If they're not using gcc or a patched gcc compiler, I'm really surprised. Many commercial compilers are nowhere near as reliable or flexible as gcc, and often fail miserably when handling cross-platform compilation in cases where gcc does so quite conveniently. It's usually much easier to patch gcc to support a new hardware than to purchase and retrain for a new proprietary compiler, especially if the hardware platform being compiled for is brand new.
This makes it interesting if they're selling or publishing for partners a developer kit with a compiler: then, they'd have to publish the patched gcc-based compiler.
Wildly off-topic, but given that more people are dying on the streets *now* in Iraq than during Sadam's reign, I don't see how it's a step up for Iraq. And given that our "ally against terror" Pakistan has been sellingn nuclear weapons facilities and technologies to any 3rd world nation that can scrape up the money and is harboring Osama bin Laden, it's not like the world is safer.
They used to be good at this: they did win against the Secret Service for the wildly mis-aimed raids against Steve Jackson Games. And they've done other useful things since they dumped Jerry Berman as their leader: that man was clearly selling their soul to get lobbying money from telecoms and have a nice office in DC.
But they've gotten better, and seem to be aiming at serious issues: even where they don't win, they're bringing awareness to serious issues, and getting people to bring up these issues at budget time in Congress.
Unfortunately, such encryption isn't ubiquitous enough to protect non-critical transmissions: normal business and personal email, cell phone conversation, and many IM-style tools aren't remotely secure enough. Even for common encryption techniques such as SSL, the keys are rarely well-protected enough to prevent "law enforcement" from stealing them and monitring traffic secretly.
There are technologically reasonable techniques, such as the so-called "Trusted Computing" tools that are growing in use, but notice that those require signed, registered keys from companies like Microsoft. That's no protection whatsoever against abuse by federal agencies with access to the fundamental key signature authorities. It's a real political and technical problem: it's very difficult to get approval for general export for any tools that include robust encryption without gaping security holes for "approved" federal access.
Any fool can write a basic tool or two. But making it robust, making it reliable, and making the interface simple to use takes serious work. The result of writing really good tools that have terrible interfaces is well-described by Eric Raymond in his famous rant on CUPS: if the interface is not good, people will not be able to use it reliably.
Have you factored in the risks and losses of Peachtree? I found it to be very unstable, and vulnerable to records being overwritten or destroyed by two people accessing it at the same time. I finally had to urge the accountant I was working with to install it on and only run it on one desktop, and arrange for regular backups which Peachtree didn't support very well at all.
Mr. Stallman is not the FSF: he is its founder. This means that the organizatin may have to acknwledge and deal with nasty situations in ways that he really wishes were not necessary, because they don't have the tools to do it a better way. This includes making very clear what the GPL license applies to, in order that it not be declared invalid because it overreaches, and dealing with distressing problems like the insistence f NVidia on keeping their drivers closed source, but seeing a big demand in the free sftware world for NVidia drivers.
X doesn't do cross platform well. And in particular, VNC tends to work better over slow connections than X. Also, VNC supports a purely Java based client, available from any Java capable web browser, so it doesn't require client installatioon.
Develpment workstations also run CentOS. You get binary compatibility with RHEL, but you can save your license fees for the production servers and not waste it on development environments.
It looks like this poster had their workgroup changed as a committee decision where some really unfortunate issues got overlooked, like the need to pay for closed source licenses for Qt, C#, and CygWin to sell their own product as closed source.
This original poster is probably trying to build a user interface for a portable device where they don't want to publish their source code. That means they need to use libraries that are not under GPL. Ergo, they need Qt or a similar license, and that's why the Qt license costs so much rather than using open source and publishing open source.
This person wants to reap the benefits of open source development, without opening their own code up for similar development. This kind of thing is *exactly* why the GPL exists: to keep closed source developers like this from harvesting the cream of open source and then locking it up for their private products.
It's also apples and oranges. RedHat server versions are comparable to Windows 2003, plus MS Office with the built-in OpenOffice, plus SNMP services, plus a print server, plus an industrial grade mail and webmail and IMAP server, plus a backup server with the built-in Amanda system, plus an industrial grade file server, plus industrial grade firewalls and security tools, plus good CD and DVD ripping sftware, and you don't have to buy additional client licenses if you have more than 25 clients. Moreover, almost all of RedHat server tools are available free as part of the CentOS distribution, if you don't want the commercial level of support and would rather use the tools free. I've actually used CentOS to demonstrate RedHat tools, before urging a client to go ahead and buy a licensed RedHat system in order to get that commercial grade of support after my contract is ver.
No, where the original poster got messed up was in trying to run commerical style software on an open source system. QT licensing is the same cost as several weeks of on-site consulting time, and should frankly be replaced in most projects with a simple web interface for portability. The Embedded Linux software is very customized, and very project specific: that's why it costs so much. Rewriting everything in C# is one of the stupidest things I've ever heard of: it's a Microsoft written language, it's *designed* to frce you to use proprietary tools, and it's a Java wannabe. If these folks had been clueful, they'd have rewritten in plain C for speed and portability, or C++ for object oriented code, or actual Java for the write-once, run-anywhere advantages. But writing anything in C# that can be avoided. And the only reason to use Qt, and its licensing, is to allow you to keep your source code unavailable.
The CygWin license cost quoted is misleading as well. The only license for CygWin that costs that much is if you want to use CygWin to publish binaries without providing source code. The licensing is described at http://www.cygwin.com/licensing.html.
This original poster was basically trying to build a closed source environment, and paying the premium to insist on keeping their environment closed. Of course that will be expensive! It's not taking advantage of the open nature of the systems at all.
Keep an eye on the BIOS. The develpers for Trusted Computing are very eager to support Trusted Computing based control of the basic haardware, to prevent the use of plug-in bootable media on secured systems. They're also eager to have the BIOS, itself, be Trusted Computing signed althugh this is much more difficult in the short term. And they're eager to have CD and DVD drives tied into it for media authentication, and have less influence but still are pursuing doing so for disk drives, to reduce access to stolen disks on secure systems. This is a long-term scheme, not a short term one: simply because the BIOS's you've worked with don't have such features doesn't mean that the next generation of drives and BIOS's won't use that.
Let me reiterate: trusted Computingn can manage keys for hardware authentication. Such abilities are key to its use for security, which is the pretext for its development, as well as for the DRM which is its real purpose. And such key management is clearly designed to authenticate bootable devices, even if such tools have not yet been implemented.
There are benefits of Trusted Computing: but those benefits are overwhelmed by really unfrtunate design choices aimed at securing vendor desires to control the use of their software and hardware, not those of consumers.
You can tie it to a host. This is one of the major purposes of Trusted Computing. It's use for encryption and reliable authentication is really secondary, its clear goal is DRM.
Whoah indeed. There are plenty of reasons not involving Microsoft for manufacturers to want DRM, and they've done it with non-Microsoft related technologies. I don't mean to cast it all on Microsoft.
But the nature of the Trusted Computing reflects a very clear plan to use it for DRM: Microsft wants it both for protecting their own software from being used in pirate copies, and to mate Windows software to other Windows software that blocks competing products from accessing them. This is particularly true for streaming media: by locking multimedia files and media to specifically, "Trusted Computing" software that has to be authorized by the mother ship, they gain the ability to prevent Putting "Trusted Computing" software on a Linux box does no good for this. Trusted Computing's approach is much more limiting than a standard DVD player: the media can be registered to a specific copy of the media, with a specific software viewer, on a specific host, andn simply refuse to allow viewing it under any software not specifically supported by the media vendor. It may be possible to do screen capture of the output, but that's about it.
Such capabilities can also be integrated, and are clearly to be integrated, in hardware. This seems reasonable for VPN identification keys or biometrics. But it's obviusly planned as well for CD and DVD burners, to force the duplicator to use authorized software to read the files, and to permit only mothership authorized software to write with that specific burner.
It's also aimed squarely at the boot system: a Trusted Computing enabled chipset on the motherboard can prevent the booting of any kernel, or bootable device, without an appropriate vendor signature. You'd better believe Microsoft wants this to prevent "Trusted Computing" hardware from booting non-authorized operating systems. It's understandable for security reasons in a corporate environment, but it's deadly for home users or experimenters who may want to install operating systems without the capital to buy and manage Trusted Computing keys.
Patents on software are like copyrights on individual words: they actively hinder development of new works, and add a serious burden to new software projects to guess whether or not some obscure patent will be held to be infringed by the new work.
Oh, yes. The changes when the US eliminated the need to register copyrighted documents were quite sufficient. And given the amount and ease of creating new documents in this digital world, 20 years seems quite sufficient.
Oh, agreed. That legislation in the US was aimed squarely at protecting the Mickey Mouse copyright: Disney is a huge sponsr of such legislation, as that darned mouse gets older and older and Disney's library of old movies grows in value. Librarians I've spoken with are amazed at the difficulties current US copyright law creates for them.
But it was a good question: the companies are the same companies as elsewhere, both US and multi-national. Sony, Disney, Samsung, and other companies interested in DRM protecting their digital investment help prevent what folks in the US call "fair rights" use.
These desires for DRM are supported heavily by Microsoft, which wants to protect its software licenses and prevent other software from writing or even reading the private formats they use, especially for "personally encrypted" files. The whole field of DRM is about to get much worse with the "Trusted Computing" software program, led by Microsoft, which locks software and media to a specific motherboard's encryption and authentication chip, and which will als be built into the next generation of AMD and Intel CPU's. These tools are aimed squarely at preventing duplication of digital media, and create a real burden for making backup or archival copies of these media. They also specifically restrict the authorized software tools to play these media, making "fair use" sampling or even using other non-specifically licensed players very difficult.
The libraries of the world are quite right to be concerned about this.
You had a bad secretary. A good one will automatically juggle your meetings, your costs, and the pain to your family of being gone on your wife's birthday. And that secretary can do that for 20 people for the same conference and get you a discount rate that helps justify his salary, along with knowing where to get decent local coffee and book a hotel room for a visiting guest on Thanksgiving weekend.
Such people are priceless, just as a really competent IT person who can make the printers work right for Powerpoint slides in a Russian edition of Windows is also priceless.
Having all the engineers do 5% of their time managing is a disaster in the offing: the one of them who enjoys office politics, and is good at gathering resources, will wind up as CEO. If they have other qualities that are really bad, they will still succeed in becoming CEO, and destroy the company. A few engineers that are also good managers can make a company wildly successful, but these are rare.
They've tried for years to prevent it, and keep trying. The regulations against exporting encryption have a chilling effect, and various clauses of the legislation that included the Telecommunications Privacy Act were amazing in what they tried to forbid.
The HitchHiker'suote is a piece of common protester slang from yhe 1960's or earlier. It was usually spouted by urban college graduates spouting the evils of capitalism and shouting the glories of socialism and the coming Marxist revolution while mommy and daddy paid for their food, tuition bills, and the gas for their VW microbus.
Douglas Adams was old enough to remember those days, and to mock them as needed.
Slashdot Mirror
That sounds like a thermal problem. Have you tried opening the case and cleaning the air vents, and making sure the fans are all working properly?
Why ever not? We post to Slashdot, you'd think we'd be used to unstable blasts of hot air.
If they're not using gcc or a patched gcc compiler, I'm really surprised. Many commercial compilers are nowhere near as reliable or flexible as gcc, and often fail miserably when handling cross-platform compilation in cases where gcc does so quite conveniently. It's usually much easier to patch gcc to support a new hardware than to purchase and retrain for a new proprietary compiler, especially if the hardware platform being compiled for is brand new.
This makes it interesting if they're selling or publishing for partners a developer kit with a compiler: then, they'd have to publish the patched gcc-based compiler.
Wildly off-topic, but given that more people are dying on the streets *now* in Iraq than during Sadam's reign, I don't see how it's a step up for Iraq. And given that our "ally against terror" Pakistan has been sellingn nuclear weapons facilities and technologies to any 3rd world nation that can scrape up the money and is harboring Osama bin Laden, it's not like the world is safer.
They used to be good at this: they did win against the Secret Service for the wildly mis-aimed raids against Steve Jackson Games. And they've done other useful things since they dumped Jerry Berman as their leader: that man was clearly selling their soul to get lobbying money from telecoms and have a nice office in DC. But they've gotten better, and seem to be aiming at serious issues: even where they don't win, they're bringing awareness to serious issues, and getting people to bring up these issues at budget time in Congress.
Unfortunately, such encryption isn't ubiquitous enough to protect non-critical transmissions: normal business and personal email, cell phone conversation, and many IM-style tools aren't remotely secure enough. Even for common encryption techniques such as SSL, the keys are rarely well-protected enough to prevent "law enforcement" from stealing them and monitring traffic secretly.
There are technologically reasonable techniques, such as the so-called "Trusted Computing" tools that are growing in use, but notice that those require signed, registered keys from companies like Microsoft. That's no protection whatsoever against abuse by federal agencies with access to the fundamental key signature authorities. It's a real political and technical problem: it's very difficult to get approval for general export for any tools that include robust encryption without gaping security holes for "approved" federal access.
Any fool can write a basic tool or two. But making it robust, making it reliable, and making the interface simple to use takes serious work. The result of writing really good tools that have terrible interfaces is well-described by Eric Raymond in his famous rant on CUPS: if the interface is not good, people will not be able to use it reliably.
Just use graduate students: they never have counted as life forms.
Have you factored in the risks and losses of Peachtree? I found it to be very unstable, and vulnerable to records being overwritten or destroyed by two people accessing it at the same time. I finally had to urge the accountant I was working with to install it on and only run it on one desktop, and arrange for regular backups which Peachtree didn't support very well at all.
Mr. Stallman is not the FSF: he is its founder. This means that the organizatin may have to acknwledge and deal with nasty situations in ways that he really wishes were not necessary, because they don't have the tools to do it a better way. This includes making very clear what the GPL license applies to, in order that it not be declared invalid because it overreaches, and dealing with distressing problems like the insistence f NVidia on keeping their drivers closed source, but seeing a big demand in the free sftware world for NVidia drivers.
X doesn't do cross platform well. And in particular, VNC tends to work better over slow connections than X. Also, VNC supports a purely Java based client, available from any Java capable web browser, so it doesn't require client installatioon.
Develpment workstations also run CentOS. You get binary compatibility with RHEL, but you can save your license fees for the production servers and not waste it on development environments. It looks like this poster had their workgroup changed as a committee decision where some really unfortunate issues got overlooked, like the need to pay for closed source licenses for Qt, C#, and CygWin to sell their own product as closed source.
This original poster is probably trying to build a user interface for a portable device where they don't want to publish their source code. That means they need to use libraries that are not under GPL. Ergo, they need Qt or a similar license, and that's why the Qt license costs so much rather than using open source and publishing open source.
This person wants to reap the benefits of open source development, without opening their own code up for similar development. This kind of thing is *exactly* why the GPL exists: to keep closed source developers like this from harvesting the cream of open source and then locking it up for their private products.
It's also apples and oranges. RedHat server versions are comparable to Windows 2003, plus MS Office with the built-in OpenOffice, plus SNMP services, plus a print server, plus an industrial grade mail and webmail and IMAP server, plus a backup server with the built-in Amanda system, plus an industrial grade file server, plus industrial grade firewalls and security tools, plus good CD and DVD ripping sftware, and you don't have to buy additional client licenses if you have more than 25 clients. Moreover, almost all of RedHat server tools are available free as part of the CentOS distribution, if you don't want the commercial level of support and would rather use the tools free. I've actually used CentOS to demonstrate RedHat tools, before urging a client to go ahead and buy a licensed RedHat system in order to get that commercial grade of support after my contract is ver.
No, where the original poster got messed up was in trying to run commerical style software on an open source system. QT licensing is the same cost as several weeks of on-site consulting time, and should frankly be replaced in most projects with a simple web interface for portability. The Embedded Linux software is very customized, and very project specific: that's why it costs so much. Rewriting everything in C# is one of the stupidest things I've ever heard of: it's a Microsoft written language, it's *designed* to frce you to use proprietary tools, and it's a Java wannabe. If these folks had been clueful, they'd have rewritten in plain C for speed and portability, or C++ for object oriented code, or actual Java for the write-once, run-anywhere advantages. But writing anything in C# that can be avoided. And the only reason to use Qt, and its licensing, is to allow you to keep your source code unavailable.
The CygWin license cost quoted is misleading as well. The only license for CygWin that costs that much is if you want to use CygWin to publish binaries without providing source code. The licensing is described at http://www.cygwin.com/licensing.html.
This original poster was basically trying to build a closed source environment, and paying the premium to insist on keeping their environment closed. Of course that will be expensive! It's not taking advantage of the open nature of the systems at all.
Keep an eye on the BIOS. The develpers for Trusted Computing are very eager to support Trusted Computing based control of the basic haardware, to prevent the use of plug-in bootable media on secured systems. They're also eager to have the BIOS, itself, be Trusted Computing signed althugh this is much more difficult in the short term. And they're eager to have CD and DVD drives tied into it for media authentication, and have less influence but still are pursuing doing so for disk drives, to reduce access to stolen disks on secure systems. This is a long-term scheme, not a short term one: simply because the BIOS's you've worked with don't have such features doesn't mean that the next generation of drives and BIOS's won't use that.
Let me reiterate: trusted Computingn can manage keys for hardware authentication. Such abilities are key to its use for security, which is the pretext for its development, as well as for the DRM which is its real purpose. And such key management is clearly designed to authenticate bootable devices, even if such tools have not yet been implemented.
There are benefits of Trusted Computing: but those benefits are overwhelmed by really unfrtunate design choices aimed at securing vendor desires to control the use of their software and hardware, not those of consumers.
You can tie it to a host. This is one of the major purposes of Trusted Computing. It's use for encryption and reliable authentication is really secondary, its clear goal is DRM.
Whoah indeed. There are plenty of reasons not involving Microsoft for manufacturers to want DRM, and they've done it with non-Microsoft related technologies. I don't mean to cast it all on Microsoft.
But the nature of the Trusted Computing reflects a very clear plan to use it for DRM: Microsft wants it both for protecting their own software from being used in pirate copies, and to mate Windows software to other Windows software that blocks competing products from accessing them. This is particularly true for streaming media: by locking multimedia files and media to specifically, "Trusted Computing" software that has to be authorized by the mother ship, they gain the ability to prevent Putting "Trusted Computing" software on a Linux box does no good for this. Trusted Computing's approach is much more limiting than a standard DVD player: the media can be registered to a specific copy of the media, with a specific software viewer, on a specific host, andn simply refuse to allow viewing it under any software not specifically supported by the media vendor. It may be possible to do screen capture of the output, but that's about it.
Such capabilities can also be integrated, and are clearly to be integrated, in hardware. This seems reasonable for VPN identification keys or biometrics. But it's obviusly planned as well for CD and DVD burners, to force the duplicator to use authorized software to read the files, and to permit only mothership authorized software to write with that specific burner.
It's also aimed squarely at the boot system: a Trusted Computing enabled chipset on the motherboard can prevent the booting of any kernel, or bootable device, without an appropriate vendor signature. You'd better believe Microsoft wants this to prevent "Trusted Computing" hardware from booting non-authorized operating systems. It's understandable for security reasons in a corporate environment, but it's deadly for home users or experimenters who may want to install operating systems without the capital to buy and manage Trusted Computing keys.
Patents on software are like copyrights on individual words: they actively hinder development of new works, and add a serious burden to new software projects to guess whether or not some obscure patent will be held to be infringed by the new work.
Oh, yes. The changes when the US eliminated the need to register copyrighted documents were quite sufficient. And given the amount and ease of creating new documents in this digital world, 20 years seems quite sufficient.
Oh, agreed. That legislation in the US was aimed squarely at protecting the Mickey Mouse copyright: Disney is a huge sponsr of such legislation, as that darned mouse gets older and older and Disney's library of old movies grows in value. Librarians I've spoken with are amazed at the difficulties current US copyright law creates for them.
But it was a good question: the companies are the same companies as elsewhere, both US and multi-national. Sony, Disney, Samsung, and other companies interested in DRM protecting their digital investment help prevent what folks in the US call "fair rights" use.
These desires for DRM are supported heavily by Microsoft, which wants to protect its software licenses and prevent other software from writing or even reading the private formats they use, especially for "personally encrypted" files. The whole field of DRM is about to get much worse with the "Trusted Computing" software program, led by Microsoft, which locks software and media to a specific motherboard's encryption and authentication chip, and which will als be built into the next generation of AMD and Intel CPU's. These tools are aimed squarely at preventing duplication of digital media, and create a real burden for making backup or archival copies of these media. They also specifically restrict the authorized software tools to play these media, making "fair use" sampling or even using other non-specifically licensed players very difficult.
The libraries of the world are quite right to be concerned about this.
You had a bad secretary. A good one will automatically juggle your meetings, your costs, and the pain to your family of being gone on your wife's birthday. And that secretary can do that for 20 people for the same conference and get you a discount rate that helps justify his salary, along with knowing where to get decent local coffee and book a hotel room for a visiting guest on Thanksgiving weekend. Such people are priceless, just as a really competent IT person who can make the printers work right for Powerpoint slides in a Russian edition of Windows is also priceless.
Having all the engineers do 5% of their time managing is a disaster in the offing: the one of them who enjoys office politics, and is good at gathering resources, will wind up as CEO. If they have other qualities that are really bad, they will still succeed in becoming CEO, and destroy the company. A few engineers that are also good managers can make a company wildly successful, but these are rare.
They've tried for years to prevent it, and keep trying. The regulations against exporting encryption have a chilling effect, and various clauses of the legislation that included the Telecommunications Privacy Act were amazing in what they tried to forbid.
Oh, my. Youngsters.
The HitchHiker'suote is a piece of common protester slang from yhe 1960's or earlier. It was usually spouted by urban college graduates spouting the evils of capitalism and shouting the glories of socialism and the coming Marxist revolution while mommy and daddy paid for their food, tuition bills, and the gas for their VW microbus.
Douglas Adams was old enough to remember those days, and to mock them as needed.