Well, from my point of view it would make sense. I am far more productive when using Java, and my code is more stable (and I am not the only one who experiences this). Java/J++ is a nice comprimise between high-level programming like VB and low-level programming like C(++). AFAIK most Windows APIs have been ported to J++, you can even write Services. MS's COMJava/J++ integration looked real clean the last time a took a look, and if I was going to write a Windows application I am sure that I would consider using VisualJ++ (and probably end up using it if the needed APIs are available and it is fast enough for the task). Not as an alternative to Java, but as an alternative to VB, VC and Delphi. Besides that J++/Java is a nice little language, the chances that MS sees are probably that Java GUIs feel real bad. Try to use any Java application in a Windows environment, and it will always look like an alien and wont support things like OLE. So when a developer wrote a program in Java and the customers start complaining about its Windows integration (and they will!), many developers wont resist the temptation to please 90% of their market. And Microsoft reached its goal: one program more for Windows, one less for the competition.
It looks like they cloned Microsofts J++ APIs, too. The reason why MS funded them is probably because they might have to remove all Sun code from their VM as a result of the MS-Sun-Java trial. I still dont care about Kaffe. The best way to ensure compatibility is to use Sun's VM or a VM based on Sun's (like the IBM VM). And if I need a open-sourced VM, I'll take Japhar or the one by the Mozilla guys. I cant see any need for Kaffe, a slow semi-proprietary Java implementation. Perhaps that's why they sold out - because MS was the only interested company...
I only have a German price list, but there shouldnt be much of a difference, besides that german prices include VAT (16%).. E10000 price list in Euro (1 Euro = 1.05 $) E10000 Base (Case, without any boards, CPUs..): 309700 EUR Control board (/w Ethernet Hub): 28200 EUR System board (for 4 CPUs): 91500 EUR UltraSparcII 250Mhz 4MB Cache: 16300 EUR UltraSparcII 400Mhz 4MB Cache: 23100 EUR PCI board (2 slots): 14100 EUR I/O board (4 SBUS slots): 10600 EUR SSP (Service Processor, Ultra 5/w monitor): 148-- EUR and so on...
Yes, the system boards are motherboards with up to 4 UltraSparcs on each. Only a little bit more expensive than normal motherboards, something like 100000$ each (without CPUs, RAM and I/O boards, of course)... i still want my personal E10000.
Journaling file systems maintain a kind of changelog, similar to the transaction logs of databases. Using this they can always recover from crashs and you are sure that you can never lose data. See the XFS whitepaper on the SGI site for more information, someone posted the link in the comments (got a 5 score so you cant miss it).
Irix seems to be very scalable. For example, their flagship Origin 2000 has up to 128 CPUs, and even on the CRAY supercomputers Irix is an option. In other words, SGI has a lot of experience in large systems, and this is what Linux needs to get into the coporate market. Beside that, supporting Linux would mean that they port OpenGL, OpenInventor and all the other nice graphic APIs to Linux. And as SGI owns Alias|Wavefront, Linux would probably get high-end rendering and design software as well. Another interresting thing to note is that AFAIK some time ago SGI announced to abandon the MIPS cpus in the long term (they already sold MIPS) and to switch to intel's IA-64 architecture. The article said that Linux would become their only Unix-like OS on intels, so this would mean that Linux would be their primary operating system, unless they really want to use NT on their high-end servers. Whoa..
The prices for the current Ultra line are rediculus. (My favorite in the price list is the "multimedia kit": 1500$ for a PCI-based video camera) So Solarix x86 is a nice alternative if you want the buzzwords (scalability, reliability) or need software that is only available for Solaris. For example, the Netscape Enterprise server is not available for Linux, and the JavaVMs for Linux are very weak compared to the Solaris versions.
We dont how much of the Playstation2's 6.4 gflops are pure hype. AFAIK the 'emotion engine' is based on the MIPS RX000 series, and I don't think that Sony is able to produce RX000-CPUs that are so much faster than MIPS's or SGI's... Perhaps they have geometry processors or similar custom logic that achieves this performance, but certainly not in a general purpose CPU.
Objectivity has released a linux version of Objectivity/DB for Linux last year. Objectivity is an very scalable, page-based OODBMS (and normal commercial software). So far I have only used the Solaris and NT versions and the Java binding. The Java bindungs on Linux are not finished yet. Objectivity works pretty well with JDK 1.2, but with JDK 1.1 we had a lot of crashes that suddely disappeared when we switched to 1.2. Probably not Objectivitys fault, but made it almost unusable.
Free (some more, some less) CORBA implementations for C++: MICO, ORBit, ILU For Java: JacORB, ILU, JavaIDL (included in JDK 1.2) For Python: FNORB, ILU, SYLO
CORBA has always suffered from the Lowest-Common-Denominator problem. Using CORBA for large-scale applications is painful, as there are many OO features it does not support, for example method overloading. But perhaps CORBAs (2.0) biggest problem is that you cannot transfer objects, only (C-like) structs. This makes it really hard to write fast distributed programs in a object oriented way. Most times you have the choice between an extremely bad performance (because you use too many calls) or losing almost any platform-neutrality (for example by seralizing objects and sending them as byte arrays). CORBAs main probelm is that is was made to support every architecture, every language. Unfortunately, if you design a standard so you can support Fortran and Cobol, you have to give up many modern features and make it almost unusable.
I can understand the author's view, but I don't think that it is fair to treat users who are less technical this way. I think you should differentiate between two kind of users. The first group is like most of us, people who spend a huge amount of time in front of their computers. The other group are people who only use computers ocassionally to do a specific things, write a letter for example. You cannot expect from the second group that they will learn how to configure Linux, they don't care for all the technical stuff beneath their word processor, they just want to write the letter. I can understand this, when I use a toaster I don't have to know how it works, and they want the same thing when writing letters. Of course, I will read the toasters manual before I call customer support, but that's a different problem I would not say that CLI tools are always superior to GUI tools. This depends on what you do. If you need a tool every day, a CLI tool will be better (at least in most cases). But for things that I do very rarely, for example install an ISDN router, I dont want to spend 10 hours of reading manuals before I am able to install it. I just want it to work as fast as possible. If I can do it in 10 minutes with a simple GUI and without reading manuals, I will do this. Unless you use a tool frequently, CLI doesn't pay off because you spend more time learning the use of the program than actually using it.
This means that either SAP doesn't make much use of relational functionality and they probably are actually caching (or copying) the DB on their middle tier (which is true AFAIK). (...) R/3 seems to be still from the age before relational DBMS. This could also mean that they have a nice object-oriented database-scheme and map this model onto a relational database to use the rdbm's transaction model and (perhaps more important) to please IT managers who want their data on "a reliable database". OODBMS->RDBMS mappings have the same characteristics.
The reason for this investment could be that SAP is afraid of Microsoft. Microsoft has shown several times that they are able to capture almost every market. And ERP seems to be a very lucrative business so they might do the same thing with SAP that they did with Oracle (SQL Server), Netscape (Internet Explorer & IIS), to Lotus (Exchange) and all the other markets. They capture huge a amount of market shares just by their presence.
Executable documents are just plain wrong. No, they can be made secure and sometimes they can even be useful (especially when you use interactive documents instead of regular applications, for example in a list that checks the consistency of new entries). The problem is that it is extremely difficult to make this secure, and it looks like Microsoft is not putting much effort in making VBA programs embedded in Office documents very secure. Netscape is quite successful at making HTML+Java/-script quite safe. They are not perfect, as the technology is evolving much too quick, but it is definately a proof-of-concept. (The above paragraph should not leave you with the impression that using Javascript on a HTML page is a good thing - of course you should use LML and Javascript is evil).
You can vote for PNG support in Java on Sun's "Bug parade", as it does not only allow you to vote for the removal of bugs, but also for new features. PNG support is currently the second most requested feature (first is Linux support). If you want to vote for PNG, you have to join the Java Developer Connection (free), the URL of the bug parade is: http://developer.java. sun.com/developer/bugParade/index.html. The PNG feature has the Bug id 4101708.
I read it, but I don't understand it. It says that Java is bloated?! Definately not, I MISS features like templates/generic objects and multiple inheritance, multi-methods... The lack of features in Java makes programming unneccessary complicated (at least for people who use it >40hrs/week, if you want to write 1000 line programs your needs will be different). I don't have the any practical experience with Dylan, but after reading the Dylan Language I had the impression that it was not very suited for designing big programs. I don't like its hybrid-oop concept (seperating methods from objects), it is even worse than Java's hybrid concept, and I cannot imagine how to design programs in a object-oriented way. It feels more like a functional language with object-oriented structures, and I don't like functional programming.
Java is almost perfect for intelligent agents, it is incredible easy to do things like serialization or remote communication (rmi) with Java. You can do this with Perl or C++, but is it complicated and not integrated in the basic libraries. Also, most agent frameworks are written for Java (Voyager is a very good example, www.objectspace.com) and Java's platform independence really makes sense for agents. Perhaps most important, agents fit very, very well into the whole OOP idea (one agent = one object), and Perl's OOP capabilities suck by the way.
SableCC is a very good alternative to most well-known Java Compiler-Compilers and it is LGPL'd. It uses LALR, is very easy to use and generates AST and so-called Tree-Walker-classes instead of putting action code in the grammar file. I haven't seen the Jikes thing yet, but I can highly recommend SableCC, it is superior to any compiler generator I've seen. You can get it at http://www.sable.mcgill.ca.
I cannot see the impact other people do. A PSN is not worse than the 48 Bit Ethernet adresses that every network cards has. When you are using HTTP, no one can read your PSN (or Ethernet address) unless you allow him to execute native code on your computer. Before a browser sends your PSN to a server, it has to ask you. It's the same thing like sending your email address. What's so great about PSN is that is prohibits software theft. I don't like closed-source software, but I also hate pirated software, because it really hurts software developers (and I am one of them). And when people suddenly really have to pay for their software, they will start to use more free software.
Slashdot Mirror
Well, from my point of view it would make sense. I am far more productive when using Java, and my code is more stable (and I am not the only one who experiences this). Java/J++ is a nice comprimise between high-level programming like VB and low-level programming like C(++).
AFAIK most Windows APIs have been ported to J++, you can even write Services. MS's COMJava/J++ integration looked real clean the last time a took a look, and if I was going to write a Windows application I am sure that I would consider using VisualJ++ (and probably end up using it if the needed APIs are available and it is fast enough for the task). Not as an alternative to Java, but as an alternative to VB, VC and Delphi.
Besides that J++/Java is a nice little language, the chances that MS sees are probably that Java GUIs feel real bad. Try to use any Java application in a Windows environment, and it will always look like an alien and wont support things like OLE. So when a developer wrote a program in Java and the customers start complaining about its Windows integration (and they will!), many developers wont resist the temptation to please 90% of their market. And Microsoft reached its goal: one program more for Windows, one less for the competition.
It looks like they cloned Microsofts J++ APIs, too. The reason why MS funded them is probably because they might have to remove all Sun code from their VM as a result of the MS-Sun-Java trial.
I still dont care about Kaffe. The best way to ensure compatibility is to use Sun's VM or a VM based on Sun's (like the IBM VM). And if I need a open-sourced VM, I'll take Japhar or the one by the Mozilla guys. I cant see any need for Kaffe, a slow semi-proprietary Java implementation. Perhaps that's why they sold out - because MS was the only interested company...
I only have a German price list, but there shouldnt be much of a difference, besides that german prices include VAT (16%).. /w monitor): 148-- EUR
E10000 price list in Euro (1 Euro = 1.05 $)
E10000 Base (Case, without any boards, CPUs..): 309700 EUR
Control board (/w Ethernet Hub): 28200 EUR
System board (for 4 CPUs): 91500 EUR
UltraSparcII 250Mhz 4MB Cache: 16300 EUR
UltraSparcII 400Mhz 4MB Cache: 23100 EUR
PCI board (2 slots): 14100 EUR
I/O board (4 SBUS slots): 10600 EUR
SSP (Service Processor, Ultra 5
and so on...
Yes, the system boards are motherboards with up to 4 UltraSparcs on each. Only a little bit more expensive than normal motherboards, something like 100000$ each (without CPUs, RAM and I/O boards, of course)... i still want my personal E10000.
Journaling file systems maintain a kind of changelog, similar to the transaction logs of databases. Using this they can always recover from crashs and you are sure that you can never lose data.
See the XFS whitepaper on the SGI site for more information, someone posted the link in the comments (got a 5 score so you cant miss it).
Irix seems to be very scalable. For example, their flagship Origin 2000 has up to 128 CPUs, and even on the CRAY supercomputers Irix is an option.
In other words, SGI has a lot of experience in large systems, and this is what Linux needs to get into the coporate market.
Beside that, supporting Linux would mean that they port OpenGL, OpenInventor and all the other nice graphic APIs to Linux. And as SGI owns Alias|Wavefront, Linux would probably get high-end rendering and design software as well.
Another interresting thing to note is that AFAIK some time ago SGI announced to abandon the MIPS cpus in the long term (they already sold MIPS) and to switch to intel's IA-64 architecture. The article said that Linux would become their only Unix-like OS on intels, so this would mean that Linux would be their primary operating system, unless they really want to use NT on their high-end servers. Whoa..
The prices for the current Ultra line are rediculus. (My favorite in the price list is the "multimedia kit": 1500$ for a PCI-based video camera)
So Solarix x86 is a nice alternative if you want the buzzwords (scalability, reliability) or need software that is only available for Solaris. For example, the Netscape Enterprise server is not available for Linux, and the JavaVMs for Linux are very weak compared to the Solaris versions.
We dont how much of the Playstation2's 6.4 gflops are pure hype. AFAIK the 'emotion engine' is based on the MIPS RX000 series, and I don't think that Sony is able to produce RX000-CPUs that are so much faster than MIPS's or SGI's...
Perhaps they have geometry processors or similar custom logic that achieves this performance, but certainly not in a general purpose CPU.
Cool, but where can I buy them (in Germany)?
Objectivity has released a linux version of Objectivity/DB for Linux last year. Objectivity is an very scalable, page-based OODBMS (and normal commercial software).
So far I have only used the Solaris and NT versions and the Java binding. The Java bindungs on Linux are not finished yet. Objectivity works pretty well with JDK 1.2, but with JDK 1.1 we had a lot of crashes that suddely disappeared when we switched to 1.2. Probably not Objectivitys fault, but made it almost unusable.
Send this article to mass media (TV, newspapers) in the hope, that somebody recognizes the problem...
Free (some more, some less) CORBA implementations for C++: MICO, ORBit, ILU For Java: JacORB, ILU, JavaIDL (included in JDK 1.2) For Python: FNORB, ILU, SYLO
CORBA has always suffered from the Lowest-Common-Denominator problem. Using CORBA for large-scale applications is painful, as there are many OO features it does not support, for example method overloading. But perhaps CORBAs (2.0) biggest problem is that you cannot transfer objects, only (C-like) structs. This makes it really hard to write fast distributed programs in a object oriented way. Most times you have the choice between an extremely bad performance (because you use too many calls) or losing almost any platform-neutrality (for example by seralizing objects and sending them as byte arrays). CORBAs main probelm is that is was made to support every architecture, every language. Unfortunately, if you design a standard so you can support Fortran and Cobol, you have to give up many modern features and make it almost unusable.
Who is San Mehat? Sorry, but I have never heard this name before...
I can understand the author's view, but I don't think that it is fair to treat users who are less technical this way.
I think you should differentiate between two kind of users. The first group is like most of us, people who spend a huge amount of time in front of their computers. The other group are people who only use computers ocassionally to do a specific things, write a letter for example. You cannot expect from the second group that they will learn how to configure Linux, they don't care for all the technical stuff beneath their word processor, they just want to write the letter. I can understand this, when I use a toaster I don't have to know how it works, and they want the same thing when writing letters. Of course, I will read the toasters manual before I call customer support, but that's a different problem
I would not say that CLI tools are always superior to GUI tools. This depends on what you do. If you need a tool every day, a CLI tool will be better (at least in most cases). But for things that I do very rarely, for example install an ISDN router, I dont want to spend 10 hours of reading manuals before I am able to install it. I just want it to work as fast as possible. If I can do it in 10 minutes with a simple GUI and without reading manuals, I will do this. Unless you use a tool frequently, CLI doesn't pay off because you spend more time learning the use of the program than actually using it.
This means that either SAP doesn't make much use of relational functionality and they probably are actually caching (or copying) the DB on their middle tier (which is true AFAIK).
(...)
R/3 seems to be still from the age before relational DBMS.
This could also mean that they have a nice object-oriented database-scheme and map this model onto a relational database to use the rdbm's transaction model and (perhaps more important) to please IT managers who want their data on "a reliable database". OODBMS->RDBMS mappings have the same characteristics.
The reason for this investment could be that SAP is afraid of Microsoft. Microsoft has shown several times that they are able to capture almost every market. And ERP seems to be a very lucrative business so they might do the same thing with SAP that they did with Oracle (SQL Server), Netscape (Internet Explorer & IIS), to Lotus (Exchange) and all the other markets. They capture huge a amount of market shares just by their presence.
Executable documents are just plain wrong.
No, they can be made secure and sometimes they can even be useful (especially when you use interactive documents instead of regular applications, for example in a list that checks the consistency of new entries). The problem is that it is extremely difficult to make this secure, and it looks like Microsoft is not putting much effort in making VBA programs embedded in Office documents very secure.
Netscape is quite successful at making HTML+Java/-script quite safe. They are not perfect, as the technology is evolving much too quick, but it is definately a proof-of-concept.
(The above paragraph should not leave you with the impression that using Javascript on a HTML page is a good thing - of course you should use LML and Javascript is evil).
You can vote for PNG support in Java on Sun's "Bug parade", as it does not only allow you to vote for the removal of bugs, but also for new features.
PNG support is currently the second most requested feature (first is Linux support).
If you want to vote for PNG, you have to join the Java Developer Connection (free), the URL of the bug parade is:
http://developer.java. sun.com/developer/bugParade/index.html. The PNG feature has the Bug id 4101708.
I read it, but I don't understand it. It says that Java is bloated?! Definately not, I MISS features like templates/generic objects and multiple inheritance, multi-methods...
The lack of features in Java makes programming unneccessary complicated (at least for people who use it >40hrs/week, if you want to write 1000 line programs your needs will be different).
I don't have the any practical experience with Dylan, but after reading the Dylan Language I had the impression that it was not very suited for designing big programs. I don't like its hybrid-oop concept (seperating methods from objects), it is even worse than Java's hybrid concept, and I cannot imagine how to design programs in a object-oriented way. It feels more like a functional language with object-oriented structures, and I don't like functional programming.
Yes, but these Wrapper Objects make everything complicated. Smalltalk is much more comfortable without losing speed in this aspect.
Java is almost perfect for intelligent agents, it is incredible easy to do things like serialization or remote communication (rmi) with Java. You can do this with Perl or C++, but is it complicated and not integrated in the basic libraries. Also, most agent frameworks are written for Java (Voyager is a very good example, www.objectspace.com) and Java's platform independence really makes sense for agents.
Perhaps most important, agents fit very, very well into the whole OOP idea (one agent = one object), and Perl's OOP capabilities suck by the way.
don't forget Revenge of the Nerds...
SableCC is a very good alternative to most well-known Java Compiler-Compilers and it is LGPL'd. It uses LALR, is very easy to use and generates AST and so-called Tree-Walker-classes instead of putting action code in the grammar file. I haven't seen the Jikes thing yet, but I can highly recommend SableCC, it is superior to any compiler generator I've seen.
You can get it at http://www.sable.mcgill.ca.
I cannot see the impact other people do. A PSN is not worse than the 48 Bit Ethernet adresses that every network cards has. When you are using HTTP, no one can read your PSN (or Ethernet address) unless you allow him to execute native code on your computer. Before a browser sends your PSN to a server, it has to ask you. It's the same thing like sending your email address.
What's so great about PSN is that is prohibits software theft. I don't like closed-source software, but I also hate pirated software, because it really hurts software developers (and I am one of them). And when people suddenly really have to pay for their software, they will start to use more free software.