The short advice to give you is most definately discuss the issue of search when entering buildings with a lawyer. The best group to contact in this case is the ACLU -- they are best equiped to answer these questions and take action if you so desire.
The long of your question is that the NIH is a special case employer since they are the federal government. The Bill of Rights does not apply to private individuals and organizations, but it does (obviously) apply to the federal government and its agents. This precendent was recently verified in the von Bulow(sp) case. Now, government agency operate in a merky space with which the ACLU has a great wealth of experience. They have lititgated a number of cases on this subject.
These types of searches are completely legal at a private company provided that the right was enumerated in a policy document or employment contract. There is not a private company that I have ever encountered that didn't give itself the right search anything and everything they wanted. Think of it this way, if they listen to your phone calls, read your email, and search your desk, they can search your person so long at its on their property. Also, bear in mind, that it is perfectly legal to sign away your Constitutional rights in a relationship through a contract -- a perfect case in point is private elementary and high schools.
The main reason I am jonesing for a color PDA is the better visibility. I find the glare on a black and white screen extremely intense -- particularly in daylight and conference rooms filled with horrible florscent light. In both of these conditions, my Palm Vx is almost unusable without contorting myself into a pretzel. Color screens are far more resistant to these problems, although not immune. Also, while the backlighting on the Palm is decent, color is still best in the dark.
I don't know if you use a handheld, but the last thing I want to have to do with mine is "graffitti" in commands via a CLI. Yes, Linux/BSD work weel as servers and for some desktop applications, but how does that make it suitable for a PDA? As an analog, consider the crappiness of the Windows CE/Pocket PC interface which uses the same metaphor as Microsoft's desktop and server(I use the term loosely) OS's. So, should we adopt the X11(whatever interface standard there is) for a PDA and commit the same comedy of errors of Microsoft?
So, on what do you basis such statements that the ultimate of PDA would be bash and X11?
Actually, you are incorrect. Both Sony and Nintendo have incorporated in the United States. Therefore, they are free to give as much money as they want to Georgy and company as long as the money was earned in the US and funneled in from Japan.
XMI is part of the OMG's Unified Modeling Language Specification 1.3, and it stands for XML Model Interchange. It is intended as a mechanism to reliably transport UML models between tools.
The OMG metadata standard is the Meta Object Factory version 1.3. It allows the specification of the meta-meta data and provides IDL interfaces for accessing a repository based on the MOF definition.
2) It is NOT being developed by Apple or AOL. These are a bunch of people who used to work for Apple and AOL, but neither company is itself directly involved.
... snip...
4) Once again, this is NOT Apple doing this. But I wouldn't be surprised if some of the people from the now-defunct Human Interface team there are now working on it.
I believe the article said former employess of Apple and AOL which implies that Apple and AOL are probably not involved. I fail to see where the misunderstandings are in these point.
Speaking of creating misunderstandings, the speculation in point four could very well lead to far worse the misunderstandings than those potentially laid out in the original ariticle. Who is to say that any of these people has ever designed a UI? I didn't see a list of resumes on the web page reporting the projects to which Apple/AOL projects they have worked.
Yes, Netscape 4 is a horrible beast, but the design for Gecko blows anything Konq right out of the water. I will not even venture down the path of discussing the fallicy the Mozilla windows-centric other than to point out that the Linux/GTK port is considered the benchmark for the team, and to point out that the whole interface is composed using a deivation of XML called XUL. How much platform indenpendence do you want?
As for the issue with the readiness of Gecko, I want to know from these guys why they didn't devote their energies into Mozilla rather than rolling their rendering engine? They could have easily concentrated on Gecko, and not have worked on the other browser pieces. Not only KDE have a far superior HTML renderer(that includes full support for XML for those app developers), but the open source community and movement would have benefited from a more stable and product product earlier. Let us not forget that Mozilla is begging for help, and here are bunch of guys re-inventing the wheel.
So, I want to know from the KDE crew, why didn't you contribute to Gecko and embed your product into KDE? Was it pure ego? or Was it something substanitive? I must say that a perspective user, it seems you wasted a lot of time that could have been better spent helping another project and expanding the functionality of your own product.
It really depresses me to see the KDE team devoting so much time developing an HTML renderer when Mozilla's Gecko engine is a far better product and is in dire need of help. Imagine if their talents had been devoted to the greater good of Mozilla. Not only would KDE have a far more mature rendering engine(read support for HTML 4.0, CSS1 & 2, XML), but they could have contributed a great deal of the stability to Mozilla. I don't buy into any integration issues. According to Mosfet's own web site, KParts make the integration issue trival(note his integration of ghostscript and dvi). It also begs the question -- how much time could have been devoted to improving the desktop environment of KDE? Last I checked, we are pissed at Microsoft for bloating their desktop with a browser. These guys needs to stop including the kitchen sink and work on the nut and bolts -- a good desktop environment.
Just to address the issue of competition. Competition is unnecessary in this situation. Obstensively, the KDE crew wants open source to succeed. Therefore, they should embrace the Mozilla project. It's success would drive their success and vica versa. But then, no one will ever accuse the KDE crew of being good, open source community members. Heaven forbid they help another project when it would also benefit them.
It seems the Mosfet's ego was placed ahead of a good product management decision which points to larger general issue in open source projects. Egos seems to hamper projects, and in some cases, kill them. Established software engineering practice states that a project should maximize reuse. A common theme of projects that "roll their own" solutions is that the "reusable" piece wasn't quite good enough. The GNOME team's complaints about imlib are perfect example of this phenomenon. A GNOME user, I am not exactly thrilled that valuable time will be wsated developing a bew imaging library instead of either working with current imlib or with Raster to correct the perceived flaws. The bottom line in my mind is that these egos are impeding the progress of these projects, and clearly points out that open source is not perfect as some would portend.
With CORBA 3.0, there will be a standard for components called CORBABeans. From what I have seen of the standard, it look extremely promising. It is platform neutral. Supports local and distributed operation. It is a mimic of the JavaBeans standard which is really very cool except for the poor performance...
It could be a contender for use in such a repository, as well as, GNOME and KDE.
CS degree is not about teaching you the latest and greatest technology. Tehre are plenty of companies which are more than willing to train you in the use of their technologies.
A CS degree is about teaching you the principles underlying a field. As such, it is intended to be very abstract. It is aimed at teaching you the logical system that underlies all computer science disciplines. There are many lessons that can only be learned in a controlled, academic setting. Therefore, the degree is not as concerned with whether the program is written in C++ or Java, but how a computer is language is constructed and how it meets its goals. The degree gives you the tools to quickly assimilate new technologies which means that going through the full 4 year ciriculum at an ACM accredited institution will make you a far better developer than simply learning each technology in an ad hoc manner.
I work as a software engineer at a company where most of the developers have not had the four year CS degree. The lack of this foundation leads to many headaches and poor designs. While the 4 year degree isn't as sexy as one would hope, it will make a far better programmer and open your eyes to things that you would not otherwise see. Finally, the upper echelon jobs in this field require at least a BA/BS in CS if not a MA or PhD, and that is where the umber cool stuff is occurring(i.e. XEROX PARC, ILM, Watson, etc.)
I read an interview you gave the Philly Inquirer in which you spoke about your use and collection of guns. What are your thoughts regarding free software vis-a-vis the 2nd amendment? vis-a-vis the Bill of Rights in general? Do you feel ideas have influenced in the Open Source movement? If so, how?
I don't mean to stir up a gun debate, but rather, derive a possible source of the Open Source philosphy.
With all due respect to Rob, SlashDot is a relatively simple application from a transactionaly perspective. It just happens to be incredibly useful. When you consider Slashdot, it does blind inserts and updates without any regard to success or failure.
The question must be viewed with the needs of through the lense of ASP/ISP's aiming to provide web hosting services. In order to make money, ASP/ISP's must be able to host eCommerce applications. Therefore, he should direct his efforts toward putting the infrastructure in place for such a transition. Mind you, such a transition is major management and technology shift wich will probably take anywhere from 12 to 18 months to draft a strategy(i.e. select hardware and software platforms), acquire the assests, and train/hire the needed human resources.
To my mind, mySQL would be a poor choice in this scenario. As far as PHP goes, it could be uses in conjunction with Java, Perl, C, etc. in the front end, but it is NOT an application server. A WebLogic, Zope, WebSphere, Gemstone, etc. should sit between the web server and te database.
MySQL is only good for simple database appilcations. It is too simple to handle complex business applications including a competent order entry system(I work on a order entry system used by Fortune 500 companies...). For these applications, Oracle, Informix, SQl Server(ick!), or DB2 are a must.
As for Cold Fusion, it is not a a scripting language, but rather, an application server. In it's most advanced iterations, it supports clustering, application partitioning, and the like. Again, if you are writing a complex, high volume business application, these are mininium features.
Try writing a order entry system for a large scale manufacturer that must implement commodity based pricing models, promotion management, real-time inventory management, and freight management -- you will discover that PHP3/mySQL will fail miserably. In fact, I would be willing to bet that we couldn't implement our software in such an environment.
So, to sum it up, your suggestion is great for little web apps. But, if your market is going to include apaps for the big boys(like this fellow obviously needs), he should strongly consider tools such an application server such as Zope, WebLogic, or SilverStream. Also, go with a Solaris or Digital box running Oracle or an RS/6000 running AIX and DB2 -- SQL Server is a contentious hunk of crap. Finally, you should trash IIS,and get Apache up on the same platform as you choose for the RDBMS.
As an aside, I don't mean to dis Linux as the platform of choice, but Linux is not ready to do DB serving for the enterprise. SMP isn't quite there yet, and the products are just too young. For example, Oracle running on Solaris is 10+ years old and solid as a rock. For this type of application, robustness and stability are the most important considerations. In the busines world, no cares if the hardware/software combo costs $500,000 if it is available 24/7/365. But, if it is $10,000 and crashs frequently, business is lost and therefore, money is lost.
GNUStep is a framework that rides atop the environment in which it sits. In order for GNUStep to become GNOME aware, a GTK/GNOME backend must be written. Once such a backend is created, any GNUStep appilcation running atop that backend would look, act, and feel like a GNOME app. The same could be done for Qt/KDE...
The GNUStep folks are looking for folks to volunteer to help in that effort.
Closed drivers are buggy and they always will be without the open source development model.
How many ways can I NO!! The quality of drivers would only increase if a company had to maintain one or, possibly, two code bases for their drivers. It's a simply rule of software engineering -- less code = higher quality. Also, to echo the sentiments of an earlier poster -- open source != high quality. Yes, it can be said that high quality products have been open source. It can not be said that all open source projects are high quality.
Binary or not, I would prefer drivers written by the company who designed the hardware. Even with open specs, the company who made the hardware knows it better than an outsider. They now about the little quirks in the silicon, and they understand where and how little optimization tricks can be had. It is not to say these insights can not be dervied from open specs, but why waste the time and resources.
Personally, I think all drivers for recent/ mainstream hardware should be written by the manufacturers. Why waste the valuable resources of the open-source community to support a company. If they want to make money on Linux, let them write the drivers and support it. To my mind, we have been a little too nice to these free loaders. Let them write the drivers and leave us to improve to the software. We should only be writing drivers for legacy or unsupported hardware.
Finally, I think the notion of a loadable driver is good thing(tm) -- even if it is implemented poorly in Windows. How can one expect to pentrate the desktop market if a user has to understand how to build a kernel when they get new hardware? While it is pretty easy for me as a developer, it is not for my mom(who is very good with computers). Plus, it takes too damn long. The Linux community is certainly cool enough and smart enough to make Linux hardware installation as easy as the Mac and as stable and powerful as Linux.
Microsoft could easily include XFS lock, stock, and barrel without a fear of releasing the NT source because NT utilizies installable file systems(IFS). Therefore, the filesystem code is external to the kernel source code. As such, all they would have to do is release the source to the IFS.
Just as an important note, only two RDBMS's support truly parallel executiion og queries -- Oracle and DB2. This feature is called MPP in Oracle and costs mucho buckos. MPP Oracle is one of the fastest(if not the), most scabale databases availible. It would be interesting to see if Oracle will port this feature to Beowulf.
Threads are merely a design issue. mySQL decided to execute mutliple queries over threads rather than processes. Oracle, Informix, DB2, and PostgresSQL use multiple processes in UNIX environments. This largely due to their age. When they were first written UNIX threads either didn't exist or weren't standard enough to work reliabily across platforms. IMHO, threads are better because they can be started and closed fater and require fewer resources, but then, processes are cool as well.
FYI, the term microkernel does not neccessarily mean that a "small" kernel. It means that the kernel provides a base set of services that operating systems built atop it can suppliment. Therefore, a microkernel is meant to a reusable core for many operating systems.
For example, the Mach microkernel is not an operating system in and of itself. It provides services such as network and disk access. Operating systems such as BSD, Linux, and NeXT were built to utilize its services. There is one thing one can not say about the Mach microkernel -- it is small.
Neither PetrOS nor NT are microkernels, although, they may utilize a microkernel-like architecture by creating Ring O level services. In this case, the folks at Trumphet have usurped a computer science concept for marketing purposes.
Actually, the first company to instroduce an idea like this one was called Lonewolf Systems. They had a chip that could go in any applicance(primary audio.video gear, but also other appliances). The chip included a fibre port. Nasically, every appliance was controlled by a central compuer on the network. It allowed a VCR or LD player in the living room to play on a TV in the bedroom because the central computer was actually routing the siganls to appropriate sources. It was a mondo cool system, but prohibitvely expensive in the early 90's as it REQUIRED a fiber optic network...
I wonder if the MS proposal will do the same kinda thing? I highly doubt it, but who knows...
Slashdot Mirror
a Beowulf cluster of these things?
...)
(Since it is Slashdot, It had to be said
The short advice to give you is most definately discuss the issue of search when entering buildings with a lawyer. The best group to contact in this case is the ACLU -- they are best equiped to answer these questions and take action if you so desire.
The long of your question is that the NIH is a special case employer since they are the federal government. The Bill of Rights does not apply to private individuals and organizations, but it does (obviously) apply to the federal government and its agents. This precendent was recently verified in the von Bulow(sp) case. Now, government agency operate in a merky space with which the ACLU has a great wealth of experience. They have lititgated a number of cases on this subject.
These types of searches are completely legal at a private company provided that the right was enumerated in a policy document or employment contract. There is not a private company that I have ever encountered that didn't give itself the right search anything and everything they wanted. Think of it this way, if they listen to your phone calls, read your email, and search your desk, they can search your person so long at its on their property. Also, bear in mind, that it is perfectly legal to sign away your Constitutional rights in a relationship through a contract -- a perfect case in point is private elementary and high schools.
I hope that helps. Good luck.
The main reason I am jonesing for a color PDA is the better visibility. I find the glare on a black and white screen extremely intense -- particularly in daylight and conference rooms filled with horrible florscent light. In both of these conditions, my Palm Vx is almost unusable without contorting myself into a pretzel. Color screens are far more resistant to these problems, although not immune. Also, while the backlighting on the Palm is decent, color is still best in the dark.
I don't know if you use a handheld, but the last thing I want to have to do with mine is "graffitti" in commands via a CLI. Yes, Linux/BSD work weel as servers and for some desktop applications, but how does that make it suitable for a PDA? As an analog, consider the crappiness of the Windows CE/Pocket PC interface which uses the same metaphor as Microsoft's desktop and server(I use the term loosely) OS's. So, should we adopt the X11(whatever interface standard there is) for a PDA and commit the same comedy of errors of Microsoft?
So, on what do you basis such statements that the ultimate of PDA would be bash and X11?
Actually, you are incorrect. Both Sony and Nintendo have incorporated in the United States. Therefore, they are free to give as much money as they want to Georgy and company as long as the money was earned in the US and funneled in from Japan.
XMI is part of the OMG's Unified Modeling Language Specification 1.3, and it stands for XML Model Interchange. It is intended as a mechanism to reliably transport UML models between tools.
The OMG metadata standard is the Meta Object Factory version 1.3. It allows the specification of the meta-meta data and provides IDL interfaces for accessing a repository based on the MOF definition.
2) It is NOT being developed by Apple or AOL. These are a bunch of people who used to work for Apple and AOL, but neither company is itself directly involved.
4) Once again, this is NOT Apple doing this. But I wouldn't be surprised if some of the people from the now-defunct Human Interface team there are now working on it.
I believe the article said former employess of Apple and AOL which implies that Apple and AOL are probably not involved. I fail to see where the misunderstandings are in these point.
Speaking of creating misunderstandings, the speculation in point four could very well lead to far worse the misunderstandings than those potentially laid out in the original ariticle. Who is to say that any of these people has ever designed a UI? I didn't see a list of resumes on the web page reporting the projects to which Apple/AOL projects they have worked.
Truly creative piece of satire -- it expresses the impotence of the FBI matter such as these.
I can say one thing, though, this man has just earned himself an FBI file and some good, ole fashioned Echelon monitoring.
Yes, Netscape 4 is a horrible beast, but the design for Gecko blows anything Konq right out of the water. I will not even venture down the path of discussing the fallicy the Mozilla windows-centric other than to point out that the Linux/GTK port is considered the benchmark for the team, and to point out that the whole interface is composed using a deivation of XML called XUL. How much platform indenpendence do you want?
As for the issue with the readiness of Gecko, I want to know from these guys why they didn't devote their energies into Mozilla rather than rolling their rendering engine? They could have easily concentrated on Gecko, and not have worked on the other browser pieces. Not only KDE have a far superior HTML renderer(that includes full support for XML for those app developers), but the open source community and movement would have benefited from a more stable and product product earlier. Let us not forget that Mozilla is begging for help, and here are bunch of guys re-inventing the wheel.
So, I want to know from the KDE crew, why didn't you contribute to Gecko and embed your product into KDE? Was it pure ego? or Was it something substanitive? I must say that a perspective user, it seems you wasted a lot of time that could have been better spent helping another project and expanding the functionality of your own product.
It really depresses me to see the KDE team devoting so much time developing an HTML renderer when Mozilla's Gecko engine is a far better product and is in dire need of help. Imagine if their talents had been devoted to the greater good of Mozilla. Not only would KDE have a far more mature rendering engine(read support for HTML 4.0, CSS1 & 2, XML), but they could have contributed a great deal of the stability to Mozilla. I don't buy into any integration issues. According to Mosfet's own web site, KParts make the integration issue trival(note his integration of ghostscript and dvi). It also begs the question -- how much time could have been devoted to improving the desktop environment of KDE? Last I checked, we are pissed at Microsoft for bloating their desktop with a browser. These guys needs to stop including the kitchen sink and work on the nut and bolts -- a good desktop environment.
Just to address the issue of competition. Competition is unnecessary in this situation. Obstensively, the KDE crew wants open source to succeed. Therefore, they should embrace the Mozilla project. It's success would drive their success and vica versa. But then, no one will ever accuse the KDE crew of being good, open source community members. Heaven forbid they help another project when it would also benefit them.
It seems the Mosfet's ego was placed ahead of a good product management decision which points to larger general issue in open source projects. Egos seems to hamper projects, and in some cases, kill them. Established software engineering practice states that a project should maximize reuse. A common theme of projects that "roll their own" solutions is that the "reusable" piece wasn't quite good enough. The GNOME team's complaints about imlib are perfect example of this phenomenon. A GNOME user, I am not exactly thrilled that valuable time will be wsated developing a bew imaging library instead of either working with current imlib or with Raster to correct the perceived flaws. The bottom line in my mind is that these egos are impeding the progress of these projects, and clearly points out that open source is not perfect as some would portend.
With CORBA 3.0, there will be a standard for components called CORBABeans. From what I have seen of the standard, it look extremely promising. ...
It is platform neutral. Supports local and distributed operation. It is a mimic of the JavaBeans standard which is really very cool except for the poor performance
It could be a contender for use in such a repository, as well as, GNOME and KDE.
CS degree is not about teaching you the latest and greatest technology. Tehre are plenty of companies which are more than willing to train you in the use of their technologies.
A CS degree is about teaching you the principles underlying a field. As such, it is intended to be very abstract. It is aimed at teaching you the logical system that underlies all computer science disciplines. There are many lessons that can only be learned in a controlled, academic setting. Therefore, the degree is not as concerned with whether the program is written in C++ or Java, but how a computer is language is constructed and how it meets its goals. The degree gives you the tools to quickly assimilate new technologies which means that going through the full 4 year ciriculum at an ACM accredited institution will make you a far better developer than simply learning each technology in an ad hoc manner.
I work as a software engineer at a company where most of the developers have not had the four year CS degree. The lack of this foundation leads to many headaches and poor designs. While the 4 year degree isn't as sexy as one would hope, it will make a far better programmer and open your eyes to things that you would not otherwise see. Finally, the upper echelon jobs in this field require at least a BA/BS in CS if not a MA or PhD, and that is where the umber cool stuff is occurring(i.e. XEROX PARC, ILM, Watson, etc.)
I read an interview you gave the Philly Inquirer in which you spoke about your use and collection of guns. What are your thoughts regarding free software vis-a-vis the 2nd amendment? vis-a-vis the Bill of Rights in general? Do you feel ideas have influenced in the Open Source movement? If so, how?
I don't mean to stir up a gun debate, but rather, derive a possible source of the Open Source philosphy.
With all due respect to Rob, SlashDot is a relatively simple application from a transactionaly perspective. It just happens to be incredibly useful. When you consider Slashdot, it does blind inserts and updates without any regard to success or failure.
The question must be viewed with the needs of through the lense of ASP/ISP's aiming to provide web hosting services. In order to make money, ASP/ISP's must be able to host eCommerce applications. Therefore, he should direct his efforts toward putting the infrastructure in place for such a transition. Mind you, such a transition is major management and technology shift wich will probably take anywhere from 12 to 18 months to draft a strategy(i.e. select hardware and software platforms), acquire the assests, and train/hire the needed human resources.
To my mind, mySQL would be a poor choice in this scenario. As far as PHP goes, it could be uses in conjunction with Java, Perl, C, etc. in the front end, but it is NOT an application server. A WebLogic, Zope, WebSphere, Gemstone, etc. should sit between the web server and te database.
MySQL is only good for simple database appilcations. It is too simple to handle complex business applications including a competent order entry system(I work on a order entry system used by Fortune 500 companies...). For these applications, Oracle, Informix, SQl Server(ick!), or DB2 are a must.
As for Cold Fusion, it is not a a scripting language, but rather, an application server. In it's most advanced iterations, it supports clustering, application partitioning, and the like. Again, if you are writing a complex, high volume business application, these are mininium features.
Try writing a order entry system for a large scale manufacturer that must implement commodity based pricing models, promotion management, real-time inventory management, and freight management -- you will discover that PHP3/mySQL will fail miserably. In fact, I would be willing to bet that we couldn't implement our software in such an environment.
So, to sum it up, your suggestion is great for little web apps. But, if your market is going to include apaps for the big boys(like this fellow obviously needs), he should strongly consider tools such an application server such as Zope, WebLogic, or SilverStream. Also, go with a Solaris or Digital box running Oracle or an RS/6000 running AIX and DB2 -- SQL Server is a contentious hunk of crap. Finally, you should trash IIS,and get Apache up on the same platform as you choose for the RDBMS.
As an aside, I don't mean to dis Linux as the platform of choice, but Linux is not ready to do DB serving for the enterprise. SMP isn't quite there yet, and the products are just too young. For example, Oracle running on Solaris is 10+ years old and solid as a rock. For this type of application, robustness and stability are the most important considerations. In the busines world, no cares if the hardware/software combo costs $500,000 if it is available 24/7/365. But, if it is $10,000 and crashs frequently, business is lost and therefore, money is lost.
GNUStep is a framework that rides atop the environment in which it sits. In order for GNUStep to become GNOME aware, a GTK/GNOME backend must be written. Once such a backend is created, any GNUStep appilcation running atop that backend would look, act, and feel like a GNOME app. The same could be done for Qt/KDE ...
The GNUStep folks are looking for folks to volunteer to help in that effort.
How many ways can I NO!! The quality of drivers would only increase if a company had to maintain one or, possibly, two code bases for their drivers. It's a simply rule of software engineering -- less code = higher quality. Also, to echo the sentiments of an earlier poster -- open source != high quality. Yes, it can be said that high quality products have been open source. It can not be said that all open source projects are high quality.
Binary or not, I would prefer drivers written by the company who designed the hardware. Even with open specs, the company who made the hardware knows it better than an outsider. They now about the little quirks in the silicon, and they understand where and how little optimization tricks can be had. It is not to say these insights can not be dervied from open specs, but why waste the time and resources.
Personally, I think all drivers for recent/ mainstream hardware should be written by the manufacturers. Why waste the valuable resources of the open-source community to support a company. If they want to make money on Linux, let them write the drivers and support it. To my mind, we have been a little too nice to these free loaders. Let them write the drivers and leave us to improve to the software. We should only be writing drivers for legacy or unsupported hardware.
Finally, I think the notion of a loadable driver is good thing(tm) -- even if it is implemented poorly in Windows. How can one expect to pentrate the desktop market if a user has to understand how to build a kernel when they get new hardware? While it is pretty easy for me as a developer, it is not for my mom(who is very good with computers). Plus, it takes too damn long. The Linux community is certainly cool enough and smart enough to make Linux hardware installation as easy as the Mac and as stable and powerful as Linux.
-MouseMicrosoft could easily include XFS lock, stock, and barrel without a fear of releasing the NT source because NT utilizies installable file systems(IFS). Therefore, the filesystem code is external to the kernel source code. As such, all they would have to do is release the source to the IFS.
-Mouse
Just as an important note, only two RDBMS's support truly parallel executiion og queries -- Oracle and DB2. This feature is called MPP in Oracle and costs mucho buckos. MPP Oracle is one of the fastest(if not the), most scabale databases availible. It would be interesting to see if Oracle will port this feature to Beowulf.
Threads are merely a design issue. mySQL decided to execute mutliple queries over threads rather than processes. Oracle, Informix, DB2, and PostgresSQL use multiple processes in UNIX environments. This largely due to their age. When they were first written UNIX threads either didn't exist or weren't standard enough to work reliabily across platforms. IMHO, threads are better because they can be started and closed fater and require fewer resources, but then, processes are cool as well.
FYI, the term microkernel does not neccessarily mean that a "small" kernel. It means that the kernel provides a base set of services that operating systems built atop it can suppliment. Therefore, a microkernel is meant to a reusable core for many operating systems.
For example, the Mach microkernel is not an operating system in and of itself. It provides services such as network and disk access. Operating systems such as BSD, Linux, and NeXT were built to utilize its services. There is one thing one can not say about the Mach microkernel -- it is small.
Neither PetrOS nor NT are microkernels, although, they may utilize a microkernel-like architecture by creating Ring O level services. In this case, the folks at Trumphet have usurped a computer science concept for marketing purposes.
Actually, the first company to instroduce an idea like this one was called Lonewolf Systems. They had a chip that could go in any applicance(primary audio.video gear, but also other appliances). The chip included a fibre port. Nasically, every appliance was controlled by a central compuer on the network. It allowed a VCR or LD player in the living room to play on a TV in the bedroom because the central computer was actually routing the siganls to appropriate sources. It was a mondo cool system, but prohibitvely expensive in the early 90's as it REQUIRED a fiber optic network ...
...
I wonder if the MS proposal will do the same kinda thing? I highly doubt it, but who knows