I'm not saying this is Red Hat's future. I am just demonstrating that the name brand has value. Red Hat sells the box set for $75 in stores, next to other boxes that sell for less. They still sell a lot of CD's. Red Hat sells (online) a guru-friendly version of Red Hat for $30. It comes wth no docs or add-ons, and yet it sells for 15x more than a cheapbytes CD. Amazingly, it sells very well. That my friend, is the power of a brand name, and in the OSS world, RedHat has the best one out there. Granted, it's a small market still, but that doesn't mean that Red Hat doesn't benefit from having that name brand out there.
My point about Deloitte & Touche and Anderson Consulting was that you CAN charge a premium in the consulting business without having proprietary knowledge, something which you said wasn't true.
Indeed, they have established impressive reputations, but to a large extent their reputation for computer systems work (to the extent that it's relevant in today's world) has been from perfromance in the past 15 years.
Cygnus Solutions, as I mentioned before, is a VERY well established consulting company with a very well respected reputation. They are THE guys that compay's go to for solutions built with GNU tools. This is a company that's been around for 11 years, with 180 employees, with offices around the world, all doing purely consulting. They are very well established in the embedded systems market, as well as various IT projects. They have an extensive client list, etc., etc. As the maintainers of gcc and glibc they have an unparalleled reputation for working with those tools (which you need to do if you want to play the Linux game). They are now owned by Red Hat. So Red Hat does indeed have a reputation. IT managers and managment HAVE ALREADY written checks based on this repuration. Small little companies you may have heard of like Nortel, SGI, Motorola, and I believe Nintendo.
I have never disagreed that the whole tech sector is overvalued, but I think that RedHat is proportionally valued to it's competition. I do think though, that the computer sector has proven that they have a significant difference with Bio Techs: initial investment is smaller, and results are more immediately available. The lack of both of these REALLY hurts the biotechs.
Other minor points: There's plenty of freely available documentation out there for Linux, and RedHat isn't the only distribution that comes with documentation. (Don't forget, when RedHat was establishing themselves, one of the most common ways to get Linux was to buy a "Learning Linux..." book with a Linux CD in the jacket.) Nonetheless, people pay a premium for RedHat. That's brand name strength. RedHat can use it to peddle more than CD's.
Coke and Pepsi are VERY similar formulas, and if you like one or another it wouldn't take you very long to make something that matches it. I don't know what you're getting your comment about carbonation from. As a kid, I worked at a fast food place and one of the things I did was work the fountains. I'm telling you, the only difference between Coke and Pepsi was the sack of syrop they came from. They both ran through the same carbonation machine. It is TERRIBLY relevant if it's people's imaginations that seperate the two, because if it is... that's a BRAND NAME that's really impacting them.
Finally, can you explain this to me: why would the market for virtually free Red Hat software be smaller than the market for $75 Red Hat software? (your point c). This makes no sense to me, unless you're saying that people will only buy Linux if they feel there's a
You can't be called a true SF reader until you've read the classics. These include:
;-)
Asimov's Foundation Series
Asimov's Robots Series
Heinlein's Stranger in a Strange Land
Clarke's 2001 (the rest of the series is a wash to me)
Herbert's Dune Series
Adam's Hitchiker's Guide To the Galaxy Series
I'd throw in Heinlein's "Moon is a Harsh Mistress" but not everybody agrees that's a classic.
Then there's a couple of modern books that make sense:
Gibson's Neuromancer
Stephenson's books, read at least one of Snow Crash, The Diamond Age, and Cryptomnicron
After that, you've established your SF credentials and are free to persue other works of the genre.
Do you really think there's much that's unique about Coke, Pepsi, or Heinz? PLEEAASE! The major differences between the first two is the amount of sugar put in to what they purchase. The profits of these companies also aren't relevant. In RedHat's sector, some of the most successful companies are not profitable, and still worth billions. (Case in point look at IPIXD... and they've never made money and a good portion of their valuation comes from bamboo.com, which has really only been in the this business model for 2 years --compared to RedHat which has been doing this for a while).
RedHat now owns Cygnus, who've essentially been THE support guys for GNU tools for the last decade. They're well established (and you'll notice that their executive team is taking over RedHat), and they are schrewd businessmen, and the history of that company demonstrates that. So RedHat's support is far from non-existent. Keep in mind they have several support contracts with OEM's as well.
If you think branding doesn't work with software, ten ask me how RedHat has sold a SINGLE CD-ROM! You can get the same thing from cheapbytes for 1/30th the price! You can download it for free off the Internet! Still, RedHat sells CD-ROMs. They also sell T-shirts, caps, etc. They actually make money doing it too. They're brand name is already worth a lot. That's why Dell signed them up, why companies pay to get the "runs on RedHat" write up, etc.
Consulting company which charges a premium for it's servicse on a large scale w/o any proprietary technology: Andersen Consulting. Indeed, if you talk to them they make money BECAUSE they don't have any proprietary technology. Deloitte & Touche have a pretty good business in this area as well. It happens. You do a good job, you get a reputation, and you charge for it.
Think of it this way, you're a company, you want to do something with Linux, who are you going to call to make a deal? If you call RedHat you know your stock will go up just that much more. RedHat can make a little bit of money off of every company that wants to do business in the Linux buxiness space if they play their cards right, and the best part is the more they do it, the more their brand name is even more valuable.
As a final point: I bet you that just about anyone can make better ketchup than Heinz if they try. They just simply don't.
Well, Redhat's valuation is crazy, but I don't think it's crazy when you compare it to it's competitors. The market is crazy right now, and so valuations reflect that. Redhat has significant first-mover advantages, and has the highest name recognition in it's space. It's also aquired Cygnus, which is a well proven business, and will be the source of a lot of their revenue.
Redhat's game isn't just to provide service. What they really provide is "Branding". "Branding" has justified the valuations of companies like Coke, Pepsi, and Heinz, so I don't think Redhat should be any different. Redhat doesn't just brand their support. They use their brand to charge a premium for services, and in the long run that can be VERY profitable.
The 'Net has the advantage that there's nothing to seize. The feds would have to seize 4/5ths of the network to prevent access to it's contents.
Yup. There's a stegonographic filesystem out there for Linux.
Man, if ever I saw a reason why steganography, this is it.
;-)
I'd like to suggest a solution to this problem. Let's call it the "Redundant Distributed Network Steganographic File System", mostly because it's an acronym that can't be made into a cute name.
Now, the idea is, everyone gives up some disk space, say 0.5gb, which they make acessible on the Internet. In exchange, they get like 0.1gb of space on the RDNSFS. The filesystem LOOKS like noise, but if you have the right key, you can extract a certain amount of data from it. If you have the wrong key, you get fake data.
Now, the big problem is how to allocate space for someone without giving away that they have data out there somewhere....
Ok. first of all, I *am* an entrepreneur. I also work with entrepreneurs on a regular basis. I still very much disagree with your premises:
a) A startup entrepreneur can definitely apply for a patent. A college student experimenting in his apartment cannot. Both are at a disadvantage to a major corp., because it's essentially out there trying to plug all the possible holes in it's IP, whereas they're going to apply for the one or two patents that really embody their idea. In the long run, the corp will have more leverage.
b) I think most people on Slashdot are aware of this. The point is that it's a tragedy. The idea of patents is SUPPOSED to be to encourage the exchange of ideas, not patent suits. What you're saying is that the Patent Office has no real credibility, and that needs to change.
c) The suggestion that we scrap old business models and go with new ones ISN'T because of RedHat's success. Indeed, it's more like the reverse. ESR has provided some very good examples of this. Let me ask you this, what is the market value of Office '97 the day after Office 2000 comes out? What was it the day before? What is it the day after MS officially desupports it? In all those scenarios, we're talking about the same product, with the same features and benefits, but the price fluctuates a lot. Why? Could it be because the real value to the consumer is the support? Now, instead of looking at the profits of RedHat, look at the costs for the industry as a whole. Is it mostly production costs or support costs? Always support costs! Either that or you give shoddy support, tell your customers they're stupid, and quickly desupport old products and sell customers "upgrades". 20 years ago, some companies tried to change the economic model for software, and succeeded. However, it's not a good model, and the problems with it are very apparent these days. THAT is why RedHat is valued so high, even though it has no track record.
d) To break down patent holders into "evil" and "not evil" is about as ludicrous as Teddy Roosevelt's comments about "good trusts" and "bad trusts". It doesn't matter whether a company is deemed evil or not, it matters what the impact of their specific actions with patents are.
Ok. So, now I'm your entreprenuer (success is questionable, but that's another story...). I'll tell you what I've learned: most VC's still won't invest in a company without IP. Every startup tries to patent what they are doing, even if six different companies are doing the same thing. The reality though is that ideas are a dime a dozen. (Remember, the whole thing with patents is that they don't protect any of the blood sweat and tears you put in to developing an idea, what they really protect is the idea itself, and you get this protection NOT because you came up with it yourself, but because you filed it first.) The real thing that seperates success from failure is all the hard work that goes in to implementing an idea.
I see tons of great ideas every week, and I talk to IP lawyers and VC's who see tons every day. None of them sign NDA's, and they back it up by saying exactly what I'm saying. You want to succeed as an entrepreneur? Do market research, write an excellent business plan, hire the best people, build an excellent organization, establish the right relationships/partnerships. But whatever you do, don't think that all you need is a good idea and the right to sue everyone else about it.
If you wanna look at the source of a good JVM you can always look at Sun's if you don't mind agreeing to SCSL. If you'd rather not be chained, then I'd suggest grabbing the VM from Transvirtual. It's not perfect, but it's not bad.
Other good VM's for other languages: Squeak and CMU Lisp (not really a VM in some ways). E-mail me if you want pointers.
Good JVM's do incremental GC, which should prevent you from experiencing any noticable delays from GC. Java2 also allows you to use references to influence the GC routine. Finally, good JVM's run the GC in a seperate thread. Most of the time GC's will happen when the system is otherwise idle.
The truth is it's all in the VM, and Java, because of it's youth, has largely been populated by poor VM's.
The biggest problem with Meta-Object-Programming is that it makes it hard to integrate 3rd party tools. This isn't necessarily a bad thing, but it's a trade off.
The biggest problem with frameworks is that you're imposing a certain flow of control on developers. Again, this isn't necessarily a bad thing, but it does limit their flexibility. J2EE is a really good example of this. Starting a thread in Java is the most natural thing in the world, but in J2EE you basically can't do it inside of an Enterprise Java Bean. Consequently, the work todo the equivalent is EXTENSIVE to say the least. Now, this isn't a problem for many J2EE projects, but you simply can't be all things to all people.
All of what he talks about as "NextGeneration" was all stuff I learned 10 years ago when I first picked up OOP. Back then this stuff was old hat.
Indeed, time has proven that Framework-based approaches and Meta-Object-Programming are not a silver bullet eithr. They all have trade offs.
I'm surprised that nobody has pointed this out yet: abortive closes ARE supported on Linux. They just aren't enabled by default, and that's a good thing.
Abortive closes are great if you're a client running a benchmark, but if you're a server, you could receive packets at a port from a previous connection that will now appera to be coming from a new connection! Not a good thing.