>
The phrase "sell software" can be interpreted as an abbreviation of either of two things: "selling a copy of software"
"Selling a copy" is misguiding. One can distribute the copies for free, but charge for the license to use them in any significant way. It should be reserved for the strict selling of copies on media, even if the medium is intangible.
>
That proves nothing.
That proves exactly what I mean, that people can charge for copies and (or) licenses (to recent improvement), even if the license is the GNU GPL. Bundled support is indeed what drives prices sky-high, but this is exactly what is expected in a rational economy; copyright is all about creating artificial scarcity.
There isn't such a thing, if you insist on strictness... you can grant a license, and you can do it for free or for money, but that is not a sale in any way.
Ergo, if you have a copy, and you distribute it, especially if there is your software in that copy, and you do it for money, this is the thing closer to a sale you can have in software. Ergo, people who are selling gcc and Ada and such are selling in the sense you used it, but if you want to be strict, not even MS is selling software.
>
They are selling support, which is a different thing. The people who sell this support are not necessarily affliated with the authors of the software.
Red Hat's Cygnus division and the Ada guys do sell a bundle of software copies and support. They typically can do this at a high price because of the credibility of being the main authors of the stuff. This is quite the thing, if you get me.
>
The statement "any speech-free program is also beer-free if a non-trivial amount of copies have already been sold" is true
This is quite a different statement from your original one. Anyway, I think it is too general; it should be "can be had beer-free" instead of "is also beer-free", as the high prices commanded by Cygnus's GNU toolchain product prove.
>
most of the drivers are for "Linux", the kernel, not "GNU/Linux", the OS
Yes and no. (1) Linux is part of the GNU/Linux system; (2) there are many other pieces of GNU/Linux using drivers, such as media players, scanner interfaces, printing, and most importantly XFree; (3) many drivers come from BSD or Mach.
This article makes the assumption that Microsoft is currently or has in the past somehow inhibited hardware vendors.
Let's see... MS all but killed three RISC platforms -- MIPS, Alpha, PowerPC -- whose vendors had spent huge amounts of money promoting. MS stole wind from their own flavors of Unix, they promoted specific models for MS WNT which never sold well for there was no version of MS Office Pro and VS Basic for it.
>
Now, there are all sorts of "hardware vendors," but I would say universally most hardware vendors have bennifited tremendously from Microsoft, especially around Plug 'n Play (once Redmon got it working).
No, PnP was always a pain. What would have benefitted hardware vendors was wider adoption of EISA, earlier adoption of PCI and its fastest flavours, and a stable OS. MS had nothing to do with the first two and prevents the third to this day.
>
supporting the Linux OS is more painful than supporting the traditional unix vendors
I fail to see why. Old Unix had each its own hardware platform with different interfaces, while GNU/Linux runs in only a few platforms -- Alpha's dying as is PA-RISC, Clipper died, there is no more DEC TurboBus or Sun SBus, everything is IDE, SCSI, PCI, AGP, USB and FireWire. Creating drivers for GNU/Linux makes them portable, and it is easy in the first place, while old Unix had a different driver model for each platform and none were easy. The Haloween documents proved that even all MS effort to facilitate drivers develpment GNU/Linux drivers are still easier, and they cover nearly all the market instead of bein confined to one platform only as MS WNT currently is.
>
there is zilch technical support for linux, outside of the open source community
First, this is wrong. IBM, HP, Red Hat, SuSE and other do give support. Technical documentation and source code are much cheaper and better than what is available for any other platform, with the possible exception of BSD, incidentally another free software OS. Second, why the community isn't enough? The rules are clear: submit your driver to Linus, if it is good enough it will get all the criticising it needs to get finished. I wonder what more is needed in support for hardware vendors...
>
Most of the boutique hardware vendors cant afford the huge support teams to handle calls on every version of linux and all distros out there.
You obviously haven't the foggiest about GNU/Linux. There is precisely one stable, up-to-date version of the kernel available at each time. At this moment it is 2.4; all the variants of it are exactly equal AFA drivers are concerned. There is no reason whatsoever for a hardware vendor to support 2.5; 2.2 is still used but its drivers are much more similar to 2.4's than are those of MS WXP, WME and WCE.
>
they have a good deal of their IP in the software and they are leary of giving that away to competitors.
You mean trade secrets, because IP has no meaning apart from the aggregation of trade secrets, trademarks, copyrights and patents; obviously the last three are protected no matter what is published. As for trade secrets, I wonder why one would want its feeble protection instead of the much more substantial protections afforded by copyrights and patents. And even then your argument is bogus, because both the Linux kernel and the X Window System accept binary drivers, evil as they are.
Obviously you ignore the evilness of binary drivers: without source code it is impossible to audit and debug them thoroughtfully, and this is one of the causes for MS W32 unstability.
>
How the heck did they manage to crash a Debian-based Linux
In a word: superuser.
Lindows leads you to be always root. Without a password. Ouch! The security goes down the drain; I wouldn't be surprised to learn its security and stability are actually worse than MS-WXP's.
>
with open source apps, you need to have Lesstif, GTK+, GNOME
No you don't. I use here only Gnome apps -- and Emacs. No Qt, no KDE, no Motif, no Lesstif. Obviously I need Gtk+, because Gnome uses Gtk+ itself, which is a good thing and not redundant. Ah, and XForms for LyX until the Gnome frontend gets ready. I don't miss no functionality, sirra.
Usually I hate pontificating about things I don't know enough about......but asides from this guy suffering from lack of commercial abilities to market his work, and being struck with a terminal case of second-system syndrome, wasn't he struck by the dreaded binary modules Linus loophole?
Meaning, lots of embedded work takes place as modules to odd devices. Companies he complain about like Lineo, Caldera and Embeddix have success by working with binary modules, what is much more difficult for an individual without the resources to develop something in-house without community participation or without credibility to sign a NDA.
What do you mean? Something like elected representative to the State Assembly?
>
This is just petty and serves to make the Assembly-people less likely to listen to this group in the future.
Why petty? When you aren't big or rich enough, the Net may be a good way to get one's attention -- assuming your are/.ted or better yet locally publicised, like getting a reference at the regional and local news channels and papers.
Your argument against relational databases was anywhere near solid. Structure excess is an argument against excessive type definition; any relational database can have parts with precise type definition (where it matters) and others with simple types (that is, in the upper levels of Date's and Darwen's proposed type hierarchy).
I wonder if you tried to discuss this with Date or Darwen, and what kind of answers you got?
>
the "journalist" who wrote it describes disk storage as "memory"
Well, it is. Mass storage, or persistent memory. No reason for it not to be solid state. Actually, the big reason for the current conceptual difference at the OS level is the cost difference between solid state and disk storage.
Indeed, in a relational system cache would be much more efficient, enabling one to extend the current benefits of virtual memory to data as well: you don't worry about how to use it, you just know that buying more RAM will speed things up.
>
current file systems are described as wasting "memory"
Guess what? They do. An ACID RDBMS as data storage would transfer lots of duplicated functionality from userspace to the DBMS.
>
I don't understand the concerns of the poster regarding performance (at least without evidence of truly dismal performance)
This is a myth from the time of CODASYL databases. People looked at their programs and thought they would never perform as well if running on a relational database. Turned out that, notwithstanding poor coding, lack of relational model compliance from SQL and the plethora of quasi-SQL products around, SQL is much better than old CODASYL and CODASYL-like systems, specially -- but not restricted to -- interactive usage and multiple different applications using the same database.
>
For most users, the performance penalty is well worth the price.
Which penalty? There is none. If a relational database system doesn't perform well, the problem is the implementation or the particular database model, not the relational model.
There is no such thing. Despite the Marketspeak, relational databases are Mathematical and Formal Logic, based on Set Theory and Predicate Logic. So you can't have more-or-less relational, because there is no such thing as more-or-less mathematical or more-or-less logic. SQL for instance isn't.
>
it's a special purpose technology and has no place in a general purpose OS
Nonsense. Which would be the "special purposes"? Every operating system needs to store data. A relational DBMS can store data faster and more logically than any other in existence. It's a wonder that IT inertia is so big that there isn't yet a fully relational OS.
This is an implementation problem, has nothing to do with the data model. One of SQL's failures is exactly confounding data model with implementation, that is the logical with the physical levels.
>
Relational databases focus on data relationships
No, relationships (between entities) are for the ERDs; relational databases are constituted of relations (sets of tuples on domains) of data. It is a quite different thing.
There is a huge misunderstanding: SQL is corrupted, not derived, from the relational model. Since SQL deviated from it, the relational model has been developed significantly by Hugh Darwen and Chris J Date among and above others.
So MS is not proposing a relational data storage instead of a file system, but a SQL data storage. Or perhaps not even SQL, as MS SQL Server is notoriously almost as little SQL-compliant as Oracle itself, the king of standards deviations, second only to MySQL.
BTW, if anyone tries to tell you SQL is really relational, or quasi-relational, don't belive the Marketspeak. You can't be quasi-Mathematical, and the relational model is really an implementation of set theory and predicate logic. Tables aren't relations, just like 2+2=5 is not Algebra.
Pragmatically, this prevents SQL from being as flexible, simple and performant as a relational system is. There are some examples, unfortunately mostly historical: Ingres QUEL was found to perform consistently better than SQL systems, and much better than pre-SQL systems that had SQL grafted upon; BS12 was much cleaner, as was QUEL, than SQL; nowadays Alphora Dataphor is extensible in a much cleaner way than current OO or SQL OO systems.
In its defense, it may be that Microsoft uses only the data storage engine of MS SQL Server in WinFS, just as it uses it now for Exchange. Getting rid of SQL looses many interesting, powerful capabilities, but also may unleash performance. It even could be perfected as the engine of a future, nice, relational system, if ever Microsoft wakens up to what Alphora and TransRelational are (trying to) do. But I very much doubt; most of the database field, both academic and corporate, is corrupted by SQL money; there simply isn't enough cultural critical mass for a big company as Microsoft to steer its course from SQL to relational.
Even so, this could be the single most important contribution of Microsoft to the Informatics field, on a par with making the x86 computer a commodity. Even a sub-relational system can be so much better than what is around; this, together with Moore's Law making inherent Wintel inefficiencies irrelevant 80% of the cases, could pose a very big threat to the competitive position of free software. Granted the GNU/Hurd, or perhaps even Eros, could implement something even better than WinFS, truly relational and with orthogonal persistence in the physical layer (Eros) and/or a functional language application layer (GNU/Hurd). But there is no talk of doing so yet, and it would take a long time to bring these systems to where GNU/Linux is now.
>
What is important is that the best aspects of Plan 9 can be incorperated into existing platforms like Linux and *BSD and generate some real innovation without too much disturbance to the existing software base.
Wrong. To improve GNU/Linux or BSD we don't need Plan 9, only trial and error and perhaps some good systems engineering. What Plan 9 (and Hurd or Eros, BTW) brings to the table is something that can't just be grafted elsewhere: it is a different architecture with a different philosophy. Trying to incorporate bits may be possible, but not that relevant; incorporating important stuff would break compatibility and still be ugly.
That is the same kind of stuff the "hey, now Intel has the Alpha technology" crowd keeps missing.
First, you assume such a phenomenon has a single cause. It probably has several, none of them sufficiently preponderant to be singled out without qualification.
Second, you assume they were disliked by everyone. There are several other possibilities, like that they were feared by the religious authorities, that they were despised by the secular ones... I have no information about they being disliked by everyone. You know, incredible as it may seem to such a competitive, ahistorical and accomodated society as the US, not every persecution is justified.
>
they spent their whole time criticising everyone else for not being religious enough
Did they? I assumed they spent their time praying, studying the Bible, preaching the Gospel, helping each other and the needy, and criticised people for being religious hypocrites and noblemen for being hedonistic.
>
like their wonderfully enlightened behaviour in Salem
How deep into folly will men go to defend their unreasonable, comfy and cozy unbelief...
Just look at how widespread was witchhunt, how much time it endured, and how it ceased. And at the then current beliefs about witches. You will see it was a relatively mild persecution as far as persecutions usually go. Not much honest, but the guys ended up repenting, that is more than what is usually expected from your standard inquisitor.
>
if the problem is one faced by some reasonable number of similarly situated companies, then just buy something.
What if there is a free software project that already does 80%? It would be criminal to be dependent on a single company without access to the source code if you could just improve something already out there, and then have the option of outsourcing it to any number of bidders.
>
Robustness, maintenance, updates, new features, support etc. will be way better.
Not necessarily so. I worked once for such a lousy vendor, I was fired for standing for the client when he was being unknowingly screwed. Later I worked for another client, where this specific vendor was the single more important cause of stress in the office.
>
the original flamewar wasn't on Usenet, but on the legendary UNIXHATERS[...]. Something like [...] http://groups.google.com/groups?q=jargon+eric+grou p:list.unix-haters should yield sufficient tidbits. Failing that, the UNIXHATERS archives are floating around online.
Thank you, but Google turns out almost nothing even with esr for eric. Won't search further 'cause I've got to find a job...
when he first started to overhaul the Hacker's Dictionary, many of the original contributors were less than pleased with the treatment he was giving it.
Any references? I found only some, which seems to be from a later date.
Slashdot Mirror
"Selling a copy" is misguiding. One can distribute the copies for free, but charge for the license to use them in any significant way. It should be reserved for the strict selling of copies on media, even if the medium is intangible.
That proves exactly what I mean, that people can charge for copies and (or) licenses (to recent improvement), even if the license is the GNU GPL. Bundled support is indeed what drives prices sky-high, but this is exactly what is expected in a rational economy; copyright is all about creating artificial scarcity.
There isn't such a thing, if you insist on strictness... you can grant a license, and you can do it for free or for money, but that is not a sale in any way.
Ergo, if you have a copy, and you distribute it, especially if there is your software in that copy, and you do it for money, this is the thing closer to a sale you can have in software. Ergo, people who are selling gcc and Ada and such are selling in the sense you used it, but if you want to be strict, not even MS is selling software.
Red Hat's Cygnus division and the Ada guys do sell a bundle of software copies and support. They typically can do this at a high price because of the credibility of being the main authors of the stuff. This is quite the thing, if you get me.
This is quite a different statement from your original one. Anyway, I think it is too general; it should be "can be had beer-free" instead of "is also beer-free", as the high prices commanded by Cygnus's GNU toolchain product prove.
Not quite. There are quite a lot of people who pay big bucks for support for GNU/Linux, gcc, GNU Ada etc. And even for unsupported installation media.
Emacs Lisp. Copyleft. The FSF and the GNU Project. GCC. The Hurd.
BTW, why one needs to be original? One doesn't. It suffices to be moral.
Yes and no. (1) Linux is part of the GNU/Linux system; (2) there are many other pieces of GNU/Linux using drivers, such as media players, scanner interfaces, printing, and most importantly XFree; (3) many drivers come from BSD or Mach.
Let's see... MS all but killed three RISC platforms -- MIPS, Alpha, PowerPC -- whose vendors had spent huge amounts of money promoting. MS stole wind from their own flavors of Unix, they promoted specific models for MS WNT which never sold well for there was no version of MS Office Pro and VS Basic for it.
No, PnP was always a pain. What would have benefitted hardware vendors was wider adoption of EISA, earlier adoption of PCI and its fastest flavours, and a stable OS. MS had nothing to do with the first two and prevents the third to this day.
I fail to see why. Old Unix had each its own hardware platform with different interfaces, while GNU/Linux runs in only a few platforms -- Alpha's dying as is PA-RISC, Clipper died, there is no more DEC TurboBus or Sun SBus, everything is IDE, SCSI, PCI, AGP, USB and FireWire. Creating drivers for GNU/Linux makes them portable, and it is easy in the first place, while old Unix had a different driver model for each platform and none were easy. The Haloween documents proved that even all MS effort to facilitate drivers develpment GNU/Linux drivers are still easier, and they cover nearly all the market instead of bein confined to one platform only as MS WNT currently is.
First, this is wrong. IBM, HP, Red Hat, SuSE and other do give support. Technical documentation and source code are much cheaper and better than what is available for any other platform, with the possible exception of BSD, incidentally another free software OS. Second, why the community isn't enough? The rules are clear: submit your driver to Linus, if it is good enough it will get all the criticising it needs to get finished. I wonder what more is needed in support for hardware vendors...
You obviously haven't the foggiest about GNU/Linux. There is precisely one stable, up-to-date version of the kernel available at each time. At this moment it is 2.4; all the variants of it are exactly equal AFA drivers are concerned. There is no reason whatsoever for a hardware vendor to support 2.5; 2.2 is still used but its drivers are much more similar to 2.4's than are those of MS WXP, WME and WCE.
You mean trade secrets, because IP has no meaning apart from the aggregation of trade secrets, trademarks, copyrights and patents; obviously the last three are protected no matter what is published. As for trade secrets, I wonder why one would want its feeble protection instead of the much more substantial protections afforded by copyrights and patents. And even then your argument is bogus, because both the Linux kernel and the X Window System accept binary drivers, evil as they are.
Obviously you ignore the evilness of binary drivers: without source code it is impossible to audit and debug them thoroughtfully, and this is one of the causes for MS W32 unstability.
No it doesn't. In
In a word: superuser.
Lindows leads you to be always root. Without a password. Ouch! The security goes down the drain; I wouldn't be surprised to learn its security and stability are actually worse than MS-WXP's.
No you don't. I use here only Gnome apps -- and Emacs. No Qt, no KDE, no Motif, no Lesstif. Obviously I need Gtk+, because Gnome uses Gtk+ itself, which is a good thing and not redundant. Ah, and XForms for LyX until the Gnome frontend gets ready. I don't miss no functionality, sirra.
Usually I hate pontificating about things I don't know enough about... ...but asides from this guy suffering from lack of commercial abilities to market his work, and being struck with a terminal case of second-system syndrome, wasn't he struck by the dreaded binary modules Linus loophole?
Meaning, lots of embedded work takes place as modules to odd devices. Companies he complain about like Lineo, Caldera and Embeddix have success by working with binary modules, what is much more difficult for an individual without the resources to develop something in-house without community participation or without credibility to sign a NDA.
Or am I just smoking?
What do you mean? Something like elected representative to the State Assembly?
Why petty? When you aren't big or rich enough, the Net may be a good way to get one's attention -- assuming your are /.ted or better yet locally publicised, like getting a reference at the regional and local news channels and papers.
Yes, I know. I just thought he could finish it up for proper publication.
Great summing up. Wouldn't you want to publish it somewhere? I'd love to have it in my bookmarks.
I couldn't be convinced yet.
Your argument against relational databases was anywhere near solid. Structure excess is an argument against excessive type definition; any relational database can have parts with precise type definition (where it matters) and others with simple types (that is, in the upper levels of Date's and Darwen's proposed type hierarchy).
I wonder if you tried to discuss this with Date or Darwen, and what kind of answers you got?
Well, it is. Mass storage, or persistent memory. No reason for it not to be solid state. Actually, the big reason for the current conceptual difference at the OS level is the cost difference between solid state and disk storage.
Indeed, in a relational system cache would be much more efficient, enabling one to extend the current benefits of virtual memory to data as well: you don't worry about how to use it, you just know that buying more RAM will speed things up.
Guess what? They do. An ACID RDBMS as data storage would transfer lots of duplicated functionality from userspace to the DBMS.
This is a myth from the time of CODASYL databases. People looked at their programs and thought they would never perform as well if running on a relational database. Turned out that, notwithstanding poor coding, lack of relational model compliance from SQL and the plethora of quasi-SQL products around, SQL is much better than old CODASYL and CODASYL-like systems, specially -- but not restricted to -- interactive usage and multiple different applications using the same database.
Which penalty? There is none. If a relational database system doesn't perform well, the problem is the implementation or the particular database model, not the relational model.
There is no such thing. Despite the Marketspeak, relational databases are Mathematical and Formal Logic, based on Set Theory and Predicate Logic. So you can't have more-or-less relational, because there is no such thing as more-or-less mathematical or more-or-less logic. SQL for instance isn't.
Nonsense. Which would be the "special purposes"? Every operating system needs to store data. A relational DBMS can store data faster and more logically than any other in existence. It's a wonder that IT inertia is so big that there isn't yet a fully relational OS.
No, they didn't. They had a SQL-like system, but SQL is not relational.
This is an implementation problem, has nothing to do with the data model. One of SQL's failures is exactly confounding data model with implementation, that is the logical with the physical levels.
No, relationships (between entities) are for the ERDs; relational databases are constituted of relations (sets of tuples on domains) of data. It is a quite different thing.
There is a huge misunderstanding: SQL is corrupted, not derived, from the relational model. Since SQL deviated from it, the relational model has been developed significantly by Hugh Darwen and Chris J Date among and above others.
So MS is not proposing a relational data storage instead of a file system, but a SQL data storage. Or perhaps not even SQL, as MS SQL Server is notoriously almost as little SQL-compliant as Oracle itself, the king of standards deviations, second only to MySQL.
BTW, if anyone tries to tell you SQL is really relational, or quasi-relational, don't belive the Marketspeak. You can't be quasi-Mathematical, and the relational model is really an implementation of set theory and predicate logic. Tables aren't relations, just like 2+2=5 is not Algebra.
Pragmatically, this prevents SQL from being as flexible, simple and performant as a relational system is. There are some examples, unfortunately mostly historical: Ingres QUEL was found to perform consistently better than SQL systems, and much better than pre-SQL systems that had SQL grafted upon; BS12 was much cleaner, as was QUEL, than SQL; nowadays Alphora Dataphor is extensible in a much cleaner way than current OO or SQL OO systems.
In its defense, it may be that Microsoft uses only the data storage engine of MS SQL Server in WinFS, just as it uses it now for Exchange. Getting rid of SQL looses many interesting, powerful capabilities, but also may unleash performance. It even could be perfected as the engine of a future, nice, relational system, if ever Microsoft wakens up to what Alphora and TransRelational are (trying to) do. But I very much doubt; most of the database field, both academic and corporate, is corrupted by SQL money; there simply isn't enough cultural critical mass for a big company as Microsoft to steer its course from SQL to relational.
Even so, this could be the single most important contribution of Microsoft to the Informatics field, on a par with making the x86 computer a commodity. Even a sub-relational system can be so much better than what is around; this, together with Moore's Law making inherent Wintel inefficiencies irrelevant 80% of the cases, could pose a very big threat to the competitive position of free software. Granted the GNU/Hurd, or perhaps even Eros, could implement something even better than WinFS, truly relational and with orthogonal persistence in the physical layer (Eros) and/or a functional language application layer (GNU/Hurd). But there is no talk of doing so yet, and it would take a long time to bring these systems to where GNU/Linux is now.
Wrong. To improve GNU/Linux or BSD we don't need Plan 9, only trial and error and perhaps some good systems engineering. What Plan 9 (and Hurd or Eros, BTW) brings to the table is something that can't just be grafted elsewhere: it is a different architecture with a different philosophy. Trying to incorporate bits may be possible, but not that relevant; incorporating important stuff would break compatibility and still be ugly.
That is the same kind of stuff the "hey, now Intel has the Alpha technology" crowd keeps missing.
First, you assume such a phenomenon has a single cause. It probably has several, none of them sufficiently preponderant to be singled out without qualification.
Second, you assume they were disliked by everyone. There are several other possibilities, like that they were feared by the religious authorities, that they were despised by the secular ones... I have no information about they being disliked by everyone. You know, incredible as it may seem to such a competitive, ahistorical and accomodated society as the US, not every persecution is justified.
Did they? I assumed they spent their time praying, studying the Bible, preaching the Gospel, helping each other and the needy, and criticised people for being religious hypocrites and noblemen for being hedonistic.
How deep into folly will men go to defend their unreasonable, comfy and cozy unbelief...
Just look at how widespread was witchhunt, how much time it endured, and how it ceased. And at the then current beliefs about witches. You will see it was a relatively mild persecution as far as persecutions usually go. Not much honest, but the guys ended up repenting, that is more than what is usually expected from your standard inquisitor.
Me too. The issue is, if vendors are evil, and if customers are dumb, then there is little else to do in our current, decadent culture.
What if there is a free software project that already does 80%? It would be criminal to be dependent on a single company without access to the source code if you could just improve something already out there, and then have the option of outsourcing it to any number of bidders.
Not necessarily so. I worked once for such a lousy vendor, I was fired for standing for the client when he was being unknowingly screwed. Later I worked for another client, where this specific vendor was the single more important cause of stress in the office.
Thank you, but Google turns out almost nothing even with esr for eric. Won't search further 'cause I've got to find a job...
Any references? I found only some, which seems to be from a later date.