>
TP monitors like Tuxedo or Encina are necessary in about 5% of database application.
Yet these tend to be quite important applications. Especially if one wants consolidate platforms, this can rule out a platform. It is even a problem to get a totally free software environment.
>
What percentage of BEA's sales do you think that product is? I'll give you a hint: It's far less than 2%. WebLogic Enterprise is a shell game application masquerading as an upgrade path for WebLogic customers who need "Enterprise" support. NONE of their customers have ever ported from WebLogic to WebLogic enterprise. Well, unless you're talking about people who would accept a SINGLE THREADED EJB container, which when coupled with the fact that the EJB spec says you can't have threads *in* a bean, basically means you have more containers running concurrently than a rabbit makes bunnies.
It might be trash, but the fact is that it has TP monitor built in. And it might be seldom used, but this is not an argument about Mac OS X not being able to cater for 80% of the users; it is about it needing more diversity to be able to cater for, say, 99%.
>
Why the hell would anyone run DB2 on an Xserve? Note I said "middle tier"
I don't care about the middle tier only. I wouldn't choose a platform for the middle tier if I can't use it for the other two as well, given GNU/Linux or BSD, and even Solaris, can do it all.
>
If you want to run a database on an Xserve, you have a few choices: Sybase, Oracle, MySQL, PostgreSQL can do some fairly heavy lifting
Yet there are some applications where I'd rather run IBM DB2 instead of any of the others. PostgreSQL is great but not yet to the same level of DB2; Oracle doesn't even have a standard type system, being kinda like the MS of DBs ('almost standard'); MySQL is still a bad joke in my book.
>
Xserve supports X11, first and foremost.
Yes, it supports X. But it's a second-class citizen. I don't want to support bad behaviour like de facto, unnecessary 'standards' and the associated bloat.
>
It plays well and it works with databases on Solaris...
Again, why would I want one more platform?
>
When you consider the price of 8 clustered Xserves running JBoss+MySQL...
I get the feeling you don't really understand databases or big iron... clusters are not a substitute to big iron, as they are more costly to run, and more complex. MySQL is not up to snuff, and their developers don't get databases.
>
I get the feeling you haven't looked at an Xserve or OS X...
I have. Great eye candy, but give me GNU/Linux, give me freedom, stop hyping and breaking promises, give me alternative sources of supply.
But it doesn't have to be Mozilla. The answer also didn't took in account the fact that many users, and even organisations, use nothing but email from MS Outlook.
>
no reason WebLogic shouldn't run on OS X. It's written entirely in Java
BEA Engine is a fundamental part of WebLogic Enterprise. It is C.
>
BEA might not support such a configuration, but it should work.
Wanna run unsupported? Good luck!
>
Care to back up your claims about high reliability, or lack thereof, with Apple Xserves?
Learn a little bit about how a real corporate server is made. MTBF, monitoring, SCSI or better, clustering, even the processors are more reliable then their desktop counterparts.
>
your real argument emerges. It ain't Free
This is part of it, but if you can't understand a platform needs diversity and alternatives to thrive, good luck...
Only if you don't need Tuxedo, WebLogic, and other insignificant stuff like DB2, etc... and you don't need scalability, you don't plan to consolidate, and you don't need high reliability.
I wonder how Apple insists on being a platform by themselves. If Mac OS X was -- all of it -- free software, or if they used the X Window Server, or if they just used Debian, or if they licensed it, then they would have a viable platform. Lots of opportunities, and Apple insists on loosing all of them.
>
Did you also offer to write a centralized, shared calendaring package, shared address book, full contact management system and meeting scheduler for them?
There are alternatives.
>
Or did you just think they used Outlook for email?
Ten years ago I was starting on MS DOS (MS Windows capable machines were too expensive at Brazil, and GNU/Linux I had just heard about from my little brother then at the University), and the following year I started working with IBM mainframes.
I had expected, especially with GNU/Linux, to combine the simplicity of MS-DOS with the mainframe's reliability and a relational database, and we are nowhere near that.
GNU/Linux is still incredibly complex and buggy, still taking loads of memory, PC hardware is still crap, and X Windows still can't resume a session like the mainframe could, and I hear the Sun XRay can. We are still stuck with SQL, and instead of going relational people are making the situation even worse with network data model disguised as OO.
I am decided to only buy top-quality hardware from now on: RISC machines with SCSI disks and so on. It doesn't matter I can get a fast CISC IDE system, I want RISC silence and SCSI reliability, and speed really isn't the issue anymore. I want to build a X host and have silent, simple X terminals, hoping XFree 5.0 will have a protocol for resuming sessions -- not the crappy session management now in Gnome, but really being able to loose power on the X terminal, turn it on again and being able to begin with the cursor where it was.
I suspect part of the complexity is due to POSIX itself. I hope the Hurd will help us migrate to a Lisp world, but meanwhile I can't really even Emacs working exactly as I want with IMAP and authenticated SMTP; I suspect the transition from GNU/Linux to GNU/Hurd will be a pain, and then the migration from POSIX to Lisp (as I am sure RMS and other hackers dream) will be another.
I also expect to have a relational OS... perhaps the Hurd can be a platform to that, where all data structures will be simple relations with arbitrarily complex data types. That, coupled with a Lisp or otherwise functional programming model, could be heaven, and much simpler and more reliable than we have now... only dreaming.
>
IBM's patents aren't transferred or licensed to SCO, but if they write code for a System V derivative work (such as AIX) they can't use the code in Linux.
Actually SCO mention their contract with Sequent. A lawyer should tell us if Sequent's terms bound IBM in Sequent-related code, or if IBM's terms take effect to all subsidiaries bought by IBM.
>
Unless publicly and freely distributing code to anonymous parties without a written contract or a requirement for monetary consideration was consider by the court as legally equivalent to putting the source code in the public domain.
What would be the precedent? Books are in the library for anyone to take, but no one can make a copy.
Anyway the code is there to take, but the GNU GPL accompanies each package stating the conditions to redistribute.
>
You are not agreeing to a contract. The GPL wants you to think that you are
The GNU GPL wants nothing, because it's not a person.
Now, there is nothing I can see in the GNU GPL to the effect that its drafters wanted to make it stick as a contract. It is a grant of rights, AKA as a license, nothing more. It is the EULAs of life that want to be contracts; that's why it is not EUL(icense), but EULA(greement).
In a contract both parties agree on something. Make a sign in the bus stop; the bus driver agrees on stopping, you agree on taking the bus or at least saying 'sorry, wrong line'.
The GNU GPL is just a grant of rights. Under copy right law you have no rights to copy, so the copy right holder grants you these rights if you abide by these conditions. Violate them, you loose the rights. It's that simple.
It's kinda like with murder. You had no contract with humankind or the government granting you the right to live. But kill someone, and your right is revoked. There is still a lengthy process to determine if it was murder, and there might be atenuant circumstances, but murderers in principle forfeit their right to live.
>
the remedy for this violation of the copyright law does not include forcing them to comply with the GPL
In this case, there would be no right to distribute. In SCO's case, they would be enforced from distributing Linux. They would have to ask nicely for the persecuting copyright holders to allow them doing so.
>
Suppose that it was a patch to the MySQL codebase.
Then use PostgreSQL.
>
The details of a hypothetical example are not particularly important.
Agreed. In general, there is always a piece of non-copylefted software one can hoard. Going for the GNU GPL stuff just makes no sense.
Now suppose the world had turned GNU GPL and all free licenses were copyleft. Even so, the cost of maintaining the patches, and the fact that you can usually work around stuff would be enough to discourage people.
Anyway, what is required from the GNU GPL is that it be good enough. Perfection is not attainable in this fallen world. Ultimately the scenario you paint is relevant if one presupposes a valuable patent; this is quite out of the scope for the GNU GPL, and must be dealt with political action to change the patent law.
It is not a contract, but a license. It asks nothing of you, it just states the rights you have. Violate the conditions, and the situation returns to the default, which under copy right law is no rights to copy...
Yes, but depending on your beliefs some decisions can be quite troubling in that they set a trend. My worries are much older, even if I'm not a citizen of the US: they come from Roe vs Wade and, later, the method patents.
We have our own media circus: ZDNet, CNet and friends... not to mention a hoard of news sites.
>
I don't know what you're talking about
The Bible calls that "a conscience seared with hot iron".
>
People and governments have always been as mean and evil as they are today
About people, agreed, at least at the depth of our hearts. But there have always been a kind of salutary hypocrisy: La Rochefoucauld's "hypocrisy is the tribute that vice pays to virtue" comes to mind. This acknowledgement of virtue is sorely missing, and confounds even supposedly conservative judges to judge based solely on economical basis and to give free rein to power. These are not the times were antitrust or public domain are given any serious consideration, and we are poorer for that.
>
the final omen would not be the courts striking down the GPL
Rome wasn't built in a day. Nor destroyed.
>
wearing a Che Guevara T-shirt right now?
I'm a culturally conservative, doctrinally orthodox, economically liberal (in the European sense) Baptist.
Not yet, but unless IBM abandons at least its claim about SCO infringing the GNU GPL, or SCO bucks, stop redistributing and negotiate a license with all interested owner of distributed code, then it might come to a test case. Which obviously isn't necessary, as per Moglen's famous paper on the GNU GPL enforceability...
>
Are the claims of either side based on the (non-)validity of the GPL
Yes. When IBM pointed finger at SCO infringing distribution of Linux, SCO tried to defend its illegal, immoral act by saying the GNU GPL is a vague license. Well, if so SCO shouldn't being redistributing any GNU GPL'd code, as it wouldn't be clear they have the rights to do so.
Slashdot Mirror
Just make them rich! Look at Europe, filthy rich, burning its riches in subsidies, and breeding like pandas.
You are asking for a level of precision and comprehensiveness that's not usually worthwhile in /. posts. Nitpicking, in short.
No, free apps are not yet perfectly integrated. But yes, they are good enough. And yes, virus avoidance and license costs are worthwhile.
Yet these tend to be quite important applications. Especially if one wants consolidate platforms, this can rule out a platform. It is even a problem to get a totally free software environment.
It might be trash, but the fact is that it has TP monitor built in. And it might be seldom used, but this is not an argument about Mac OS X not being able to cater for 80% of the users; it is about it needing more diversity to be able to cater for, say, 99%.
I don't care about the middle tier only. I wouldn't choose a platform for the middle tier if I can't use it for the other two as well, given GNU/Linux or BSD, and even Solaris, can do it all.
Yet there are some applications where I'd rather run IBM DB2 instead of any of the others. PostgreSQL is great but not yet to the same level of DB2; Oracle doesn't even have a standard type system, being kinda like the MS of DBs ('almost standard'); MySQL is still a bad joke in my book.
Yes, it supports X. But it's a second-class citizen. I don't want to support bad behaviour like de facto, unnecessary 'standards' and the associated bloat.
Again, why would I want one more platform?
I get the feeling you don't really understand databases or big iron... clusters are not a substitute to big iron, as they are more costly to run, and more complex. MySQL is not up to snuff, and their developers don't get databases.
I have. Great eye candy, but give me GNU/Linux, give me freedom, stop hyping and breaking promises, give me alternative sources of supply.
May be... so what? The fact is that WebLogic and other stuff is a sine qua non to many. My argument isn't about only a product.
But it doesn't have to be Mozilla. The answer also didn't took in account the fact that many users, and even organisations, use nothing but email from MS Outlook.
BEA Engine is a fundamental part of WebLogic Enterprise. It is C.
Wanna run unsupported? Good luck!
Learn a little bit about how a real corporate server is made. MTBF, monitoring, SCSI or better, clustering, even the processors are more reliable then their desktop counterparts.
This is part of it, but if you can't understand a platform needs diversity and alternatives to thrive, good luck...
Only if you don't need Tuxedo, WebLogic, and other insignificant stuff like DB2, etc... and you don't need scalability, you don't plan to consolidate, and you don't need high reliability.
I wonder how Apple insists on being a platform by themselves. If Mac OS X was -- all of it -- free software, or if they used the X Window Server, or if they just used Debian, or if they licensed it, then they would have a viable platform. Lots of opportunities, and Apple insists on loosing all of them.
There are alternatives.
Very often.
Ten years ago I was starting on MS DOS (MS Windows capable machines were too expensive at Brazil, and GNU/Linux I had just heard about from my little brother then at the University), and the following year I started working with IBM mainframes.
I had expected, especially with GNU/Linux, to combine the simplicity of MS-DOS with the mainframe's reliability and a relational database, and we are nowhere near that.
GNU/Linux is still incredibly complex and buggy, still taking loads of memory, PC hardware is still crap, and X Windows still can't resume a session like the mainframe could, and I hear the Sun XRay can. We are still stuck with SQL, and instead of going relational people are making the situation even worse with network data model disguised as OO.
I am decided to only buy top-quality hardware from now on: RISC machines with SCSI disks and so on. It doesn't matter I can get a fast CISC IDE system, I want RISC silence and SCSI reliability, and speed really isn't the issue anymore. I want to build a X host and have silent, simple X terminals, hoping XFree 5.0 will have a protocol for resuming sessions -- not the crappy session management now in Gnome, but really being able to loose power on the X terminal, turn it on again and being able to begin with the cursor where it was.
I suspect part of the complexity is due to POSIX itself. I hope the Hurd will help us migrate to a Lisp world, but meanwhile I can't really even Emacs working exactly as I want with IMAP and authenticated SMTP; I suspect the transition from GNU/Linux to GNU/Hurd will be a pain, and then the migration from POSIX to Lisp (as I am sure RMS and other hackers dream) will be another.
I also expect to have a relational OS... perhaps the Hurd can be a platform to that, where all data structures will be simple relations with arbitrarily complex data types. That, coupled with a Lisp or otherwise functional programming model, could be heaven, and much simpler and more reliable than we have now... only dreaming.
Ah, and formal methods such as Z.
Great, thanks! Already sent.
Suggestion: each one please do change the text as to look like you actually thought about the issues... we don't want to be called astroturfers.
Not quite... only the GNU GPL asserts the code must be unencumbered.
What would be the precedent? Books are in the library for anyone to take, but no one can make a copy.
Anyway the code is there to take, but the GNU GPL accompanies each package stating the conditions to redistribute.
Your imagination is running too fast...
The GNU GPL wants nothing, because it's not a person.
Now, there is nothing I can see in the GNU GPL to the effect that its drafters wanted to make it stick as a contract. It is a grant of rights, AKA as a license, nothing more. It is the EULAs of life that want to be contracts; that's why it is not EUL(icense), but EULA(greement).
In a contract both parties agree on something. Make a sign in the bus stop; the bus driver agrees on stopping, you agree on taking the bus or at least saying 'sorry, wrong line'.
The GNU GPL is just a grant of rights. Under copy right law you have no rights to copy, so the copy right holder grants you these rights if you abide by these conditions. Violate them, you loose the rights. It's that simple.
It's kinda like with murder. You had no contract with humankind or the government granting you the right to live. But kill someone, and your right is revoked. There is still a lengthy process to determine if it was murder, and there might be atenuant circumstances, but murderers in principle forfeit their right to live.
In this case, there would be no right to distribute. In SCO's case, they would be enforced from distributing Linux. They would have to ask nicely for the persecuting copyright holders to allow them doing so.
Then use PostgreSQL.
Agreed. In general, there is always a piece of non-copylefted software one can hoard. Going for the GNU GPL stuff just makes no sense.
Now suppose the world had turned GNU GPL and all free licenses were copyleft. Even so, the cost of maintaining the patches, and the fact that you can usually work around stuff would be enough to discourage people.
Anyway, what is required from the GNU GPL is that it be good enough. Perfection is not attainable in this fallen world. Ultimately the scenario you paint is relevant if one presupposes a valuable patent; this is quite out of the scope for the GNU GPL, and must be dealt with political action to change the patent law.
It is not a contract, but a license. It asks nothing of you, it just states the rights you have. Violate the conditions, and the situation returns to the default, which under copy right law is no rights to copy...
Yes, but depending on your beliefs some decisions can be quite troubling in that they set a trend. My worries are much older, even if I'm not a citizen of the US: they come from Roe vs Wade and, later, the method patents.
We have our own media circus: ZDNet, CNet and friends... not to mention a hoard of news sites.
The Bible calls that "a conscience seared with hot iron".
About people, agreed, at least at the depth of our hearts. But there have always been a kind of salutary hypocrisy: La Rochefoucauld's "hypocrisy is the tribute that vice pays to virtue" comes to mind. This acknowledgement of virtue is sorely missing, and confounds even supposedly conservative judges to judge based solely on economical basis and to give free rein to power. These are not the times were antitrust or public domain are given any serious consideration, and we are poorer for that.
Rome wasn't built in a day. Nor destroyed.
I'm a culturally conservative, doctrinally orthodox, economically liberal (in the European sense) Baptist.
In this case it would be much simpler to just use BSD. Not so important.
Yes, and it will be very hard to keep the patches against an ever-changing codebase. Not impossible, but unlikely.
How on Earth can this be rated +1 Funny?!?
Not yet, but unless IBM abandons at least its claim about SCO infringing the GNU GPL, or SCO bucks, stop redistributing and negotiate a license with all interested owner of distributed code, then it might come to a test case. Which obviously isn't necessary, as per Moglen's famous paper on the GNU GPL enforceability...
Yes. When IBM pointed finger at SCO infringing distribution of Linux, SCO tried to defend its illegal, immoral act by saying the GNU GPL is a vague license. Well, if so SCO shouldn't being redistributing any GNU GPL'd code, as it wouldn't be clear they have the rights to do so.
Some of us like to eat. Actually, even GNU needs money, which Stallman initially provided by lecturing on the gcc.