Java is almost a set of open standards. Language, Bytecode, Libraries, etc are well documented, and there is the JCP (Java Community Process), where everybody can participate in the evolution of java. (Okay, Sun is still more equal than everybody else is, because they have a veto right.) I don't know, why Sun withdraw the standard requests, but in my eyes they look, like trying to keep Java unpolluted and unforked in the first place.
If you take a look at the enterprise sector, Sun does only provide a spec and leaves the implementation to others like BEA or Orion.
As far as I understand, the XPU128 itself doesn't run Linux:
The GNU C Compiler and related tools allow programming of sequential parts of algorithms for the Risc processor. The API for Communication with Host supports NT and Linux operating systems.
Parallel parts of the algorithms are programmed in the high level language LELA or Assembler language. A communication class library provides simple integration of C programs and XPU programs.
Linux is only running at the risc processor, used as an interface between the XP128 and the PC.
I don't believe, you can translate C to efficient parallel code. applications and algorithms have been optimized for the common sequential architectures over the last 20, 30 years. Switching to massive parallel devices will demand new software.
Yes, Sun seems to be a little bit boring at first, but I think this is exaclty what the industry wants. And no changing roadmaps twice a year.;-)
Itanium: Probable winner
I agree, too. Not because the're good, but because the're from Intel.
Sledgehammer: Expected no-show
We will see. But I don't expect gread marketshare. Well, maybe, if AMD get's more industry support for their x86 stuff in the meantime.
Alpha: Dying, Also Ran
Yes, Compaq seems to be quite happy with Intel chips, no make competition inhouse.;-)
MIPS: Out of production, Also Ran
Still in production, but IMHO only with a vague future. SGI is bouncing from MIPS to Intel (NT and Linux workstations) and back to MIPS (Origin 3000). Looks for me like a looser in the last years.
HPPA: Possible contender if anyone knew about 'em.
Nope. Intel developed the Itanium together with HP, and even if HP has extended its HP-HP roadmap, the future at HP is IA64. For example, the new SuperDome is already prepared for IA64.
It is not well defined, what an operating system is or has to be. The most common definition could be an interface between the user(s) and the hardware. While we don't have to argue about the low level funktions provided by the kernel, there is much room to discuss wether the user level applications, packaged together with the kernel, are part of the os, too.
If you do this (and the original author does it), I don't see why you should differ between the "classic" Unix utilities and the graphical ones running under X11.
No OS is just the same as some years ago. Neither MacOS nor Windows had a browser included five years ago. Now they have. Browsers are included in almost any Unix software box, you can buy today. Why count with differend measures?
Unix is, in costrast to Windows or MacOS, much more modular and versatile. The core parts got some additional layers. X11 for example is such a layer and the foundation for graphical applications. Adding window management and an email client is the purpose of destop environments like Gnome.
Unix has also been a multiuser/multitasking OS for a very, very long time. So it has a strong position in the server market. There you don't need (primary) an web client, but a web server. (Linux, as a flavor of the Unix idea, is so versatile, it run's almost everywhere from a IBM mainframe to a wrist watch)
Unix has, and this is the main point, in which I could agree with the original author, a weak history on the desktop. The desktop environments and the applications are not as mature as they are for Windows or MacOS. But the topic was not about the quality of the Unix desktops, but about Unix not being a OS at all.
And I think, you need a pretty weird definition of an OS to declare Unix as being not being one.
It wouldn't just help the Apps company, it's the only way to get Office ported.
If the integration between os and app is so tight, as everybody says, they (the apps company) will have a very hard time to get ports done without the source (and the most interesting details) of the OS libraries like COM, OLE, etc).
I remember times, when IBM sold their own version of Netware. The box looked exactly the same than the original one from Novell, but it was not red, it was blue.:-)
ohh great yet another STANDARDISED linux UI, with any luck we will have a whole plethora od STANDARDISED UI's
I think, it's not another Linux UI, but Sun and others joining their forces to push Gnome. This is remarkable in my eyes because
Sun has commited itself to Solaris and not to Linux. It's the first among the big Unix players to drop CDE as the default desktop for their own Unix. So others like IBM or HP will hopefully follow. One standard desktop for all Unix flavors has been a dream for many, many years.
Gnome and KDE are still lacking real applications which integrate with the desktop. If I understand the anouncement right, Gnome will get a bunch of very valuable components. This is not so new news, because Sun announced to use Gnome for StarOffice 6 before.
If they succeed to build a standardised component architecture, where you can replace the StarOffice spreadsheet component with the Gnumeric component without problems, their have some clear advantage over Microsoft Windows. Even if COM components are old and mature technology there, you can't replace, say, Excel with anonther spreadsheet. This would be the first advantage for a Unix desktop compared to windows and not a try to catch up after years of stagnation.
I disagree. I've read some stuff about the O3000 last week and it's maybe a nice box from the technical point of view, but as a ISP I wouldn't buy one.
You can get COD (capacity on demand) for most Sun servers now, not only the E10k. I haven't found something similar at SGIs website. They seem to have the more flexible hardware, but if your demands grow, you have to order bricks, wait for delivery, shut down and install them. With Sun, you have the hardware right there and only need an additional licence. IMHO Sun has the better solution here.
Repartitioning works with our E10k nice and dynamic. No downtimes. I thing this is a important feature, if you have to run 24/7.
I haven't ever worked with SGI boxes and this is just a quick shot. I think, the SGI boxes are nice, but not tailored to our demands. Maybe they are nice for universities or somewhere else, where availability isn't that important.
I think, this was an interesting interview. Even if much of Anders speak is anti-Java, he's more a technician than a manager or marketing man. (*) Most pasts sound logic for me, even if I see some things different, there's a smart guy behind C#.
But when it comes to standardisation, I'm very uncertain: even if C# and the core libraries are free available, what about COM? He says, that componenents are a core technology of C#, and components at Microsoft means COM. Period.
When Microsoft keeps control over such an important part, they keep control over the hole thing.
Just my 2 Pfennige;-)
Waldmeister
(*) This is different for Tony Goodhew, who clearly told the untruth about the write once, run anywhere of java. Nobody can misinterpret the statement from J. Gosling that much. Even a Microsoft manager.;-)
I don't make a translating here, but the main points are:
The didn't sue everyone, but only one company
They invested lots of money over many years in their software called SAMBA
They only want to prevent confiusion between their SAMBA and the SMB Samba.
They don't want to battle with the open source community, contrary they welcome this kind of development.
They stop suing the german company, who is only providing service for Samba, and trying to find a compromise solution with the international Samba team.
I think, their action was a little bit hasty and big companies tend to have a different view than the open source community, but they seem to be on the right way now.:-)
This is a big major pain in the *ss here in Germany and an obvious abuse of German law. They are going after the little guys by sending their bills to site owners and those who merely mirror a site.
Yes, it is. And they going after the little ones, because these don't have the money for a big legal battle. It's easier to pay the charges for the dissuasiveness (I hope this is the correct translation for Abmahnung) than to risk to pay the jurisdictional amount after (potentially) loosing at court.
As far as I remember past statements from intel about the IA64, it's targeted to the server market first. There is the biggest demand for a huge deal of data. 32 Bit should still be sufficient for the desktop for several years from now. Maybe this is similar to cars and trucks: you would never try to implant a truck engine into a car, would you?:-)
Crusoes dynamic translation could be a good way to get future 64 Bit software on simple (and portable) hardware. But the P6 architecure is portable, too, and it's far away from being power efficient.
I've installed Corel Linux on my Desktop PC several weeks ago and made several experiences:
The install is okay. I'm not religious about a graphical install, and, well, maybe people are used to that from Microsoft, but it adds some extra complexity to the installer. For example, you must have a supported graphics card.
The packages are mostly from Debian GNU/Linux 2.1 (aka slink), so you can get all those nice little programs, which are not on the CD, through the internet from your favourite Debian mirror. slink has been released some time ago, so some packages are not the newest version. Well, that's fine for me, because I like a stable system. But I know, most people prefer brand new software over stable and Corel has updated the kernel to 2.2.
Corel added some software to the desktop. The file manager has been stated before, another example is Corel Update, a KDE frontend for dpkg and apt-get. This is a very nice piece of software. Maybe I should try to extract the source and make a tarball for other distributions. (I think, Corel used a licence derieved from the MPL, so redistributing the source should be legal)
A major disadvantage is the packaging of KDE. Corel put everything in one huge package with lots of dependancies. Also, there a some bugs, which are not severe, but disturbing.
The hardware support out of the box - not only graphics, but also SCSI or sound cards - seems to be not that broads as with other distributions. But if you have good standard Linux hardware (like me, I had no problems here), Corel Linx is worth a look.
For me it works just fine. I'm using smit since 1994 and AIX 3.2.5 and didn't run into trouble. Modifying text files thought to be edited by humans with a stupid program isn't harmless. I remember problems with sam under HP-UX years ago. But that is nothing compared with the fun I had with SuSE Linux.:-(
YaST uses a text file with variables. But YaST creates the configuration files in/etc. I'm not that deep in this topic, but AFAIK you don't get a warning everytime an edited file will be overwritten. So you can deceide: you use YaST for everything or you disable it.
As you said, AIX has a binary object repository, but that is only used for AIX specific things like device configuration or package management. Almost everything else is done as usual with text files in/etc. And these text files are edited, not rewritten or yet created from some kind of database.
But - independently from using text or binary files - AIX has a complete set of homogeneous command line utilities for almost every administration task. And these utilities change existing text files, they don't overwrite them. Beside that, the set is much more complete than YaST is.
Okay, so you have a set of command line utilities with smit glueing everything together vs. a menu driven binary. I prefer the first approach. IBM did that years ago and it is still a very good concept.
Well, SuSE is in Germany very popular, but I don't like it. It's a good idea to hide system administration behind an easy interface, expecially for new users. But yast demands to be exclusive, it doesn't cooperate with other tools or manually edited config files. The complete configuration is stored in one file,/etc/rc.config and everything is regenerated from there. Poor one who edited something manually. OK, you get a warning if some files had been touched manually and you can disable rc.config, but I think that this is not the right way. There are other tools which do that job better. My favourite is smit for AIX. The interface just collects the data an passes it to a command line driven program and documents that in a logfile. So you have hidden everything below a nice interface, documented the changes and you can see what happens. There are some other points I dislike with SuSE Linux (not only since 6.0, which is already available in germany for some time), for example the braindead command completion in tcsh. Ever used that with "cp fubar.c fubar.c.bak"? But I think YaST is a good example how you get patronized by SuSE.
Slashdot Mirror
Java is almost a set of open standards. Language, Bytecode, Libraries, etc are well documented, and there is the JCP (Java Community Process), where everybody can participate in the evolution of java. (Okay, Sun is still more equal than everybody else is, because they have a veto right.) I don't know, why Sun withdraw the standard requests, but in my eyes they look, like trying to keep Java unpolluted and unforked in the first place.
If you take a look at the enterprise sector, Sun does only provide a spec and leaves the implementation to others like BEA or Orion.
As far as I understand, the XPU128 itself doesn't run Linux:
The GNU C Compiler and related tools allow programming of sequential parts of algorithms for the Risc processor. The API for Communication with Host supports NT and Linux operating systems.
Parallel parts of the algorithms are programmed in the high level language LELA or Assembler language. A communication class library provides simple integration of C programs and XPU programs.
Linux is only running at the risc processor, used as an interface between the XP128 and the PC.
I don't believe, you can translate C to efficient parallel code. applications and algorithms have been optimized for the common sequential architectures over the last 20, 30 years. Switching to massive parallel devices will demand new software.
Yes, Sun seems to be a little bit boring at first, but I think this is exaclty what the industry wants. And no changing roadmaps twice a year. ;-)
Itanium: Probable winnerI agree, too. Not because the're good, but because the're from Intel.
Sledgehammer: Expected no-showWe will see. But I don't expect gread marketshare. Well, maybe, if AMD get's more industry support for their x86 stuff in the meantime.
Alpha: Dying, Also RanYes, Compaq seems to be quite happy with Intel chips, no make competition inhouse. ;-)
MIPS: Out of production, Also RanStill in production, but IMHO only with a vague future. SGI is bouncing from MIPS to Intel (NT and Linux workstations) and back to MIPS (Origin 3000). Looks for me like a looser in the last years.
HPPA: Possible contender if anyone knew about 'em. Nope. Intel developed the Itanium together with HP, and even if HP has extended its HP-HP roadmap, the future at HP is IA64. For example, the new SuperDome is already prepared for IA64.It is not well defined, what an operating system is or has to be. The most common definition could be an interface between the user(s) and the hardware. While we don't have to argue about the low level funktions provided by the kernel, there is much room to discuss wether the user level applications, packaged together with the kernel, are part of the os, too.
If you do this (and the original author does it), I don't see why you should differ between the "classic" Unix utilities and the graphical ones running under X11.
No OS is just the same as some years ago. Neither MacOS nor Windows had a browser included five years ago. Now they have. Browsers are included in almost any Unix software box, you can buy today. Why count with differend measures?
Unix is, in costrast to Windows or MacOS, much more modular and versatile. The core parts got some additional layers. X11 for example is such a layer and the foundation for graphical applications. Adding window management and an email client is the purpose of destop environments like Gnome.
Unix has also been a multiuser/multitasking OS for a very, very long time. So it has a strong position in the server market. There you don't need (primary) an web client, but a web server. (Linux, as a flavor of the Unix idea, is so versatile, it run's almost everywhere from a IBM mainframe to a wrist watch)
Unix has, and this is the main point, in which I could agree with the original author, a weak history on the desktop. The desktop environments and the applications are not as mature as they are for Windows or MacOS. But the topic was not about the quality of the Unix desktops, but about Unix not being a OS at all.
And I think, you need a pretty weird definition of an OS to declare Unix as being not being one.
It wouldn't just help the Apps company, it's the only way to get Office ported.
If the integration between os and app is so tight, as everybody says, they (the apps company) will have a very hard time to get ports done without the source (and the most interesting details) of the OS libraries like COM, OLE, etc).
I remember times, when IBM sold their own version of Netware. The box looked exactly the same than the original one from Novell, but it was not red, it was blue. :-)
I think, it's not another Linux UI, but Sun and others joining their forces to push Gnome. This is remarkable in my eyes because
I think, this is very good news.
I disagree. I've read some stuff about the O3000 last week and it's maybe a nice box from the technical point of view, but as a ISP I wouldn't buy one.
I haven't ever worked with SGI boxes and this is just a quick shot. I think, the SGI boxes are nice, but not tailored to our demands. Maybe they are nice for universities or somewhere else, where availability isn't that important.
Oh, dont worry. Wait for additional features and you will get the bugs for free. ;-)
I think, this was an interesting interview. Even if much of Anders speak is anti-Java, he's more a technician than a manager or marketing man. (*) Most pasts sound logic for me, even if I see some things different, there's a smart guy behind C#.
;-)
;-)
But when it comes to standardisation, I'm very uncertain: even if C# and the core libraries are free available, what about COM? He says, that componenents are a core technology of C#, and components at Microsoft means COM. Period.
When Microsoft keeps control over such an important part, they keep control over the hole thing.
Just my 2 Pfennige
Waldmeister
(*) This is different for Tony Goodhew, who clearly told the untruth about the write once, run anywhere of java. Nobody can misinterpret the statement from J. Gosling that much. Even a Microsoft manager.
The CMG has made a press release. You can read it at http://www.cmg.de/aktuelles/aktuell. htm#P7_138.
I don't make a translating here, but the main points are:
I think, their action was a little bit hasty and big companies tend to have a different view than the open source community, but they seem to be on the right way now. :-)
cu
Waldmeister
This is a big major pain in the *ss here in Germany and an obvious abuse of German law. They are going after the little guys by sending their bills to site owners and those who merely mirror a site.
Yes, it is. And they going after the little ones, because these don't have the money for a big legal battle. It's easier to pay the charges for the dissuasiveness (I hope this is the correct translation for Abmahnung) than to risk to pay the jurisdictional amount after (potentially) loosing at court.
Nope, they mean _all_ revenue. I've read this in an article about a similar case against DaimlerChrysler.
As far as I remember past statements from intel about the IA64, it's targeted to the server market first. There is the biggest demand for a huge deal of data. 32 Bit should still be sufficient for the desktop for several years from now. Maybe this is similar to cars and trucks: you would never try to implant a truck engine into a car, would you? :-)
Crusoes dynamic translation could be a good way to get future 64 Bit software on simple (and portable) hardware. But the P6 architecure is portable, too, and it's far away from being power efficient.
I've installed Corel Linux on my Desktop PC several weeks ago and made several experiences:
The install is okay. I'm not religious about a graphical install, and, well, maybe people are used to that from Microsoft, but it adds some extra complexity to the installer. For example, you must have a supported graphics card.
The packages are mostly from Debian GNU/Linux 2.1 (aka slink), so you can get all those nice little programs, which are not on the CD, through the internet from your favourite Debian mirror. slink has been released some time ago, so some packages are not the newest version. Well, that's fine for me, because I like a stable system. But I know, most people prefer brand new software over stable and Corel has updated the kernel to 2.2.
Corel added some software to the desktop. The file manager has been stated before, another example is Corel Update, a KDE frontend for dpkg and apt-get. This is a very nice piece of software. Maybe I should try to extract the source and make a tarball for other distributions. (I think, Corel used a licence derieved from the MPL, so redistributing the source should be legal)
A major disadvantage is the packaging of KDE. Corel put everything in one huge package with lots of dependancies. Also, there a some bugs, which are not severe, but disturbing.
The hardware support out of the box - not only graphics, but also SCSI or sound cards - seems to be not that broads as with other distributions. But if you have good standard Linux hardware (like me, I had no problems here), Corel Linx is worth a look.
cu
Waldmeister
For me it works just fine. I'm using smit since 1994 and AIX 3.2.5 and didn't run into trouble. :-(
Modifying text files thought to be edited by humans with a stupid program isn't harmless. I remember problems with sam under HP-UX years ago. But that is nothing compared with the fun I had with SuSE Linux.
Yes and no.
/etc. I'm not that deep in this topic, but AFAIK you don't get a warning everytime an edited file will be overwritten. So you can deceide: you use YaST for everything or you disable it.
/etc. And these text files are edited, not rewritten or yet created from some kind of database.
YaST uses a text file with variables. But YaST creates the configuration files in
As you said, AIX has a binary object repository, but that is only used for AIX specific things like device configuration or package management. Almost everything else is done as usual with text files in
But - independently from using text or binary files - AIX has a complete set of homogeneous command line utilities for almost every administration task. And these utilities change existing text files, they don't overwrite them. Beside that, the set is much more complete than YaST is.
Okay, so you have a set of command line utilities with smit glueing everything together vs. a menu driven binary. I prefer the first approach. IBM did that years ago and it is still a very good concept.
Well, SuSE is in Germany very popular, but I don't like it. /etc/rc.config and everything is regenerated from there. Poor one who edited something manually. OK, you get a warning if some files had been touched manually and you can disable rc.config, but I think that this is not the right way.
It's a good idea to hide system administration behind an easy interface, expecially for new users. But yast demands to be exclusive, it doesn't cooperate with other tools or manually edited config files. The complete configuration is stored in one file,
There are other tools which do that job better. My favourite is smit for AIX. The interface just collects the data an passes it to a command line driven program and documents that in a logfile. So you have hidden everything below a nice interface, documented the changes and you can see what happens.
There are some other points I dislike with SuSE Linux (not only since 6.0, which is already available in germany for some time), for example the braindead command completion in tcsh. Ever used that with "cp fubar.c fubar.c.bak"? But I think YaST is a good example how you get patronized by SuSE.