>
Otherwise the division should be spun off as a research company, purely focussed on IP. IBM is a good example in point.
IBM may be a good example, but it is a bad model ethically and pragmatically example. Its patents hoarding heightens the barrier to entry, because basically to do real business in IT now you have to have enough fundamental patents to bargain on IBM out of paying them huge, confiscatory royalties on everything under the Sun.
Actually, according to Cringley's standards, IBM patents are part of the problem, not of the solution.
>
MySQL (commercial license) will be shipped as standard with NetWare according to this announcement. I consider it a follow-up to the Slashdot story about the PostgreSQL port for NetWare.
Let's see if I got it right. MySQL gets money to deliver something far inferior on features, programming integrity and scalability to PostgreSQL which was free.
Incidentally, that shows one of the copyleft virtues: using dual licensing with GNU GPL, MySQL gets more resources even being far inferior to its non-copyleft counterpart.
The most important question, besides if the MS Word XML format will be well-documented enough, is if it will be the default saving format. Most MS Office users simply don't care enough to save MS Word documents in RTF, for example, even if it's more than good enough for the vast majority of the documents.
Not the main issue on the article, but it is unfair to single someone as the inventor of XML, which is just a streamlined version of SGML which is an evolution from IBM's GML.
I worked for three years at the biggest telephony billing system vendor, Amdocs, the Israeli company that support Ensemble, the billing system that operates most of US POTS calls and most of the European GSM ones, besides being the dominant player at cell phones in places as South Korea and elsewhere.
Amdocs started as a pretty much technical savy company, with a Unix system to do yellow pages layout. Delivering this system leveraged them to do the billing systems for the Baby Bells, when they got mainframe and data processing knowledge. It was GNU friendly, using most of the GNU toolchain such as gcc, GNU make, RCS and so on.
Supporting cell phone billing could have deepened open systems commitment, because these systems again run in Unix systems, mainly HP-UX. But the interface was done in Sybase PowerBuilder, as opposed to the POTS "Philishave" 3278 terminals.
After big money arrived, MBAs took power. Text processing was migrated from Unix WordPerfect to MS WinWord, and it is a pain to browse technical documentation in MS Word, I tell ya. Technically-savy people migrated to better jobs or higher up in the corporate foodchain, and today very little of the staff has even an IT background. Most don't even know the tools they use enough to be able to evaluate something else, and even the GNU tools in use are left to rot. People are still using GNU Emacs 19, RCS use was never upgraded to CVS or something else, and people are generally wary of the GNU commitment due to a lack of understanding.
The option of porting the PowerBuilder programs to run under POSIX and X Window System was never explored, instead "smart" clients are going Citrix MetaFrame, that at least gives them the option of using X Window System at the desktop, but I never heard of one that does.
The billing system back-end is always a commercial Unix, generally HP-UX. It is seriously misused, with a stupid multilingual setup that complicates administration and consumes resources as many times as there are languages supported. I've seen a billing system in a small GSM operator that needed to support interfaces in four different languages using up two SuperDomes, where a more intelligent system would have taken at most half such a system!
Needless to say, the predominance of HP as a hardware and OS vendor does not encourage alternatives to MS, as HP itself is deeply commited to a "MS everywhere, Unix where strictly necessary" policy. That's why they need and support Samba so much: they gave up on open systems desktops a long time ago.
There were unofficial talks of supporting GNU/Linux, but until I left there nothing came of it except a small intranet webcam server... when I setup a spare PC as squid cache to save on bandwidth it was no sooner discovered by supervision than took offline, even if half the office, including system analysts, programmer and project managers depended on it to read technical documentation and surf.
All telephony operators I visited while at Amdocs, in South America and Europe, were pretty much married to MS, using MS Exchange as email servers and that is it. They generally use MS all over the place, eventually migrating some web server or other non-critical system to Apache or HP-UX because of reliability; but when something delicate as MS Exchange breaks, the standard answer is to just throw more resources at it. Technical and historical discussion on open vs proprietary systems is generally discouraged as disruptive of chosen directions. The preferred platform for development is Java, but there is strong pressure to consider MS.Net instead.
Many managers are completely uneducated on both general culture and Informatics, so all this is quite unlikely to change soon, unless GNU systems make a big splash either on cost, support and scalability on the back end or on ease-of-use, resources consumption and interoperability on the desktop.
The bright side is that there is one or other small non-critical back end system being evaluated on GNU/Linux, usually SuSe, Mandrake or Red Hat. Few people even heard of free software or Debian, mostly it comes from some daring software vendor or a curious underling as myself.
Why oh why they don't just support a real SQL DBMS?
As always people coding to MySQL will have to code around all sorts of limitations that wouldn't be there with real SQL DBMSs as Firebird, SAPdb or PostgreSQL.
Their code will be more complex, less reliable, their databases will get all sorts of inconsistent data due to lack, or underutilization, of integrity constraints and transaction control.
And once they go over three concurrent users, performance will suffer... then they will have to recode to ANSI SQL, because MySQL isn't standards-compliant at all, and will never achieve the level of elegance that SQL provides -- and that is already less than a truly, but not yet widely available, fully RDBMS could give.
>
Replace copyright law with any other law (gun control, abortion, stealing, fraud, etc.) and the story will remain the same.
I think you misread it. Copyright law has the potential to block that most naturally free of things, communication, and therefore it will ultimately require a police state if it is not tamed. No other law restricts such a thing or would actually require so much enforcement.
>
The goal of the patents (according to the article) is to prevent the use of end to end control as a copyright enforment method
Hope you are right. As I read it, the goal of the aforementioned patents is to prevent installation of software in violation of its licensing. I hope they will cover some fundamental aspects of DRM, but I doubt it: MS will most likely long have covered all its DRM bases with patents of its own.
>
My husband and I chose to have a child when we were quite young - 18 and 20, respectively, and our daughter has Asperger's Syndrome.
Mutations are a matter of probability, and isn't advanced as the only genetical cause. So your daughter's problem at such age does not disprove the mutation possibility in general, nor the possibility that her problem might be due to a mutation.
That said, mutations are just a theoretical possibility, not an observed fact statistically correlated to the supposed consequence.
>
These "misspellings" of useful DNA aren't bad or good, but simply different. The differences could give your child a genetic advantage just as easily as a genetic disadvantage.
A little knowledge is indeed the most dangerous of things...
This is absolutely false. The overwhelming majority of mutations are malefic. Only a very few DNA copy deviations will create anything at least on par with the original, something like one in millions. The chance at an actual improvement is theoretical only, none ever observed.
BTW, this is just one reason of miriad reasons why some people, religiously informed or not, still doubt that the Theory of Biological Evolution will ever be proved a fact. The main reason being that in speculative Philosophy it has existed for millenia before Darwin applied it to Biology: see CS Lewis' The Funeral of a Great Myth.
Even if he successfully prevents MS from enforcing only licensed software on its OSs, it still does not addresses the issue raised by RMS in The Right to Read, namely that copyright enforcement thru technology can turn all the World in a global police state in copyright owners' benefit.
Agreed. But that is because many of the protocols, APIs, file formats and hardware specifications are secret. The more we use GNU/Linux, more of this information get available and is put into good use.
Additionally, MS Windows can't keep your privacy, protect you from viruses, save you on hardware and software costs, give you the information and freedom GNU/Linux does.
Not quite. Government agencies, departments and state-owned companies do, but generally speaking they can't share your data with other branches of government. So while these may be seen as orthogonal issues, the digital certificates may be seen as a PR spin to convince citizens to allow the government to centralize all this data.
Now centralising all data *is* dangerous, prone to abuse and data security issues. But no one can protect citizens from their own stupidity if we allow government the leeway it naturally wants, because it is a human organization and all big enough human organizations usually almost have a mind of their own bent on ever growing and extending their power and influence.
>
It would be 5000% more effective for his post to lkml to say "I've decided to support the (subversion|arch|aegis|pcrs|whatever) project as an official GNU project
He cannot do that unless all source code in the project is donated to the FSF. Obviously he can still profer his personal support.
>
i dont find linus's perspective to be completely contradictory.
It is not that it is contradictory. It is just that he is advancing proprietary software, when it only was possible for him to go so far by building on free software. The point is that he would never have got where he is if other people had applied to kernels and OSs what he now applies to source code control systems.
>
Nor were there other freely available compilers, or anything.
BSD had its own compiler. Anyway, the point here is that Linux depended on GNU tools to exist, and now does not upheld the same liberties it has enjoyed. Granted it is not so bad as plain software hoarding that almost killed BSD, but still is annoying and contradictory.
Re:Server down for obvious reasons
on
Blender Is GPL
·
· Score: 2
>
If your code needs foreign key and check constraints
It is not code that requires integrity constraints, but data integrity. Data integrity is a function of the DBMS that is declared. That makes it much simpler, faster, more consistent than if one tries to do that with procedural coding, which ends up not covering up ad hoc use of the database anyway.
>
The kernel developers decided that BitKeeper was the best tool for the job so they used it
Interestingly enough, if everyone used the same arguments Linux, the kernel, would never had coming into being.
Just imagine that gcc and the other GNU utilities carried a similar license. Linus would have had to buy a real compiler and developer toolbox before even starting, and these cost real money in those days, much more than he could afford. And then no one would use his work or collaborate with him, because BSD and the other Unices were so much better for the task. The few tinkerers would have stayed with BSD.
Re:Server down for obvious reasons
on
Blender Is GPL
·
· Score: 2
>
if you require that much extra coding, then you might need to think about your database approach
Always when dealing with data, application code must check for constraints procedurally. This is extra, uncessarily complex code that is much better, simpler and more consistently done at the DBMS.
>
If you rely on things like constraint checks to keep your data straight, you're doing a pretty half assed job to begin with.
Ignore History at your own peril... the relational model for database management was exactly intended to provide declarative, centralised integrity constraints, because doing so in the application, no matter how good the application, is a sure recipe for failure. You don't cover integrity assurance for interactive, direct users of the database, and it is next to impossible to keep track of all integrity constraints and enforce them in all application programs.
Re:Server down for obvious reasons
on
Blender Is GPL
·
· Score: 2
>
guess you've never seen the/. code. The whole thing is convoluted.
That is my point. Having a real DBMS makes coding much, much simpler.
>
Oracle never releases patches or new versions tho because their code is so "tight" right?
Oracle is a bad example of quality. Try PostgreSQL instead.
>
mySQL practically admin's itself. Not everyone has a need for a useless 100K+/year DBA who has TOMES of oracle books to do something that *should* be a simple process.
Who's talking about Oracle? PostgreSQL is as easy to manage as MySQL, but requires less coding and makes much more complex tasks easier for both developer and administrator.
>
I would say transactions but MySQL has transactions now
Yes, in a totally unproven, and optional, implementation! While still lacking all kinds of declared integrity constraints, scalability and even sane documentation.
>
mySQL is good at the personal/small-medium business level.
Nothing that makes data integrity optional or procedural is good for holding any organisational data...
Re:Why oh why are you such a snob?
on
Blender Is GPL
·
· Score: 2
>
There are plenty of applications where a flat file is sufficient and a database of any sort is overkill
Certainly dynamic web sites aren't among them. They call for concurrent data access and updateability, which is a typical DBMS use. Also, a good DBMS will ease the complexity of coding and improve reliability over simple file access.
>
MySQL does have some advantages for web based applications that demand low connection latency - an area where MySQL excels over "real" databases that cost "real" bucks. Also MySQL has a very easy to use command line management client that facilitate remote admin.
A real DBMS -- a database is the organised data, not the software to control it -- does not necessarily cost anything, see PostgreSQL and others. And PostgreSQL does have all the advantages of MySQL, including speed and command line management, plus reliability and while requiring far simpler coding and keeping the data more consistent and accessible.
>
I just hope that MySQL or our file-systems can rise to the task.
In a way yes, but not ultimately. Because the people behind MySQL and filesystems do not really grok the task, which is ultimately a database one and thus should be handled under the relational model.
Slashdot Mirror
IBM may be a good example, but it is a bad model ethically and pragmatically example. Its patents hoarding heightens the barrier to entry, because basically to do real business in IT now you have to have enough fundamental patents to bargain on IBM out of paying them huge, confiscatory royalties on everything under the Sun.
Actually, according to Cringley's standards, IBM patents are part of the problem, not of the solution.
And remember, there is no such thing as IP!
According to Bill Parish, CPA, it isn't and hasn't been for quite some time now.
Let's see if I got it right. MySQL gets money to deliver something far inferior on features, programming integrity and scalability to PostgreSQL which was free.
Incidentally, that shows one of the copyleft virtues: using dual licensing with GNU GPL, MySQL gets more resources even being far inferior to its non-copyleft counterpart.
Too pricey for something that doesn't play Ogg, doesn't have Firewire connector.
The most important question, besides if the MS Word XML format will be well-documented enough, is if it will be the default saving format. Most MS Office users simply don't care enough to save MS Word documents in RTF, for example, even if it's more than good enough for the vast majority of the documents.
Not the main issue on the article, but it is unfair to single someone as the inventor of XML, which is just a streamlined version of SGML which is an evolution from IBM's GML.
I worked for three years at the biggest telephony billing system vendor, Amdocs, the Israeli company that support Ensemble, the billing system that operates most of US POTS calls and most of the European GSM ones, besides being the dominant player at cell phones in places as South Korea and elsewhere.
.Net instead.
Amdocs started as a pretty much technical savy company, with a Unix system to do yellow pages layout. Delivering this system leveraged them to do the billing systems for the Baby Bells, when they got mainframe and data processing knowledge. It was GNU friendly, using most of the GNU toolchain such as gcc, GNU make, RCS and so on.
Supporting cell phone billing could have deepened open systems commitment, because these systems again run in Unix systems, mainly HP-UX. But the interface was done in Sybase PowerBuilder, as opposed to the POTS "Philishave" 3278 terminals.
After big money arrived, MBAs took power. Text processing was migrated from Unix WordPerfect to MS WinWord, and it is a pain to browse technical documentation in MS Word, I tell ya. Technically-savy people migrated to better jobs or higher up in the corporate foodchain, and today very little of the staff has even an IT background. Most don't even know the tools they use enough to be able to evaluate something else, and even the GNU tools in use are left to rot. People are still using GNU Emacs 19, RCS use was never upgraded to CVS or something else, and people are generally wary of the GNU commitment due to a lack of understanding.
The option of porting the PowerBuilder programs to run under POSIX and X Window System was never explored, instead "smart" clients are going Citrix MetaFrame, that at least gives them the option of using X Window System at the desktop, but I never heard of one that does.
The billing system back-end is always a commercial Unix, generally HP-UX. It is seriously misused, with a stupid multilingual setup that complicates administration and consumes resources as many times as there are languages supported. I've seen a billing system in a small GSM operator that needed to support interfaces in four different languages using up two SuperDomes, where a more intelligent system would have taken at most half such a system!
Needless to say, the predominance of HP as a hardware and OS vendor does not encourage alternatives to MS, as HP itself is deeply commited to a "MS everywhere, Unix where strictly necessary" policy. That's why they need and support Samba so much: they gave up on open systems desktops a long time ago.
There were unofficial talks of supporting GNU/Linux, but until I left there nothing came of it except a small intranet webcam server... when I setup a spare PC as squid cache to save on bandwidth it was no sooner discovered by supervision than took offline, even if half the office, including system analysts, programmer and project managers depended on it to read technical documentation and surf.
All telephony operators I visited while at Amdocs, in South America and Europe, were pretty much married to MS, using MS Exchange as email servers and that is it. They generally use MS all over the place, eventually migrating some web server or other non-critical system to Apache or HP-UX because of reliability; but when something delicate as MS Exchange breaks, the standard answer is to just throw more resources at it. Technical and historical discussion on open vs proprietary systems is generally discouraged as disruptive of chosen directions. The preferred platform for development is Java, but there is strong pressure to consider MS
Many managers are completely uneducated on both general culture and Informatics, so all this is quite unlikely to change soon, unless GNU systems make a big splash either on cost, support and scalability on the back end or on ease-of-use, resources consumption and interoperability on the desktop.
The bright side is that there is one or other small non-critical back end system being evaluated on GNU/Linux, usually SuSe, Mandrake or Red Hat. Few people even heard of free software or Debian, mostly it comes from some daring software vendor or a curious underling as myself.
Why oh why they don't just support a real SQL DBMS?
As always people coding to MySQL will have to code around all sorts of limitations that wouldn't be there with real SQL DBMSs as Firebird, SAPdb or PostgreSQL.
Their code will be more complex, less reliable, their databases will get all sorts of inconsistent data due to lack, or underutilization, of integrity constraints and transaction control.
And once they go over three concurrent users, performance will suffer... then they will have to recode to ANSI SQL, because MySQL isn't standards-compliant at all, and will never achieve the level of elegance that SQL provides -- and that is already less than a truly, but not yet widely available, fully RDBMS could give.
I think you misread it. Copyright law has the potential to block that most naturally free of things, communication, and therefore it will ultimately require a police state if it is not tamed. No other law restricts such a thing or would actually require so much enforcement.
Moderators, parent up, please!
Hope you are right. As I read it, the goal of the aforementioned patents is to prevent installation of software in violation of its licensing. I hope they will cover some fundamental aspects of DRM, but I doubt it: MS will most likely long have covered all its DRM bases with patents of its own.
Assuming people will get a say. With so much PR-turned-Newspeak and Practical Materialism around, I very much doubt it.
Mutations are a matter of probability, and isn't advanced as the only genetical cause. So your daughter's problem at such age does not disprove the mutation possibility in general, nor the possibility that her problem might be due to a mutation.
That said, mutations are just a theoretical possibility, not an observed fact statistically correlated to the supposed consequence.
A little knowledge is indeed the most dangerous of things...
This is absolutely false. The overwhelming majority of mutations are malefic. Only a very few DNA copy deviations will create anything at least on par with the original, something like one in millions. The chance at an actual improvement is theoretical only, none ever observed.
BTW, this is just one reason of miriad reasons why some people, religiously informed or not, still doubt that the Theory of Biological Evolution will ever be proved a fact. The main reason being that in speculative Philosophy it has existed for millenia before Darwin applied it to Biology: see CS Lewis' The Funeral of a Great Myth.
Even if he successfully prevents MS from enforcing only licensed software on its OSs, it still does not addresses the issue raised by RMS in The Right to Read, namely that copyright enforcement thru technology can turn all the World in a global police state in copyright owners' benefit.
Agreed. But that is because many of the protocols, APIs, file formats and hardware specifications are secret. The more we use GNU/Linux, more of this information get available and is put into good use.
Additionally, MS Windows can't keep your privacy, protect you from viruses, save you on hardware and software costs, give you the information and freedom GNU/Linux does.
Not quite. Government agencies, departments and state-owned companies do, but generally speaking they can't share your data with other branches of government. So while these may be seen as orthogonal issues, the digital certificates may be seen as a PR spin to convince citizens to allow the government to centralize all this data.
Now centralising all data *is* dangerous, prone to abuse and data security issues. But no one can protect citizens from their own stupidity if we allow government the leeway it naturally wants, because it is a human organization and all big enough human organizations usually almost have a mind of their own bent on ever growing and extending their power and influence.
He cannot do that unless all source code in the project is donated to the FSF. Obviously he can still profer his personal support.
It is not that it is contradictory. It is just that he is advancing proprietary software, when it only was possible for him to go so far by building on free software. The point is that he would never have got where he is if other people had applied to kernels and OSs what he now applies to source code control systems.
BSD had its own compiler. Anyway, the point here is that Linux depended on GNU tools to exist, and now does not upheld the same liberties it has enjoyed. Granted it is not so bad as plain software hoarding that almost killed BSD, but still is annoying and contradictory.
It is not code that requires integrity constraints, but data integrity. Data integrity is a function of the DBMS that is declared. That makes it much simpler, faster, more consistent than if one tries to do that with procedural coding, which ends up not covering up ad hoc use of the database anyway.
Interestingly enough, if everyone used the same arguments Linux, the kernel, would never had coming into being.
Just imagine that gcc and the other GNU utilities carried a similar license. Linus would have had to buy a real compiler and developer toolbox before even starting, and these cost real money in those days, much more than he could afford. And then no one would use his work or collaborate with him, because BSD and the other Unices were so much better for the task. The few tinkerers would have stayed with BSD.
Always when dealing with data, application code must check for constraints procedurally. This is extra, uncessarily complex code that is much better, simpler and more consistently done at the DBMS.
Ignore History at your own peril... the relational model for database management was exactly intended to provide declarative, centralised integrity constraints, because doing so in the application, no matter how good the application, is a sure recipe for failure. You don't cover integrity assurance for interactive, direct users of the database, and it is next to impossible to keep track of all integrity constraints and enforce them in all application programs.
That is my point. Having a real DBMS makes coding much, much simpler.
Oracle is a bad example of quality. Try PostgreSQL instead.
Who's talking about Oracle? PostgreSQL is as easy to manage as MySQL, but requires less coding and makes much more complex tasks easier for both developer and administrator.
Yes, in a totally unproven, and optional, implementation! While still lacking all kinds of declared integrity constraints, scalability and even sane documentation.
Nothing that makes data integrity optional or procedural is good for holding any organisational data...
Certainly dynamic web sites aren't among them. They call for concurrent data access and updateability, which is a typical DBMS use. Also, a good DBMS will ease the complexity of coding and improve reliability over simple file access.
A real DBMS -- a database is the organised data, not the software to control it -- does not necessarily cost anything, see PostgreSQL and others. And PostgreSQL does have all the advantages of MySQL, including speed and command line management, plus reliability and while requiring far simpler coding and keeping the data more consistent and accessible.
In a way yes, but not ultimately. Because the people behind MySQL and filesystems do not really grok the task, which is ultimately a database one and thus should be handled under the relational model.