For classical music the answer is definitely "yes". It's likely that people in the 23rd Century will be very interested in 21st Century performance styles and premieres of new pieces for example.
I agree with these comments re:libraries. A great language is nothing without class libraries that reflect the reality of the work you want to do; the problems you need to fix.
I'd like to add that I one approach is simply to avoid languages where no ecosystem exists around the language. A really balanced language choice will have an ecosystem that delivers across the board from IDEs, build tooling, great runtimes and choice of runtimes and good RAS / PD tooling for when things do wrong. That's one of the reasons Java is in widespread use - it can tick all of these boxes and is balanced between these develop - deploy - debug areas.
the database access functionality is built in to 5.1 and above. I wasn't sure from what you said whether what you want to do is possible with this approach, but it might be worth considering ?
I doubt if many people outside the US go to the NYT webpage or editorial.
I don't know how you can possibly say this. Speaking as a Brit I regularly read the NYT on the web and read the op-ed pages that will now be denied to me. I'm really surprised at this decision, since it will probably not generate much money, but will cut off a lot of interest abroad. If you compare this with the Guardian, they are conciously appealing via the web to a global progressive audience for news, and I'd have thought that was a better model for the NYT as well. Lame.
Many of the ancient IBM mainframes are being replaced with servers running Java applications on them
Businesses don't have to do this - they can deploy J2EE apps directly on your so called ancient IBM mainframes. The same Java and WebSphere code runs directly on big blue iron.
I can see a lot in your arguments, which is obviously based on experience at a high level in IBM and I respect that. My experience is probably lower down;-), but it does give me the benefit of seeing a lot of things close too. I think your view of IBM has grains of truth, but it really does miss the complexity inherent in a big organization. I worked within IBM in the organization that worked on JDKs. I saw competent technical leadership - managers with PhDs and patents to their name - leading large cross country teams of good developers to create JDKs across a much wider range of platforms than Sun was willing to do. I saw the addition of good technologies like concurrent GC, great JITs, the persistent resuable JVM on zSeries. In other words - not just a port of the Sun codebase.
This is all fairly low level - these things are then embedded in IBM middleware which is the reason for doing it - and given away for free to whoever wants them. This is still techincal and not contracted out to some OSS organization. I know this is unlikely to change your view, but in my experience is that IBM is no shell.
Well that's a bracing view of things for sure. I suppose where we differ is that you are setting the benchmark much higher for what constitutes "technology" or a "technical success". I'd argue that a lot of great work is done in turning ideas "not invented here" into code and products that do actually meet the needs, requirements and expectations of businesses (which after all is still the main focus of IBM).
That's not going to win a Nobel prize, but it's still technical in my book.
Where's your evidence for this assertion that IBM is weaker technically now than then ? Since those days IBM has jettisoned old and unloved technologies like SNA and has got some pride and technical leadership back, often by working with people rather then bullying them. It's actually delivered and extending on good ideas that were "not invented here" (like J2SE and J2EE) where it has a very successful app server in WebSphere. Look at Eclipse - becoming a defacto tools environment which will benefit a lot of people. Mainframe tech. wrenched in to the 21st Century - from 3270/ISPF/CICS etc. to J2EE apps on a nice unix base in 8 years of continuous development. Hardware: POWER, IBM chips power Macs (G5), the Gamecube. The NG consoles will all have IBM technology inside. Too early to tell about Cell, but it's hardly an example of a company on the slide. Still the best at virtualization and partitioning. Still top of the patent league (for what that's worth I agree, but still a metric). The Research division continues to innovate, and is no worse then it was under Akers - in fact much more focused on delivery and real world problems. What about the Community Grid ? I could go on and on...
Yep. Also, the compiler can encapsulate the accumulated knowledge of the ages about the target platform. There is no way on God's earth any normal person will know enough about the instruction set of the target platform to beat that accumulated knowledge. Anyone who does should be writing compilers, not application progams. Or maybe taking up another interest in life.
"I was astonished when I saw how beautiful Birmingham was. The buildings, the art, the use of the water. It is an extraordinary jewel of a city.... I was bowled over when I was there. It is quite wonderful." Bill Clinton, The Times, October 2nd, 2002
yours, a slashdot reading Brummie.
PS. Slagging off Birmingham is about the oldest meme you could pick up. Congrats on your originality.
But cash alone is not enough for survival. IBM had to change itself significantly to make it through the bad years. That took new leadership at the top - this I think would be firm evidence the MS was under pressure.
You could just ship a JRE at a specific version and service level with your code ? This is what some middleware vendors do at the server side, so they are shipping their code with the JVM and Class library code they have tested against.
I think this is a very good point. How about Linux on an IBM zSeries box ? This would get plus points from management for scalability, managability and decent service, all on reliable hardware. I think there are three distros that are packaged for the "Iron Penguin" : Suse, Redhat or Turbolinux.
Yes, I see what you are saying, but I think it underestimates the extent to which what you might call the "corporate model" for developing software can deliver software people are willing to pay for. If you take Middleware today for example, it's undeniable that despite great alternatives (JBoss, LAMP things, even Zope) IBM makes a fortune in software revenues from WebSphere.
I honestly don't think this is right. For sure we can all forsee a future where OSes and large areas of middleware are commoditized and OSS versions are dominant, but I don't for a moment think IBM's intention is to wind up its Software Group, with its n labs round the world and strong links to research. I think the view rather is (a) we can remain ahead of the curve, at least enough that people will want to part with hard cash to buy products that we ship and (b) there are still things you can do with 1000s of internally distributed developers (coupled with good management and good project managers) that you can't do as easily in typical Open Source projects based outside of large companies. I'm thinking in particular of Servicing many releases concurrently on multiple platforms, or having high focus on RAS (Reliability, Availibility and Serviceability), or having specialist performance groups keeping an eye on product performance. It's these "nuts and bolts" aspects of software development that big companies like to know about and l33t hackers are not that concerned about (forgive the generalization).
Ouch, cheap shot:-) IBM produces many 100s of Redbooks a year and there is a standard toolset to maintain consistency. However, the source material given by the authors to the editors could have been done on any platform, in most any format...
I think Sony are getting gradually closer, and Apple must be at least a bit worried. Let's imagine a world where Sony (finally) drop ATRAC3+ only, and their Hi-MD devices and new HDD devices support MP3 natively and ATRAC3+ at whatever bitrate you desire. Given that the low end Hi-MD ships at 140UKP Sony could wipe out the Compact Flash competition in months with their cheap 1GB removable media. Much higher capacities are in the offing for MO as well. They are the only other vendor to offer integration with a online music store, which is a serious advantage over vendors like iRiver (despite their excellent hardware). The Connect store may suck now in comparison with ITMS, but heh, it's software, it can change quickly.
Other: not clearly defined & buzzword like Java Platform, EJB etc EJB isn't a buzzword, it's defined in the J2EE Spec.
You missed out the JDK. No implementation would be complete without the basic tooling e.g. javac, javap etc.
Because currently there are no Java VM implentation which is as complete as Sun Java VM That's wrong. The IBM JDKs pass the JCK and have VM implementations just as complete (but different) to the Sun JDKs.
Re: the end of the article. I'm no historian, but I'm certain Condoms go back at least 200 years, so a scientist who mentioned this in 1800 wouldn't haven't been thought absurd.
This isn't as high risk as it sounds, for at least two reasons. The first is that IBM has lots of people with 50 pound brains employed to think about what's next, rather then what's here already. They already realize that OSS can make a commodity out of what was previously sold for a profit. Since it's happened for the OS, it will happen for middleware as well. Secondly, the organizations they sell to aren't impressed one way or another by some standlone discrete piece of software like MySQL or JBoss. They want to know: where's the tooling that integrates closely with the runtime ? What's your 5 year development plan for this product ? How does this help me integrate different legacy systems ? Are you playing with industry standard bodies ? Lots of different questions that a disperate group of OSS developers may or may not be able to answer.
Slashdot Mirror
"Eat Food. Mainly Plants. Not too much"
For classical music the answer is definitely "yes". It's likely that people in the 23rd Century will be very interested in 21st Century performance styles and premieres of new pieces for example.
I agree with these comments re:libraries. A great language is nothing without class libraries that reflect the reality of the work you want to do; the problems you need to fix.
I'd like to add that I one approach is simply to avoid languages where no ecosystem exists around the language. A really balanced language choice will have an ecosystem that delivers across the board from IDEs, build tooling, great runtimes and choice of runtimes and good RAS / PD tooling for when things do wrong. That's one of the reasons Java is in widespread use - it can tick all of these boxes and is balanced between these develop - deploy - debug areas.
If something is dearer it's more expensive than something else, so the opposite of cheaper. It's standard UK English as well.
What don't you try Mathematica on OS X ? According to this link
http://www.wolfram.com/news/dbak.html
the database access functionality is built in to 5.1 and above. I wasn't sure from what you said whether what you want to do is possible with this approach, but it might be worth considering ?
I doubt if many people outside the US go to the NYT webpage or editorial.
I don't know how you can possibly say this. Speaking as a Brit I regularly read the NYT on the web and read the op-ed pages that will now be denied to me. I'm really surprised at this decision, since it will probably not generate much money, but will cut off a lot of interest abroad.
If you compare this with the Guardian, they are conciously appealing via the web to a global progressive audience for news, and I'd have thought that was a better model for the NYT as well. Lame.
Many of the ancient IBM mainframes are being replaced with servers running Java applications on them
Businesses don't have to do this - they can deploy J2EE apps directly on your so called ancient IBM mainframes. The same Java and WebSphere code runs directly on big blue iron.
I can see a lot in your arguments, which is obviously based on experience at a high level in IBM and I respect that. My experience is probably lower down ;-), but it does give me the benefit of seeing a lot of things close too. I think your view of IBM has grains of truth, but it really does miss the complexity inherent in a big organization. I worked within IBM in the organization that worked on JDKs. I saw competent technical leadership - managers with PhDs and patents to their name - leading large cross country teams of good developers to create JDKs across a much wider range of platforms than Sun was willing to do. I saw the addition of good technologies like concurrent GC, great JITs, the persistent resuable JVM on zSeries. In other words - not just a port of the Sun codebase.
This is all fairly low level - these things are then embedded in IBM middleware which is the reason for doing it - and given away for free to whoever wants them. This is still techincal and not contracted out to some OSS organization. I know this is unlikely to change your view, but in my experience is that IBM is no shell.
Well that's a bracing view of things for sure. I suppose where we differ is that you are setting the benchmark much higher for what constitutes "technology" or a "technical success". I'd argue that a lot of great work is done in turning ideas "not invented here" into code and products that do actually meet the needs, requirements and expectations of businesses (which after all is still the main focus of IBM).
That's not going to win a Nobel prize, but it's still technical in my book.
Where's your evidence for this assertion that IBM is weaker technically now than then ? Since those days IBM has jettisoned old and unloved technologies like SNA and has got some pride and technical leadership back, often by working with people rather then bullying them. It's actually delivered and extending on good ideas that were "not invented here" (like J2SE and J2EE) where it has a very successful app server in WebSphere. Look at Eclipse - becoming a defacto tools environment which will benefit a lot of people.
Mainframe tech. wrenched in to the 21st Century - from 3270/ISPF/CICS etc. to J2EE apps on a nice unix base in 8 years of continuous development.
Hardware: POWER, IBM chips power Macs (G5), the Gamecube. The NG consoles will all have IBM technology inside. Too early to tell about Cell, but it's hardly an example of a company on the slide. Still the best at virtualization and partitioning. Still top of the patent league (for what that's worth I agree, but still a metric).
The Research division continues to innovate, and is no worse then it was under Akers - in fact much more focused on delivery and real world problems. What about the Community Grid ? I could go on and on...
Yep. Also, the compiler can encapsulate the accumulated knowledge of the ages about the target platform. There is no way on God's earth any normal person will know enough about the instruction set of the target platform to beat that accumulated knowledge. Anyone who does should be writing compilers, not application progams. Or maybe taking up another interest in life.
You beat me to it, but this is common practice and a good tip - I was surprised to see the original example written the other way round.
"I was astonished when I saw how beautiful Birmingham was. The buildings, the art, the use of the water. It is an extraordinary jewel of a city.... I was bowled over when I was there. It is quite wonderful." Bill Clinton, The Times, October 2nd, 2002
yours, a slashdot reading Brummie.
PS. Slagging off Birmingham is about the oldest meme you could pick up. Congrats on your originality.
But cash alone is not enough for survival. IBM had to change itself significantly to make it through the bad years. That took new leadership at the top - this I think would be firm evidence the MS was under pressure.
You could just ship a JRE at a specific version and service level with your code ? This is what some middleware vendors do at the server side, so they are shipping their code with the JVM and Class library code they have tested against.
I think this is a very good point. How about Linux on an IBM zSeries box ? This would get plus points from management for scalability, managability and decent service, all on reliable hardware. I think there are three distros that are packaged for the "Iron Penguin" : Suse, Redhat or Turbolinux.
See http://www-106.ibm.com/developerworks/java/jdk/lin ux/tested.html
for a list of supported distros and architectures for the IBM JDKs. You'll see that IBM supports Linux on PPC both 32 and 64 bit.
Yes, I see what you are saying, but I think it underestimates the extent to which what you might call the "corporate model" for developing software can deliver software people are willing to pay for. If you take Middleware today for example, it's undeniable that despite great alternatives (JBoss, LAMP things, even Zope) IBM makes a fortune in software revenues from WebSphere.
I honestly don't think this is right. For sure we can all forsee a future where OSes and large areas of middleware are commoditized and OSS versions are dominant, but I don't for a moment think IBM's intention is to wind up its Software Group, with its n labs round the world and strong links to research. I think the view rather is (a) we can remain ahead of the curve, at least enough that people will want to part with hard cash to buy products that we ship and (b) there are still things you can do with 1000s of internally distributed developers (coupled with good management and good project managers) that you can't do as easily in typical Open Source projects based outside of large companies. I'm thinking in particular of Servicing many releases concurrently on multiple platforms, or having high focus on RAS (Reliability, Availibility and Serviceability), or having specialist performance groups keeping an eye on product performance. It's these "nuts and bolts" aspects of software development that big companies like to know about and l33t hackers are not that concerned about (forgive the generalization).
RMI is somewhat limited, for example
- The server application must be written in java
- No object activation policies
RMI/IIOP (as implemented by WebSphere amongst others) at least gives you language neutrality.
Ouch, cheap shot :-) IBM produces many 100s of Redbooks a year and there is a standard toolset to maintain consistency. However, the source material given by the authors to the editors could have been done on any platform, in most any format...
I think Sony are getting gradually closer, and Apple must be at least a bit worried. Let's imagine a world where Sony (finally) drop ATRAC3+ only, and their Hi-MD devices and new HDD devices support MP3 natively and ATRAC3+ at whatever bitrate you desire. Given that the low end Hi-MD ships at 140UKP Sony could wipe out the Compact Flash competition in months with their cheap 1GB removable media. Much higher capacities are in the offing for MO as well. They are the only other vendor to offer integration with a online music store, which is a serious advantage over vendors like iRiver (despite their excellent hardware). The Connect store may suck now in comparison with ITMS, but heh, it's software, it can change quickly.
Other: not clearly defined & buzzword like Java Platform, EJB etc
EJB isn't a buzzword, it's defined in the J2EE Spec.
You missed out the JDK. No implementation would be complete without the basic tooling e.g. javac, javap etc.
Because currently there are no Java VM implentation which is as complete as Sun Java VM
That's wrong. The IBM JDKs pass the JCK and have VM implementations just as complete (but different) to the Sun JDKs.
Re: the end of the article. I'm no historian, but I'm certain Condoms go back at least 200 years, so a scientist who mentioned this in 1800 wouldn't haven't been thought absurd.
This isn't as high risk as it sounds, for at least two reasons. The first is that IBM has lots of people with 50 pound brains employed to think about what's next, rather then what's here already. They already realize that OSS can make a commodity out of what was previously sold for a profit. Since it's happened for the OS, it will happen for middleware as well. Secondly, the organizations they sell to aren't impressed one way or another by some standlone discrete piece of software like MySQL or JBoss. They want to know: where's the tooling that integrates closely with the runtime ? What's your 5 year development plan for this product ? How does this help me integrate different legacy systems ? Are you playing with industry standard bodies ? Lots of different questions that a disperate group of OSS developers may or may not be able to answer.