>
Could you please elaborate a bit on why SQL is not relational?
You'd be better reading Date, Darwen and Pascal; but in a general way, SQL is full of conceptual errors and arbitrary limitations.
As for arbitrary limitations, the biggest one is lack of user-defined types. But the real problem is the conceptual side; for example, SQL allows for tables that are not relations, because a relation is a set and without a primary key an SQL table is just a bag.
The end result is that SQL is a lot less powerful and more complicated than it should be. And incidentally this is why people keep thinking they need an OODB or XMLDB or whatever, each of which are actually necessarily less powerful and more complicated than the relational model too.
It is not much as compared to other transgressors like MySQL or even PostgreSQL, but it is sad how developers can't grasp that SQL is in violation of the relational model. Claiming anything SQL is relation is like claiming 2+2=5 is some kind of 'improved' or 'practical' or 'good enough' algebra.
>
my microprocessor will probably never even have a C compiler for it so this wouldn't make much of a difference anyway
While I appreciate a Lisp or Smalltalk based microprocessor should be much more efficient, how would one get the network economical effects? I'd love to see such a system, but even in the US they couldn't push it: Lisp systems and Smalltalk workstations died, at least for now.
>
it is not possible to make a really modern OS if you demand POSIX compatibility
I like to think of the Hurd as a bridge to a reborn Lisp system. The other side of the river may be greener, but you need a bridge to the other side, that will carry you there while keeping access to your old shackle this side while you build your mansion at the other one.
In other words, if we can just finish the GNU Hurd, then we can start building a Lisp (or Smalltalk) codebase. First by porting existing Lisp-on-C systems such as GNU Emacs, GNUCash, the GIMP, TeXmacs, perhaps an AutoCAD clone, the Sawfish window manager and whatever else -- take Squeak and whatever else if that's your idea of code Paradise -- then by creating new apps there that will mirror their C equivalents, including file format compatibility with OpenOffice.org and the such; and finally enabling people to get rid of the C personality. Somewhere after we start having Lisp or Smalltalk personality programs (not only Lisp-on-C or Smalltalk-on-C programs) it might be a better economical proposition to have specialised uprocessors. Not before.
But I have a hunch against OO. While it is not hard to be better than C, I still think it lacks functional programming elegance.
>
By that logic, mm should be "meter meter" or meters squared. But it's not, it's defined to mean millimeter.
Thanks, I just forgot the initial minor m is mili, or a thousandth. So the error in question is still worse, because isolated it should mean Meter but as a quantifier it should mean a thousandth.
>
The reason people just use m instead of M in those cases is because at the end of a word, people are more accustomed to using a non-capital letter.
Just that it is not a word. The real reason here, I suspect, is the bad state of education in general and teaching of the metric system in particular, at least in the US but elsewhere too.
Sorry if this is in any way pedantic - just FYI since I used to work in a capital markets trading environment... The abbreviation in most currency markets is not UKP, it's GBP
Also, the SI unit abbreviation for million is M, not m. m is Meter, M is million (mega), so a mM is a million meters (a thousand Km), but Mm, MM or mm don't make any sense at all, nor mGBP.
If this is real it is good for the next generations. IT is just too stressful for would-be mothers.
Oh wait, women in the First World don't want to be mothers until they're rich or realised or what not, and then only one child... don't worry, in a few generations the Third World will take over by sheer numbers. Remember, the Barbarians did that to Rome once, we'll do it again.
>
Would you be complaining that Linux can't run Internet Explorer just because some websites are IE-only?
I am not complaining about GNU/Linux, but about proprietary vendors.
>
the problem is with those individual sites. I know you probably hear it too often, but complaining does produce change
You'd be surprised to know how people can be callous, and how often I do complain.
>
there are many banks out there, and it's usually easy to switch from one to another
Not so fast... when you're at Brazil it is no always necessarily possible or even easy to avoid Banco do Brasil. Not a monopoly, but they do somethings no other bank will or can do.
>
I'm on x86, and I go out of my way to uninstall them
My wife's web banking is off limits for us because of lack of j2re. Many employment agencies and hiring companies are off limits because of Flash sites. I haven't found anything but Adobe Acrobat that allows me to fill PDF forms.
It is funny but disheartening to see how even Slashdot editors can't remember yesterday... before Intel (or HP or Compaq, you name 'em) killed the Alpha, Linus was given a four-way Alpha workstation he used for quite some time, I think it was two or three years until x86 hardware took over in performance (over his three-years old system!) or Alpha was seen as a dead end or whatever.
So he's just doing the same, this time with a platform not so fancy but with a safer future.
It means an easier life for us Linuxers on PPC, but we were already blessed with great hackers both on the kernel and in other parts; for example the leader(s?) of the Debian X Strike Force are Linux on PPC users.
Now what would be great is if proprietary vendors start porting their stuff... every day I miss things like j2re plugin for Mozilla, a Flash player, Adobe Acrobat and NX. Granted there are alternatives and clones, but gcjwebplugin still crashes Epiphany and ain't Java 2 level yet, swf_player is only playback, no interaction and takes way too much CPU, Evince doesn't do PDF forms and X.Fast (LBX) simply can't work in POTS dial-up situations where NX shines.
Guess what? Not quite. In Mac OS X Mach was merged with the BSD kernel, loosing any advantage -- or even characteristic -- of the microkernel design but keeping a few liabilities.
So booting Mac OS X would help nothing, but the GNU Hurd yes.
>
Counting parenthesis on a vax (iirc) was a freaking pain.
As Borland proved with the Turbo Pascal IDE and Microsoft with its Visual line, the IDE can make all the difference... try Emacs parenthesis highlighting.
>
that ain't gonna happen anytime soon, so we must learn to live with it
So why waste mediocre programmers in write-only languages such as the C-derived ones? Better give them COBOL or Pascal and let them toil with known quantities...
>
the same problem, just shifted to a slightly different area (library design instead of language).
At least you get decent languages that can be extended in different directions instead of changing the languages of choice every few years...
Standardising libraries is not a big deal.
The jump really is getting managers to value technical excellence instead of just wanting cheap programmers. There are quite a few people using IBM's Visual Age Smalltalk, or some commercial Lisp, but the hype is always dumber.
>
It is because of what Pascal was developed for that it's structure is far too inflexible for a lot of serious software development.
What is serious? Obviously you won't create an OS or Office suite in Pascal. But if you have average programmers to create or extend custom commercial applications, it's far more interesting than, say, Microsoft Visual BASIC. No wonder Borland Delphi came so much later and still got a sizeable chunk of the market. If your programmers are dumber, go COBOL; if smarter, go functional or OO.
>
C was created to write portable operating systems
So use it for that. And for things like Office suites, basic utilities... not for custom commercial apps.
So you take a mess and make it messier still... worse, VB is used mainly to create DB apps, and what we need is a relational DB to substitute SQL; until then, all this language shifting will have only marginal improvements.
>
it has its place... mainly for entry-level programmers to stumble about their code./i>
So no need for OO... and BTW, these guys could use Pascal or COBOL and have a better language at reach, better portability... should guys who need BASIC to do something be allowed anywhere near commercial apps?
>
every programmer should have to learn a billion programming languages
Who said that? Just take an approach... if it's custom development by dumb programmers, take COBOL; if you have slightly better programmers, Pascal; if you really have good programmers, go OO with Smalltalk or functional with any Lisp derivative, Haskell or ML. Reserve BASIC for teaching COBOL wannabes and C derivatives for systems programming.
I am old enough to remember the original VB columnist at some high-profile magazine (was it Dr Dobb's itself?) throwing the towel on the column because he couldn't stand the bloating of the language by MS... and the C++ Advisor-or-something-the-like columnist (was it Unix Magazine or what?) quitting the column because C++, being designed by committee, required a language lawyer and was only getting worse.
No news here. If you don't care for elegance, you go awok with evolution. ISO SQL, Perl, there are many many examples.
Now if only people would rethink and take the pain of learning a real, elegant language... a functional (Lisp, Scheme, Haskell, ML) or pure OO (Smalltalk, Squeak) or truly relational (Tutorial D, D4) one.
Instead of just trying to keep extending known languages into unknown fields. C is just structure, platform-independent Assembly; how come people want to create custom applications in it or its Java, C++, C#, ObjectiveC? This comes only as an indictment of the alternatives, or worse still of programmers and their managers.
And BASIC, it was only a stepping stone in learning COBOL. How come it is used to deploy anything more than a prototype? Don't get me started with excuses.
It is high time managers and programmers get real and start using languages designed to do what they want. COBOL, Pascal, Smalltalk, Lisp... each in their niche, they are better than C or BASIC and their overextended derivatives.
>
how are they going to enforce/prove it if someone looks at their code and gets ideas from it
That's what patents are for, and MS has been known file quite some lately. Also, they have the option of isolating what they consider their 'most innovative' pieces in libraries still hidden from view. Finally, if you are good enough to get ideas from them without incurring in copyright infringement by inadvertently doing derivative work by inconscient memory afterwards, you are probably worth your weight in gold.
But seriously, how innovative is MS SQL Server? It is as decent a Transact SQL implementation as it can get in a substandard OS as MS Windows. It does not conform fully to the ISO SQL standards, it contradicts the relational model, the only interesting thing it has is integration with and extension of MS.Net's type system -- not exactly rocket science, PostgreSQL for example has had an extensible type system for years and neither of them conform to, for example, Date's and Darwen's specialisation by constraint type inheritance system.
In the Brazilian Southern region we have a telephony operator that gives us a second POTS line dedicated to their in-house ISP. Just dial 0, no dial tone, basic PPP authentication and you're in for a flat fee.
If I'm not mistaken in Europe there are places with similar schemes for ISDN, where one gets digital, real 64Kbps instead of a maximum theoretical 56Kbps with analog.
Nice, but how current it is? I'd assume it is older than UltraSPARC, being that Leon is 32 bits, but might still be useful if it can reach, say, 600 MHz or so.
And, can I buy a SPARC workstation with a chip manufactured by anyone else than Sun?
>
there's also an open source implementation: the Leon and it's supported by Linux.
What are the chances of someone using that, or a derivative, into a general-purpose computing product?
Slashdot Mirror
You'd be better reading Date, Darwen and Pascal; but in a general way, SQL is full of conceptual errors and arbitrary limitations.
As for arbitrary limitations, the biggest one is lack of user-defined types. But the real problem is the conceptual side; for example, SQL allows for tables that are not relations, because a relation is a set and without a primary key an SQL table is just a bag.
The end result is that SQL is a lot less powerful and more complicated than it should be. And incidentally this is why people keep thinking they need an OODB or XMLDB or whatever, each of which are actually necessarily less powerful and more complicated than the relational model too.
It is not much as compared to other transgressors like MySQL or even PostgreSQL, but it is sad how developers can't grasp that SQL is in violation of the relational model. Claiming anything SQL is relation is like claiming 2+2=5 is some kind of 'improved' or 'practical' or 'good enough' algebra.
While I appreciate a Lisp or Smalltalk based microprocessor should be much more efficient, how would one get the network economical effects? I'd love to see such a system, but even in the US they couldn't push it: Lisp systems and Smalltalk workstations died, at least for now.
I like to think of the Hurd as a bridge to a reborn Lisp system. The other side of the river may be greener, but you need a bridge to the other side, that will carry you there while keeping access to your old shackle this side while you build your mansion at the other one.
In other words, if we can just finish the GNU Hurd, then we can start building a Lisp (or Smalltalk) codebase. First by porting existing Lisp-on-C systems such as GNU Emacs, GNUCash, the GIMP, TeXmacs, perhaps an AutoCAD clone, the Sawfish window manager and whatever else -- take Squeak and whatever else if that's your idea of code Paradise -- then by creating new apps there that will mirror their C equivalents, including file format compatibility with OpenOffice.org and the such; and finally enabling people to get rid of the C personality. Somewhere after we start having Lisp or Smalltalk personality programs (not only Lisp-on-C or Smalltalk-on-C programs) it might be a better economical proposition to have specialised uprocessors. Not before.
But I have a hunch against OO. While it is not hard to be better than C, I still think it lacks functional programming elegance.
But is it POSIX? The government shouldn't be in the business of dictating implementations, but interfaces. Links?
Support isn't superfluous, urging is.
I wonder why, if it is an already estabilished State policy in Brazil.
And no, I didn't RTFineA.
Thanks, I just forgot the initial minor m is mili, or a thousandth. So the error in question is still worse, because isolated it should mean Meter but as a quantifier it should mean a thousandth.
Just that it is not a word. The real reason here, I suspect, is the bad state of education in general and teaching of the metric system in particular, at least in the US but elsewhere too.
Also, the SI unit abbreviation for million is M, not m. m is Meter, M is million (mega), so a mM is a million meters (a thousand Km), but Mm, MM or mm don't make any sense at all, nor mGBP.
If this is real it is good for the next generations. IT is just too stressful for would-be mothers.
Oh wait, women in the First World don't want to be mothers until they're rich or realised or what not, and then only one child... don't worry, in a few generations the Third World will take over by sheer numbers. Remember, the Barbarians did that to Rome once, we'll do it again.
I am not complaining about GNU/Linux, but about proprietary vendors.
You'd be surprised to know how people can be callous, and how often I do complain.
Not so fast... when you're at Brazil it is no always necessarily possible or even easy to avoid Banco do Brasil. Not a monopoly, but they do somethings no other bank will or can do.
My wife's web banking is off limits for us because of lack of j2re. Many employment agencies and hiring companies are off limits because of Flash sites. I haven't found anything but Adobe Acrobat that allows me to fill PDF forms.
Any tips welcome.
It is funny but disheartening to see how even Slashdot editors can't remember yesterday... before Intel (or HP or Compaq, you name 'em) killed the Alpha, Linus was given a four-way Alpha workstation he used for quite some time, I think it was two or three years until x86 hardware took over in performance (over his three-years old system!) or Alpha was seen as a dead end or whatever.
So he's just doing the same, this time with a platform not so fancy but with a safer future.
It means an easier life for us Linuxers on PPC, but we were already blessed with great hackers both on the kernel and in other parts; for example the leader(s?) of the Debian X Strike Force are Linux on PPC users.
Now what would be great is if proprietary vendors start porting their stuff... every day I miss things like j2re plugin for Mozilla, a Flash player, Adobe Acrobat and NX. Granted there are alternatives and clones, but gcjwebplugin still crashes Epiphany and ain't Java 2 level yet, swf_player is only playback, no interaction and takes way too much CPU, Evince doesn't do PDF forms and X.Fast (LBX) simply can't work in POTS dial-up situations where NX shines.
Guess what? Not quite. In Mac OS X Mach was merged with the BSD kernel, loosing any advantage -- or even characteristic -- of the microkernel design but keeping a few liabilities.
So booting Mac OS X would help nothing, but the GNU Hurd yes.
As Borland proved with the Turbo Pascal IDE and Microsoft with its Visual line, the IDE can make all the difference... try Emacs parenthesis highlighting.
So why waste mediocre programmers in write-only languages such as the C-derived ones? Better give them COBOL or Pascal and let them toil with known quantities...
At least you get decent languages that can be extended in different directions instead of changing the languages of choice every few years...
Standardising libraries is not a big deal.
The jump really is getting managers to value technical excellence instead of just wanting cheap programmers. There are quite a few people using IBM's Visual Age Smalltalk, or some commercial Lisp, but the hype is always dumber.
What is serious? Obviously you won't create an OS or Office suite in Pascal. But if you have average programmers to create or extend custom commercial applications, it's far more interesting than, say, Microsoft Visual BASIC. No wonder Borland Delphi came so much later and still got a sizeable chunk of the market. If your programmers are dumber, go COBOL; if smarter, go functional or OO.
So use it for that. And for things like Office suites, basic utilities... not for custom commercial apps.
So what? It's still BASIC at heart.
So you take a mess and make it messier still... worse, VB is used mainly to create DB apps, and what we need is a relational DB to substitute SQL; until then, all this language shifting will have only marginal improvements.
So no need for OO... and BTW, these guys could use Pascal or COBOL and have a better language at reach, better portability... should guys who need BASIC to do something be allowed anywhere near commercial apps?
Who said that? Just take an approach... if it's custom development by dumb programmers, take COBOL; if you have slightly better programmers, Pascal; if you really have good programmers, go OO with Smalltalk or functional with any Lisp derivative, Haskell or ML. Reserve BASIC for teaching COBOL wannabes and C derivatives for systems programming.
I am old enough to remember the original VB columnist at some high-profile magazine (was it Dr Dobb's itself?) throwing the towel on the column because he couldn't stand the bloating of the language by MS... and the C++ Advisor-or-something-the-like columnist (was it Unix Magazine or what?) quitting the column because C++, being designed by committee, required a language lawyer and was only getting worse.
No news here. If you don't care for elegance, you go awok with evolution. ISO SQL, Perl, there are many many examples.
Now if only people would rethink and take the pain of learning a real, elegant language... a functional (Lisp, Scheme, Haskell, ML) or pure OO (Smalltalk, Squeak) or truly relational (Tutorial D, D4) one.
Instead of just trying to keep extending known languages into unknown fields. C is just structure, platform-independent Assembly; how come people want to create custom applications in it or its Java, C++, C#, ObjectiveC? This comes only as an indictment of the alternatives, or worse still of programmers and their managers.
And BASIC, it was only a stepping stone in learning COBOL. How come it is used to deploy anything more than a prototype? Don't get me started with excuses.
It is high time managers and programmers get real and start using languages designed to do what they want. COBOL, Pascal, Smalltalk, Lisp... each in their niche, they are better than C or BASIC and their overextended derivatives.
That's what patents are for, and MS has been known file quite some lately. Also, they have the option of isolating what they consider their 'most innovative' pieces in libraries still hidden from view. Finally, if you are good enough to get ideas from them without incurring in copyright infringement by inadvertently doing derivative work by inconscient memory afterwards, you are probably worth your weight in gold.
But seriously, how innovative is MS SQL Server? It is as decent a Transact SQL implementation as it can get in a substandard OS as MS Windows. It does not conform fully to the ISO SQL standards, it contradicts the relational model, the only interesting thing it has is integration with and extension of MS .Net's type system -- not exactly rocket science, PostgreSQL for example has had an extensible type system for years and neither of them conform to, for example, Date's and Darwen's specialisation by constraint type inheritance system.
In the Brazilian Southern region we have a telephony operator that gives us a second POTS line dedicated to their in-house ISP. Just dial 0, no dial tone, basic PPP authentication and you're in for a flat fee.
If I'm not mistaken in Europe there are places with similar schemes for ISDN, where one gets digital, real 64Kbps instead of a maximum theoretical 56Kbps with analog.
Nice summing up, but I could find nothing at Power.org, and I'd definetly want to read something by either them or IBM itself.
No you can't. IBM gets quite some money to license out POWER.
Nice, but how current it is? I'd assume it is older than UltraSPARC, being that Leon is 32 bits, but might still be useful if it can reach, say, 600 MHz or so.
And, can I buy a SPARC workstation with a chip manufactured by anyone else than Sun?
What are the chances of someone using that, or a derivative, into a general-purpose computing product?
You mean KiB of RAM.