>
Common arguments such as "Sun was there first!" or "Alpha has been around..." are not accurate, since both of those product lines were marketed as servers or workstations.
Not really. At a time when SPARCs were at a sweet price spot Sun tried marketing them as personal systems, and Digital had always tried to make a dent in the desktop with Personal Workstantions and Multias. MS's failure to port its apps and tools hindered Digital.
>
I can understand Sun's initial reluctance to open-source Java years ago -- Microsoft would have (probably successfully) embraced and extended Java, as they indeed tried to do.
Wrong.
If Java was GNU GPL'd, MS would want distance from it.
MS tried to hijack Java 'cause they had a non-copyleft license to it.
What I'd like to know is if it would be possible to overwrite a, say, Apple Power Macintosh beige OldWorld G3 with LinuxBIOS or OpenBIOS and thus get to use x86 SCSI and VGA adapters.
As much as US costs are inflated by protectionism, it is still quite a rich country. It is not government funding missing, but rigorous teachers, interested pupils and demanding parents. Ah, and a more general social pressure.
Many ages ago, when people were still using Dr Halo for graphics and Ventura for publishing, I had two Englishmen for bosses. They told me the English power grid was already resilient, so that one never needed the battery in the alarm clock. Needless to say I didn't believe them.
In fact even if Fedora, Red Hat and every other RPM distro out there ceased to exist, even if SCO and MS had their way, still free software and open standards would make it easy to migrate to Unix or BSD. Now if MS folded where would its users go?
>
Hence this article's topic
This article is a nice example on how we're safer: we don't hide dirt under the rug.
>
a proven fact of how markets work
Are you trying to prove your economical ignorance? Markets don't inexorably tend to consolidation. There is always an exuberant phase in any market where it's fragmented, then it matures, but unless there is a strong network effect or monopolist practices unchecked it never kills diversity.
More than that, big corporations tend to senility, and new initiatives tend to carve niches or explore related markets. So it's more of a up-and-down movement than just consolidation.
>
an account of some vaporous scheme to wire the third world
Are you claiming Sao Paulo's 100 Telecenters with 3000+ X terminals serving 300000 people are vapourware? And that the Brazilian governmental program to reproduce Sao Paulo's initiative is just a fake? That LTSP.org doesn't exist? BTW, it is not just the Third World, even First World schools and community centres are using this 20-years-old architecture.
>
Let me know when you transition 120 million consumer PCs to dumb terminals running "GNU/Linux" and controlled from their ISPs central Brainiac super mainframe.
Actually by alternative platforms I was not referring to host-and-X-terminals as opposed to client-server, but to small and efficient RISCs as opposed to obsolete and bloated CISC.
>
with the nature of open source what benefits will Debian have that, say, Apache won't pick up on?
There is lots of integration in any given installation. The distro is where these issues gets hashed out by users, packagers and developers. Eventually the upstream gets it, but then he has to release, users have to learn about the fix and recompile. With aptitude it's a command line, no pressing need to follow each advisory unless you are a really sensitive site.
>
Choosing the proper version and keeping up with security updates is a very reasonable responsibility.
I don't think so. Installations can easily have dozens or hundreds of software providers to keep track of.
>
a minimalist distro
This is nice for learning and spending your time, but it isn't realistic professionally. One simply does not have the time to play his own distro packager.
Now if you were talking BSDs, where the whole ports mechanism is much more mature and the OS core is much simpler...
Consolidation is a phenomenon, therefore it dictates nothing.
Moreover it is in part a result of massification and proprietary lock-in, both of which are actually reversed by free software and open standards.
>
"killer apps" tend to cut down on diversity as well
Open standards apps are inherently safer, due to open discussion in specification phase.
>
Except in the consumer market
Where a market does exist. In Third World countries people have computing access in community centres operated with hosts and X terminals. This has potential to get bigger than First World consumer markets, especially with more and more companies going X terminals. Right now 2.6 is getting better and better at serving terminals, with scaling and scheduling improvements. Java, Flash, Gecko and OpenOffice.org are still problematic apps, but not inherently so.
>
See above.
Wrong. Servers were always viable in alternative platforms. It is free software and open standards that enable alternative platforms to the consumer market. See that what got IBM PowerPC efforts rolling in the XXI century wasn't neither MS WNT nor IBM OS/2, but Debian GNU/Linux which comes preinstalled with every PoP system sold (Eyetech and Genesi). GNU tools gave a boost to proprietary Unices and to BSDs, and GNU/Linux is giving StrongARM a lease on life, as well being the platform for both IA-64 and AMD-64. Also it is BSD which powers the Macintosh now.
>
How is this different than automatic updates on Windows
Automatic updates covers MS Windows, MS Office, MS BackOffice. aptitude covers all apps, and draws from a fundamentally saner development and deployment process to boot.
>
If it's built from the ground up with no services, what does 'Gentoo' have to do with security?
Integration and support. Debian and the 'Enterprise' distros are thoroughly tested and supported. With 'do-it-yourself' distros, good luck is needed to choose versions, testing and integrating them, and keeping up with security updates. In the end it takes time better spent actually auditing systems.
>
We _can not_ keep brushing things off and pretending they are not significant
Fully agree, but...
Other than going for OpenBSD and lacking some functionality, what else do you propose?
I do happen to think we should use vastly simpler systems: functional programming, perhaps Lisp, certainly all data relationally organised down to kernel level, multisserver microkernel, RISC implementation... but how realistic is this when POSIX simply has so much critical mass? This is not a technically-driven world, not even in free software or academia.
Slashdot Mirror
Not really. At a time when SPARCs were at a sweet price spot Sun tried marketing them as personal systems, and Digital had always tried to make a dent in the desktop with Personal Workstantions and Multias. MS's failure to port its apps and tools hindered Digital.
Check David McGoveran, Chris(topher) J Date, Fabian Pascal, Hugh Darwen, and BTW Eduard F 'Ted' Codd. Start at DMoz, then Google around...
The Third Manifesto, DBDebunk, DMoz.
Not as powerful and simple as a relational language would be.
It ain't. It is less powerful and more complex than a relational language would be.
Badly designed, too complex, not relational.
Better forget it, and anything non-relational too.
No, it hurts because it is not relational and because it was badly designed as a language anyway.
Looking like COBOL is not a problem in itself.
Any references about this CCPM?
Wrong.
If Java was GNU GPL'd, MS would want distance from it.
MS tried to hijack Java 'cause they had a non-copyleft license to it.
The 2940U is quite different. It never had a PowerDomain version.
Works on Matrox and 3dfx, but not ATi or nVidia.
And doesn't work on my Adaptec AHA-2940U.
What I'd like to know is if it would be possible to overwrite a, say, Apple Power Macintosh beige OldWorld G3 with LinuxBIOS or OpenBIOS and thus get to use x86 SCSI and VGA adapters.
As much as US costs are inflated by protectionism, it is still quite a rich country. It is not government funding missing, but rigorous teachers, interested pupils and demanding parents. Ah, and a more general social pressure.
Face it, the US goes downhill, slowly but surely.
And to think that HP was supposed to be the Debian champion...
Now why would PDF need replacement?
I bet any PDF page will have a smaller file size and better performance than the SVG equivalent.
Not to mention EPS.
Many ages ago, when people were still using Dr Halo for graphics and Ventura for publishing, I had two Englishmen for bosses. They told me the English power grid was already resilient, so that one never needed the battery in the alarm clock. Needless to say I didn't believe them.
If they didn't actually invested and therefore just want something for free, there is Fedora and a host of other RPM-based desktop distros.
But if they actually invested, they already have received Red Hat Enterprise Linux WS "for desktop/client systems" or Red Hat Professional Workstation "Enterprise Linux for personal use".
In fact even if Fedora, Red Hat and every other RPM distro out there ceased to exist, even if SCO and MS had their way, still free software and open standards would make it easy to migrate to Unix or BSD. Now if MS folded where would its users go?
This article is a nice example on how we're safer: we don't hide dirt under the rug.
Are you trying to prove your economical ignorance? Markets don't inexorably tend to consolidation. There is always an exuberant phase in any market where it's fragmented, then it matures, but unless there is a strong network effect or monopolist practices unchecked it never kills diversity.
More than that, big corporations tend to senility, and new initiatives tend to carve niches or explore related markets. So it's more of a up-and-down movement than just consolidation.
Are you claiming Sao Paulo's 100 Telecenters with 3000+ X terminals serving 300000 people are vapourware? And that the Brazilian governmental program to reproduce Sao Paulo's initiative is just a fake? That LTSP.org doesn't exist? BTW, it is not just the Third World, even First World schools and community centres are using this 20-years-old architecture.
Actually by alternative platforms I was not referring to host-and-X-terminals as opposed to client-server, but to small and efficient RISCs as opposed to obsolete and bloated CISC.
I am not a developer, and am not bashing.
There is lots of integration in any given installation. The distro is where these issues gets hashed out by users, packagers and developers. Eventually the upstream gets it, but then he has to release, users have to learn about the fix and recompile. With aptitude it's a command line, no pressing need to follow each advisory unless you are a really sensitive site.
I don't think so. Installations can easily have dozens or hundreds of software providers to keep track of.
This is nice for learning and spending your time, but it isn't realistic professionally. One simply does not have the time to play his own distro packager.
Now if you were talking BSDs, where the whole ports mechanism is much more mature and the OS core is much simpler...
Consolidation is a phenomenon, therefore it dictates nothing.
Moreover it is in part a result of massification and proprietary lock-in, both of which are actually reversed by free software and open standards.
Open standards apps are inherently safer, due to open discussion in specification phase.
Where a market does exist. In Third World countries people have computing access in community centres operated with hosts and X terminals. This has potential to get bigger than First World consumer markets, especially with more and more companies going X terminals. Right now 2.6 is getting better and better at serving terminals, with scaling and scheduling improvements. Java, Flash, Gecko and OpenOffice.org are still problematic apps, but not inherently so.
Wrong. Servers were always viable in alternative platforms. It is free software and open standards that enable alternative platforms to the consumer market. See that what got IBM PowerPC efforts rolling in the XXI century wasn't neither MS WNT nor IBM OS/2, but Debian GNU/Linux which comes preinstalled with every PoP system sold (Eyetech and Genesi). GNU tools gave a boost to proprietary Unices and to BSDs, and GNU/Linux is giving StrongARM a lease on life, as well being the platform for both IA-64 and AMD-64. Also it is BSD which powers the Macintosh now.
Automatic updates covers MS Windows, MS Office, MS BackOffice. aptitude covers all apps, and draws from a fundamentally saner development and deployment process to boot.
Perhaps. I'd propose you help Debian maintainers with this. But will it need attention from sysadmins? If so, one more issue to deal with...
Yet it does not solve what I think are the basic issues: excess complexity, lack of code auditing and lack of maintenance.
Integration and support. Debian and the 'Enterprise' distros are thoroughly tested and supported. With 'do-it-yourself' distros, good luck is needed to choose versions, testing and integrating them, and keeping up with security updates. In the end it takes time better spent actually auditing systems.
Fully agree, but...
Other than going for OpenBSD and lacking some functionality, what else do you propose?
I do happen to think we should use vastly simpler systems: functional programming, perhaps Lisp, certainly all data relationally organised down to kernel level, multisserver microkernel, RISC implementation... but how realistic is this when POSIX simply has so much critical mass? This is not a technically-driven world, not even in free software or academia.
This ain't exactly source code.