Re:SQL, XML not a Holy Grail: relational would be.
on
CNet on WinFS
·
· Score: 1
>
Care to elaborate on the "corruption of the relational model" bit?
The best would be to read the latest editions of Date's and Pascal's books, but just to sum it up:
Tables are not relations: they may contain duplicates and NULLs, thus breaking the power of and complicating significantly relation 2-valued logic and relational operators.
Lack of view updateability: thus lacking data independence.
Lack of user-defined types: thus making people think they need OODBMSs.
But don't take my word for it. If you can't get the books, some good arguments are delineated at http://dmoz.org./Computers/Software/Databases/Rela tional/
>
I know little of the relational calculus, but always assumed SQL was just a way of expressing it.
Actually SQL was just a half-baked prototype gone awry, and it is a mix of calculus and algebra to boot.
SQL, XML not a Holy Grail: relational would be.
on
CNet on WinFS
·
· Score: 1
SQL is just a corruption of the relational model, loosing power and simplicity. XML is just markup, and perhaps a nice programming convenience, but attempts to force feed it into databases and filesystems are bound to fail as miserably and costly as the similar attempts to force feed OO.
On the other hand, even SQL would be better than the current hierarchical file systems. And a truly relational database system such as Alphora Dataphor as a filesystem would be my technical Holy Grail, yes -- just not in MS-WNT. Make that work in GNU and I'd be happy as happy can be.
Noticing this is becoming more and more about Apple... sounds like what happened to the Mustang, as its first costumers grew older and richer the car grew and was tamed. Is this the fate of geeks, graduating (degrading?) from GNU/Linux on cheap Intel boxen to an Apple Macintosh?
I love the Mac, but the machine. I still run GNU/Linux and Gnome on it.
Copyrights are a concept, they are retained no matter what. But in copyright law, copyright notices have to be retained. This has nothing to do with the GNU GPL.
Moderators, how can you justify this being rated 4?
>
how does using KDE put money into McBride's pocket, if everything is GPL?
Trolltech sells licenses to non-free platforms ports of Qt, and support for all ports. So the more KDE users there are, the more proprietary programs using Qt, specially because KDE users have higher tolerance for non-free software than Gnome ones. Besides, free Gtk+ programs port just OK to non-free platforms, while for Qt ones, again, you need a license.
That said, the few shares Canopy has on Trolltech probably don't justify a KDE boycott... perhaps freedom preoccupation would, but then you'd be more GNU than RMS, since he's happy with Trolltech's dual-licensing of Qt.
>
Nobody is required to write QT-based software. Openoffice, Gnumeric, Mozilla all run fine in KDE
But KDE itself is based on KT. Need it running on a non-free platform, or its applications? Pay a fee.
>
Is Evolution bad just because Ximian has control over it?
AFAIR the only proprietary product of Ximian is a connector to other proprietary software.
I want for my managers the same I want for my colleagues and subordinates: education, not training. I want them to get the concepts so that we can have intelligent discussions on solutions, not that they dumbly memorise specific tools that will change sooner than later.
>
if the tons of smart people working on wine cant keep full compatibility
Because MS-W32 is a pain. POSIX is much easier.
>
sometimes we need a fresh start
Why?
Point is, as far as alternatives go, the GNU Hurd is a much more interesting one, and it is already here... efforts spent on pie-in-the-sky efforts should go towards helping GNU Hurd reach its 1.0 release.
>
XML is a good choice for storing data when you might want to extend the attributes of a record/properties of an object without breaking existing applications.
No, this would be the relational model for database management -- not SQL. XML fails due to its hierarchical nature and complexity.
The real sad thing is that once again x86 is the only architecture left with a fundamental market feature, alternative sources of supply.
I really want to see PowerPC or some other RISC succeeding in the mass market with GNU. But we need alternative sources of supply: SPARC is not targetting the mass market or GNU, PowerPC is left with IBM, Alpha is dying, MIPS also misses the mass market, ARM doesn't scale up... and x86 is fragmenting, and no one know if the future generations will be proprietary as too prevent alternatives.
Re: Everything, including tools, in moderation!
on
Software Fashion
·
· Score: 1
>>
Design patterns and UML were designed as practical tools, not dogma.
> note that the UML/patterns/OO newbie is in no position to determine that.
The thing is that all that, and XML too, are promoted as either dogma or magic bullets. Meanwhile, the real bullets are not magic at all but conceptual in nature: functional programming, relational database management, specification and testing, Brook's chirurgical development teams and so on.
But people won't realise that, because the current decadent culture wishes only HOWTOs, not textbooks, and centers so much on products and the short term that fails to ask developers for the tools they really need in the long term.
High-end Parkers -- not Vectors -- are reliable, agreeable pens. If you don't want cartdriges prices nor the mess of ink bottles, get a Schaffer snorkel (PFM) from eBay, they get their ink from the bottle thru a retractable tube that keeps the nip clean.
There are lots of misguided rants on the Net, and even some good stuff... but there are lots not yet there, specially if the authors are the kind of ethical lone rangers who don't get alond with corporate sponsors, or if the subject matter simply isn't simple or fashionable enough to be cheap to produce and publish on spare time.
For example, there is lots of stuff about various SQL and pseudo-SQL (think MySQL) implementations, but woefully little about the relational model of database management; there is lots about fashionable application servers, but little about their more solid antecessors, namely TP monitors, and the whole theory of transaction processing.
Not to mention real books that convey in-depth information are so much more pleasurable to read, at least until we have GNU/Linux UMTS handsets to read digital text.
>
I'm running several 16-bit Windows applications under Win2k right now
I'd be curious which, since what I have tried to run under MS W2K from MS W3.0 never worked. Granted it was not a huge sample.
But what do you mean by 16-bits MS Windows? MS has been known to keep broken behaviour and deviate from published APIs to keep compatibility with their own apps, at the expense of everyone else.
Also there were big changes between MS W3.0 and the final 16-bits release, MS WfW 3.11. MS WfW 3.11 was a stepping stone to MS W95, including the MS W32s API which is quite compatible with MS W2K WoW.
But MS WoW means you are running MS W16 inside of MS W2K, thus contributing to the already enormous bloat of the platform.
Meanwhile far older than MS W3.0 apps that were coded to X still work, without no more bloat than the old widgets themselves -- and very often simple recompilations bring them to newer versions which are shared by other apps, and you can recompile too because you have the source code.
A more valid comparision would be MS W2.X apps, but try finding these...
Yes... sad. Remember the Itsy? Digital had this cool GNU/Linux prototype in their labs for years, when Compaq bought them it was launched running MS WCE, named iPaq...
>
Do you mean that X is efficient for flat-color rectangles but inefficient for pixmap-laden interfaces?
No, he means that pixmaps over a network aren't nice, regardless of the underlying protocol.
In fact this problem is starting to go away with Gnome's SVG icons and other vectorial images.
>
X widget sets *also have* broken user level compatibility on a regular basis.
But X doesn't force you to use any particular version. With MS Windows, try running today that little MS W3.0 app... good luck. With Mac OS refusing to open up, without X MS Windows would have stagnated if MS's history is any guide.
>
Show me a good way of pasting one selection over another selection under X without retyping.
Erase first. I find it a great trade-off for the convenience of automatic copying and middle button pasting.
Why the Linux obsession? Linux the kernel isn't relevant to standards, no one codes to it. People code to the C library, which is the GNU libc, which follows closely the POSIX standard.
And speaking about the GNU/Linux system, it is far more standard than MS Windows, following not only POSIX but several other standards which MS simply ignores or corrupts.
>
the way I read the comment I was replying to, the software in question was the kernel itself
Well, if you think kernel, then indeed Linux is much more "standard" than MS Windows. You can use Linux the kernel wherever you want, for a big variety of platforms and applications, and implement different subsystems as well as exchange drivers with other kernels such as the BSDs, Machs and the OS Kit. It scales up and down, and you can tune or change it as you wish, including things MS Windows can't do such as realtime.
Meanwhile with MS Windows you simply can't use the kernel, only the APIs are documented and the same across different editions -- more or less. So as a kernel it isn't even there to be considered as a standard or even a product indeed.
That's why you need to have a strategy... such as periodically reviewing your storage and the means of reading it, as well as availability of readers; to be able to change when the particular medium becomes unsupported.
>
that doesn't wash the problem away
It is an explanation, not baptismal waters.
>
managers aren't going to change
Indeed, and taking the longer view that is why our culture is doomed.
>
Windows, by it's prevalance and the varied implementations (9x & NT families), is sort of a de facto standard
No, it isn't. For a de facto standard, you need at least two interoperable, independent implementations. While MS Windows is composed of many components, like SMB for file sharing, AD for authentication and Win32 or.Net for programming, there aren't yet independent, fully interoperable implementations of any. Samba, Wine and Mono try, but they aren't there yet and it is doubtful if they will ever be. Samba seems to be the best, but it is still partial and outdated, while better performant and more compatible than any of MS's own implementations.
On the other hand, POSIX and associated standards like NFS, NIS, LDAP, the X Window System and the such, and even Java, do have independent, interoperable implementations, suchs as Unix, GNU/Linux, BSD and the such, and a large body of software using them; while there is little Win32 or Mono or MFC software being compiled for independent implementations such as Wine or Mono.
Slashdot Mirror
The best would be to read the latest editions of Date's and Pascal's books, but just to sum it up:
Tables are not relations: they may contain duplicates and NULLs, thus breaking the power of and complicating significantly relation 2-valued logic and relational operators.
Lack of view updateability: thus lacking data independence.
Lack of user-defined types: thus making people think they need OODBMSs.
But don't take my word for it. If you can't get the books, some good arguments are delineated at http://dmoz.org./Computers/Software/Databases/Rela tional/
Actually SQL was just a half-baked prototype gone awry, and it is a mix of calculus and algebra to boot.
SQL is just a corruption of the relational model, loosing power and simplicity. XML is just markup, and perhaps a nice programming convenience, but attempts to force feed it into databases and filesystems are bound to fail as miserably and costly as the similar attempts to force feed OO.
On the other hand, even SQL would be better than the current hierarchical file systems. And a truly relational database system such as Alphora Dataphor as a filesystem would be my technical Holy Grail, yes -- just not in MS-WNT. Make that work in GNU and I'd be happy as happy can be.
Noticing this is becoming more and more about Apple... sounds like what happened to the Mustang, as its first costumers grew older and richer the car grew and was tamed. Is this the fate of geeks, graduating (degrading?) from GNU/Linux on cheap Intel boxen to an Apple Macintosh?
I love the Mac, but the machine. I still run GNU/Linux and Gnome on it.
But it is not public. It is a private contract.
No, it can't.
Copyrights are a concept, they are retained no matter what. But in copyright law, copyright notices have to be retained. This has nothing to do with the GNU GPL.
Moderators, how can you justify this being rated 4?
Trolltech sells licenses to non-free platforms ports of Qt, and support for all ports. So the more KDE users there are, the more proprietary programs using Qt, specially because KDE users have higher tolerance for non-free software than Gnome ones. Besides, free Gtk+ programs port just OK to non-free platforms, while for Qt ones, again, you need a license.
That said, the few shares Canopy has on Trolltech probably don't justify a KDE boycott... perhaps freedom preoccupation would, but then you'd be more GNU than RMS, since he's happy with Trolltech's dual-licensing of Qt.
But KDE itself is based on KT. Need it running on a non-free platform, or its applications? Pay a fee.
AFAIR the only proprietary product of Ximian is a connector to other proprietary software.
Pacifism is good for aggresive tyrants, d'ye know?
OTOH, if you mean to ask peace from Canopy, it is not a matter of asking. Either they are converted, or conquered.
I want for my managers the same I want for my colleagues and subordinates: education, not training. I want them to get the concepts so that we can have intelligent discussions on solutions, not that they dumbly memorise specific tools that will change sooner than later.
Not at all. Why all the computational and bandwidth complexity, plus hierarchical structure, if all you need for interchange is a schema definition?
But these are commonplace systems -- good and useful, but not flexible as this project strives to be. GNU Hurd is.
Because MS-W32 is a pain. POSIX is much easier.
Why?
Point is, as far as alternatives go, the GNU Hurd is a much more interesting one, and it is already here... efforts spent on pie-in-the-sky efforts should go towards helping GNU Hurd reach its 1.0 release.
No, this would be the relational model for database management -- not SQL. XML fails due to its hierarchical nature and complexity.
The real sad thing is that once again x86 is the only architecture left with a fundamental market feature, alternative sources of supply.
I really want to see PowerPC or some other RISC succeeding in the mass market with GNU. But we need alternative sources of supply: SPARC is not targetting the mass market or GNU, PowerPC is left with IBM, Alpha is dying, MIPS also misses the mass market, ARM doesn't scale up... and x86 is fragmenting, and no one know if the future generations will be proprietary as too prevent alternatives.
The thing is that all that, and XML too, are promoted as either dogma or magic bullets. Meanwhile, the real bullets are not magic at all but conceptual in nature: functional programming, relational database management, specification and testing, Brook's chirurgical development teams and so on.
But people won't realise that, because the current decadent culture wishes only HOWTOs, not textbooks, and centers so much on products and the short term that fails to ask developers for the tools they really need in the long term.
Yet it is but rehashing and hyping of 30-years old ideas like Fred Brooks' chyrurgical programming teams and good specification and testing practices.
High-end Parkers -- not Vectors -- are reliable, agreeable pens. If you don't want cartdriges prices nor the mess of ink bottles, get a Schaffer snorkel (PFM) from eBay, they get their ink from the bottle thru a retractable tube that keeps the nip clean.
Good cartdridge pens do have adaptors, they are much more convenient than syringes and usually have higher capacity than the original cartdriges.
I couldn't find in the article anything about SCO claiming breaches couldn't be fixed.
There are lots of misguided rants on the Net, and even some good stuff... but there are lots not yet there, specially if the authors are the kind of ethical lone rangers who don't get alond with corporate sponsors, or if the subject matter simply isn't simple or fashionable enough to be cheap to produce and publish on spare time.
For example, there is lots of stuff about various SQL and pseudo-SQL (think MySQL) implementations, but woefully little about the relational model of database management; there is lots about fashionable application servers, but little about their more solid antecessors, namely TP monitors, and the whole theory of transaction processing.
Not to mention real books that convey in-depth information are so much more pleasurable to read, at least until we have GNU/Linux UMTS handsets to read digital text.
I'd be curious which, since what I have tried to run under MS W2K from MS W3.0 never worked. Granted it was not a huge sample.
But what do you mean by 16-bits MS Windows? MS has been known to keep broken behaviour and deviate from published APIs to keep compatibility with their own apps, at the expense of everyone else.
Also there were big changes between MS W3.0 and the final 16-bits release, MS WfW 3.11. MS WfW 3.11 was a stepping stone to MS W95, including the MS W32s API which is quite compatible with MS W2K WoW.
But MS WoW means you are running MS W16 inside of MS W2K, thus contributing to the already enormous bloat of the platform.
Meanwhile far older than MS W3.0 apps that were coded to X still work, without no more bloat than the old widgets themselves -- and very often simple recompilations bring them to newer versions which are shared by other apps, and you can recompile too because you have the source code.
A more valid comparision would be MS W2.X apps, but try finding these...
Yes... sad. Remember the Itsy? Digital had this cool GNU/Linux prototype in their labs for years, when Compaq bought them it was launched running MS WCE, named iPaq...
No, he means that pixmaps over a network aren't nice, regardless of the underlying protocol.
In fact this problem is starting to go away with Gnome's SVG icons and other vectorial images.
But X doesn't force you to use any particular version. With MS Windows, try running today that little MS W3.0 app... good luck. With Mac OS refusing to open up, without X MS Windows would have stagnated if MS's history is any guide.
Erase first. I find it a great trade-off for the convenience of automatic copying and middle button pasting.
Why the Linux obsession? Linux the kernel isn't relevant to standards, no one codes to it. People code to the C library, which is the GNU libc, which follows closely the POSIX standard.
And speaking about the GNU/Linux system, it is far more standard than MS Windows, following not only POSIX but several other standards which MS simply ignores or corrupts.
Well, if you think kernel, then indeed Linux is much more "standard" than MS Windows. You can use Linux the kernel wherever you want, for a big variety of platforms and applications, and implement different subsystems as well as exchange drivers with other kernels such as the BSDs, Machs and the OS Kit. It scales up and down, and you can tune or change it as you wish, including things MS Windows can't do such as realtime.
Meanwhile with MS Windows you simply can't use the kernel, only the APIs are documented and the same across different editions -- more or less. So as a kernel it isn't even there to be considered as a standard or even a product indeed.
That's why you need to have a strategy... such as periodically reviewing your storage and the means of reading it, as well as availability of readers; to be able to change when the particular medium becomes unsupported.
It is an explanation, not baptismal waters.
Indeed, and taking the longer view that is why our culture is doomed.
No, it isn't. For a de facto standard, you need at least two interoperable, independent implementations. While MS Windows is composed of many components, like SMB for file sharing, AD for authentication and Win32 or .Net for programming, there aren't yet independent, fully interoperable implementations of any. Samba, Wine and Mono try, but they aren't there yet and it is doubtful if they will ever be. Samba seems to be the best, but it is still partial and outdated, while better performant and more compatible than any of MS's own implementations.
On the other hand, POSIX and associated standards like NFS, NIS, LDAP, the X Window System and the such, and even Java, do have independent, interoperable implementations, suchs as Unix, GNU/Linux, BSD and the such, and a large body of software using them; while there is little Win32 or Mono or MFC software being compiled for independent implementations such as Wine or Mono.