It took over 5 years for Sun to add NBIO to Java. The Java Community Process simply does not work.
would you prefer a system with no community process? Like I said, managing people's desires is much harder than making the code changes. Sure the JCP might need better managers, but I'd much rather work to improve it than just say it sucks or that it doesn't work. There's plenty of flaws in the JCP process, but atleast there is one. It's up to the community to make it known to Sun what are the most important issues. Have you ever participated and sent in comments? It's a lot like voting right. You get out of it what you put into it.
The fact that java expert are free to criticize and influence the direction of Java is one of the most attractive part of the language for me. Look at NBIO in jdk 1.4.0. From my understanding, it came about because others outside of Sun saw a need and pushed for it. Now it exists and is part of jdk1.4. If we take the statement "no language is perfect" to be true, the article does make some compelling arguments.
Sure people are going to disagree over what is more important, but the fact these discussions occur in the expert groups regularly shows the community is open. Reguardless of who or what company invented the language, the more important thing is it include a process by which the community can drive its direction. Balancing the needs of all the users isn't an easy task and in most cases if much harder than writing the code.
Microsoft has agreed with the 8 remaining states and will begin to break Microsoft into three different divisions. Bill gates has voluntarily submitted himself to the state police and is awaiting for sentencing. Steve jobs went on record stating "I love windows. All this time I was simply jealous of Bill Gates and his billions."
If you have experience and facts to back up your assertions, I'm more than willing to admit my own ignorance/misunderstanding. Saying something without one piece of technical information is pretty easy. All of the facts I stated are verifiable on mappoint soap message board on msdn. I'm not the only person who finds mappoint.NET soap to be frustrating for non-windows platform.
I've also used the non-soap version of mappoint and in my mind it's a step backward feature wise. With the older version, you can get a URL to a map, whereas the newer version requires the application to send it directly to the browser, or save the image locally. This causes more problems and work for the developer because it means there has to be a thread to clean up the images, or set an arbitrary number of images and over write the old ones. Other mapping applications support the ability to get a remote URL like the older version. As far as I know, MapPoint.NET SOAP is the version.NET service released by MS. Of course no one knows if that is going to be the dominant model for MS, but from this particular model, it's far from enterprise class service. I a person wanted to integrate mapping into a site that gets 500K map hits a day, retrieving the map every time is going to be expensive bandwidth wise and require a lot more work to build a scalable website. Others may have different experiences, but that's what I've seen first hand.
I don't claim to every single thing about how microsoft sees web services, but I've tried to playing microsoft's MapPoint.NET and went through most of their online docs.
If you go on MSDN and look at the MapPoint.NET docs, examples and the problems people have been having you'll concede there's mixed message coming from microsoft. What do I mean? First off, the way MapPoint.NET SOAP works, it requires authentication using NTLM. Some will recognize it as IIS Challenge/response. Problem is when you try to connect to MapPoint.NET soap servers, it may take multiple connections. In MS MapPoint SOAP press release they claim to be better than Mapquest. I've used both. With Mapquest it takes one connection because the authentication and request happen with the same connection. Not only that, you can get a remote URL to a map. With MapPoint.NET you get map encoded in base64. This means any developer integrating MapPoint SOAP has save the image locally and manage those resources. If MS went to a stub model where the request generates a stub to an image, the response would be faster and reduce the traffic to the client's server. The application can print the URL to the browser and the browser would get the image from mappoint servers. You might think the MS model is better, but having worked with several mapping products, the MS model of a "web service" sucks. Following the MapPoint SOAP model would mean more work and more hardware to support what would be a simple application with Mapquest. Now I don't that this is how MS sees web services, but MapPoint.NET soap is the first official.NET service from MS that I know of.
Now lets look at this closer. MS recommends using UTF16 encoding for SOAP messages, but digest authentication with IIS requires base64 encoding. Having the ability to use different encoding provides great flexibility, but doing that over HTTP seems lame to me. Sure SOAP is not dependent on HTTP, but according to MapPoint SOAP in its current form it is tied HTTP. MS could have easily used encryption to hash request and make req/resp one single connection, but instead they chose to turn on challenge/response.
There were several questions on msdn about using other languages with MapPoint SOAP. If someone is trying to integrate mappoint soap with something written in C or C++ on unix, aix, irx, or vms it would be a pain. when developers asked about other platforms and languages, the response was "it's not supported, you're on your own." How's mappoint soap suppose to make a developer's life simplier? If MS is really interested in making loosely coupled web services, then it should support other languages and platforms, so that someone on a vms mainframe who wants to add mapping to an application can do it without writing their own SOAP, UDDI and WSDL drivers. Technologies like castor and IBM's WSIF appear more balanced and flexible. Even though MS uses standard protocols, their implementations and use is different.
Lets not confuse implementation with protocol specification please.
this article on ibm is honest about web services and it's short comings.
send a stateful request for info to a server of some kind. Since developers have been working around the Browser's lack of state for years and non browser applications can use a myriad of protocols etc to communicate to their server, such as XML-RPC,RMI,CORBA or whatever.
No where in the current WSDL, or SOAP specification mention stateful management or framework. If you look at IBM's perspective on web services, it includes both stateless and stateful applications. If you look at MS's application center and application server, it explicitly says:
Accessing Stored Data
Unlike ADO, which makes it easy to build heavy-touch clients that don't scale well, ADO.NET is biased toward building light-touch clients. An ADO.NET client uses forward-only, read-only cursors to read data. Stateful server-side cursors aren't supported, so the programming model encourages short connection lifetimes. Clients that read and process data directly can use the ADO.NET DataReader object, which provides no caching for returned data. Alternatively, data can be read into a DataSet object, which acts as a cache for data returned from SQL queries and other sources. Unlike an ADO Recordset, however, a DataSet cannot explicitly maintain an open connection to a database. the original page
I still a lot of conflicting/confusing/contrasting marketing bs about what web services really means. Until everyone agrees and makes all the necessary protocols and frameworks free, web services is going to have a hard time finding wide acceptance.
Can't get to the damn article, so rather than talk about what Scott Deitzen said, I'll talk about what I do know. There's a lot of confusion about web services and most of it is marketing BS. Web services the technology, language, and framework isn't BS. Loosely, it an attempt to move away from tightly integrated systems. What does that mean? To me that means rather than have clients and servers all designed specifically to use a limited set of protocols, applications are designed to see each other as services. where a traditional client server application like point of service connects several dumb terminals to a central server using a proprietary protocol, web services defines standard ways for discovering and using services. A service is defined by WSDL (web services description language) which defines the server, port and message structure. when two applications need to talk to each other, they look at each other's WSDL to figure out how to transform a message into a desired binary format.
IBM has been steadily putting lots of good code and ideas about how to approach web service. Having read through most of the white papers on microsoft's site and worked a little on.NET application which uses SOAP and WSDL, microsoft's model is not ready for prime time. IBM recently release a Web Services Innvocation framework and web services toolkit.
there are also other efforts like castor, which are designed with web services model. IBM has a protocol called Web Services Flow Language, which borrows ideas and techniques from pi calculus. Web services as a technology will mature, but not at the rate microsoft claims or in the way microsoft thinks.
It's good to know someone sees C# and java for what it is, A Tool! IMO, it's good that those devices are switching to standard protocols, but it's hardly new. Wireless point of sales systems for restaurants have existed since early 90's. I know for a fact because in southern California, I went to a restaurant where a friend worked that had a wireless sytem. My friend would take the order and it would be sent wireless to a terminal for the cooks to see.
Having standard ways to do remote communication is good, but all the bloated hype from all the various camps does a lot of dis-service. As far as I'm concerned, all the big tech companies are guilty of it, so it's not like MS or Sun are worse than others. The real benefit for businesses with technologies like C#, Java,.NET and web services is standard protocols and frameworks. The chances of a company finding skilled replacement when staff leave is more likely than a non-standard proprietary system. For me, the hassle of marketoid gibberish is worth it to move towards more standard and flexible ways of developing software.
I feel it is a positive step, but the administration needs to be more clear about what exactly they mean. Talk is cheap. When I see some legislation that improves/encourages/balances the research/report/fix/disclosure of bugs I'll smile. Until then, I'll take the perspective of hope for the best and expect the worst.
Big business owns the government, so getting tough laws passed to measureably improve software security is a very tough task. The key here is measurable. Not some bs statistics that politicians can throw around. I want results.
Did you use connection pooling in the db hello test? If not, I would be curious to see if using connection pooling for resin and tomcat improve the performance. Depending on the driver, there may be a 0-5ms wait time. I know that this was true of ODBC back in the sql server 5 days. thanks for posting your results. Even though it doesn't provide answers as to which server is more scalable, it does provide a baseline control for others to compare to.
In my own benchmarks of web applications, I try to include a super simple test to establish a baseline, so that there is a point of reference for comparing the real application.
Other's haven't mentioned this, but I think the real issue here is persistence. If I buy a piece of software from company A, I can't just sit around and read the source code to figure why it works a certain way to figure how to best use the application. This means the only eye balls reading the code for fun or for real reasons are new programmers joining company A. How likely is a closed environment going to encourage programmers and user to explore and look at the code? Trojans and virii will happen, it's a given. Encouraging people to look at the code is your best bet.
How many companies are going to tell a new programmer to go ahead and spend 6 months reading through all the code? How many companies encourage all the programmers to look at old code, check every line every couple weeks and perform extensive regression testing? From my own experience, few companies look at old code and the regression testing is typically a narrow focus on the functional aspects. Things like a trojan aren't going to be caught by the typical regression testing procedures.
On my free time, I do read through open source code for fun. From my own biased experience, open source code tends to be much cleaner and better documented than closed source projects. This isn't including all the PERL code I've seen written in creative ways to make visual art. I've also seen clean PERL code, but that's another story. Back to the point. Persistence is what wins in the end when it comes to security. The minute a person get lazy is when an attack will happen. But I seriously doubt this will change in the near future, since it's really a matter of culture. Businesses can't afford to have a team of programmers to sit around and audit their security every couple months. So unless our culture changes and realizes quality is more important than convienance, things like this will continue to increase in frequency. Of course everyone living in a modern techno society is guilty of it. But that's not to say technology is the cause of it, though they are related.
Reguardless of whether there is global warming, (which no real scientist can prove beyond reasonable doubt) does not negate the fact humans are majorly F_ _ _ing up this planet. Using resources too quickly, poluting the air/water we breathe/drink and basically killing every freakin thing in site. Any body stupid enough to buy into this whole global warming argument is falling into a political brain trap. Instead of arguing about how to actually change society in measurable ways, the debate ends being about global warming, which typically results in more fossil fuels/resources being used to try to prove something unprovable.
One might believe that is the case, but if you go back in recent history, one can find examples that prove otherwise. This is semi-OT, but it is relevant. My wife was reading a recent article in MS magazine by Gloria Steinem that addressed this issue. When MS magazine was driven by advertising, they were pressured to drop articles that was critical of the clothing and makeup industry. Now that MS has changed their revenue model, they are back to publishing articles driven by honest reporting. Here's a recent editor's note. I honestly believe consumers are willing to pay more for a magazine if it is well written with solid content. If I had to choose between paying 3.00 for a magazine full of adds and PR junk or 12.00 for objective reporting, I'll pay 12.00.
This whole pattern of corporate america dictating was is news worthy needs to swing back towards objective reporting. I think many people are sick and tired of the drivel on the news stands today.
I hate to see people loose their jobs, but zdnet and pcmag have been useless for over 5 years now. PCMag in particular stop reporting back in 95 and only rehashed PR junk. It is time for new crop of magazines to take the place of all the dead weight in the tech reporting industry. Zd used to have solid in depth articles that were fairly objective. But soon after win95 came out, that all changed. Of course these are my own biased opinions, but I know other techies share similar perspectives.
IBM announced a deal a few months ago that would see the NYSE move to DB2. Take that, Sybase! I stand corrected. Why wasn't that article posted on/.? That's real news, not this other junk.
I told by a CEO once when asked about the future of the company. He drew the following comparison. "We have no revenue right now, so we can only go up. Whereas some one like MS has to always push to sustain the growth, because if they don't the stock market will think they've peaked and are on their way down."
Microsoft has to become a major player in enterprises services. If they don't they won't be able to sustain their past growth rates. The OS is rapidly becoming a commodity, now that win 2K is stable enough that that most people don't need to upgrade ever again. Win 2K already does what most people want and more, so the only way to continue to grow is new markets.
That everyone already knows. Microsoft is doing all these interviews to paint a picture that.NET really is ready for the enterprise world of 24/7 computing. Back in 98 MS commissioned some company to prove SQL Server was good enough to run the NY EX, but everyone in the RDBMS business knows Sybase ASE run the stock market. Is it possible that if MS can't get it's act together with.NET, that they have reached their peek?
Doh! I didn't expect that many responses to such a simple question. I'd have to agree with a lot of the posts and say most politicians lie, but I am still puzzled as to why so many people believe statistics, and opinions from people in power as fact. It's not isolated to politicians and executives. I've had similar experiences in church where the pastor, priest, or monk says "this is the interpretation accepted as the best."
Maybe I'm an optimist, but I tend to think people aren't that dumb. Perhaps people get lost in the moment, go along with the flow and forget there's a real consequence. But I'm sure if you ask Clinton, Bush or Reagan, they'll tell you the "true" reason:)
I remember president Ronald Reagan pulling the same kind of stuff when he was in office. Some of the statistics Reagan quoted in his public speeches often were wrong or had no data supporting the claim. Why is it that people buy into BS when it comes out of the president or some CEO?
Are people being stupid, or simply letting themselves get caught up in the excitement?
Re:Your links say the opposite.
on
.NET for Apache
·
· Score: 1
excuse the run on sentence, but what's so hard to understand? persistence engine is just a state machine. stateful session framework is simply a framework for developing stateful applications that follow a consistent pattern to ensure applications can work together smoothly.
I'm not saying every freakin app needs to be stateful, since most of the current web sites don't. News, e-commerce and search engines don't really need full blown stateful sessions or transactional processing. It's just lame to excuse Microsoft's poor performance on JDBC drivers when it's their own fault. What Oracle and Sun do to smear MS is also not excusable. Those results only show the person doing the benchmarks were biased and purposely handicapped each set of tests for their own agenda.
Re:Your links say the opposite.
on
.NET for Apache
·
· Score: 1
I think you're interpreting the data erroneously. The first two charts clearly state Oracle and Mysql beat SQL Server in throughput and response time. The 3rd and 4th charts show a.NET based application. If you compare the first two charts with the last two, it's obvious some caching is occuring on both SQL Server and.NET Application server, or the request buffer is set high enough to pace the incoming requests. Anyone who has done n-tier application development will say "no duh!". Having worked with ASP for E-Commerce, it's nice to see ASP.NET performing better than the older non-scaling ASP. Microsoft still has a long way to go to reach full n-tier architecture with a full fledged persistence engine and generalized stateful session framework.
Since when is apache == covalent?
on
.NET for Apache
·
· Score: 1
Covalent developers have contributed tremendously to apache, but let's make one thing clear. Covalent working with apache != apache.org working with microsoft. There's nothing wrong with apache.org working with micrsoft as long as MS agrees to apache's license agreements.
Re:maybe this is a good thing?
on
.NET for Apache
·
· Score: 1
According to those charts, SQL Server 2000 sucked compared to Oracle. The comparison between Microsoft's.NET and BEA weblogic is totally pointless. If you look at Microsofts.NET application server page, it states emphatically it does not support state or state replication. If you look at the config.xml file for BEA, you will see they've set up weblogic to cluster. Gee, I wonder why ASP.NET beat BEA.
Also, if you look at the clustering described on.NET application server, you will see it is basically round robin load balancing. What Microsoft calls clustering isn't application clustering. It's load balancing.
My take on it is this. For a simple bookstore app, you don't really need stateful transactions. Not unless you have a very complex system that absolutely needs to be transactional and have to be totally fail safe. If someone wanted to do the opposite and show how bad.NET application server is, they could write a stateful application and make.NET crawl.
Not to defend microsoft, but their test case is absolutely on target because they're going for medium sized businesses that want something which scales better than IIS with Microsoft Transaction Server. And by far,.NET does that without a doubt. But comparing a full fledged application server that is mature and industry tested to a glorified stateless transaction server is just plain old wrong and stupid. Like anyone with real experience with transactional systems is going to believe that one. Now I could be wrong, but from the examples files the page provided, it's pointless to setup clustering in BEA. Anyone with BEA experience will tell you this particular example it's better to throw a local director infront and leave clustering off.
Given the huge population of China and potential market, Microsoft's chance of continuing the same level of growth just got worse. Perhaps that's why microsoft is pushing into the service world. If they don't, they won't be able to continue the same level of growth over the next few years and will no longer be the darling of the stock market.
Slashdot Mirror
would you prefer a system with no community process? Like I said, managing people's desires is much harder than making the code changes. Sure the JCP might need better managers, but I'd much rather work to improve it than just say it sucks or that it doesn't work. There's plenty of flaws in the JCP process, but atleast there is one. It's up to the community to make it known to Sun what are the most important issues. Have you ever participated and sent in comments? It's a lot like voting right. You get out of it what you put into it.
Sure people are going to disagree over what is more important, but the fact these discussions occur in the expert groups regularly shows the community is open. Reguardless of who or what company invented the language, the more important thing is it include a process by which the community can drive its direction. Balancing the needs of all the users isn't an easy task and in most cases if much harder than writing the code.
Microsoft has agreed with the 8 remaining states and will begin to break Microsoft into three different divisions. Bill gates has voluntarily submitted himself to the state police and is awaiting for sentencing. Steve jobs went on record stating "I love windows. All this time I was simply jealous of Bill Gates and his billions."
I've also used the non-soap version of mappoint and in my mind it's a step backward feature wise. With the older version, you can get a URL to a map, whereas the newer version requires the application to send it directly to the browser, or save the image locally. This causes more problems and work for the developer because it means there has to be a thread to clean up the images, or set an arbitrary number of images and over write the old ones. Other mapping applications support the ability to get a remote URL like the older version. As far as I know, MapPoint.NET SOAP is the version .NET service released by MS. Of course no one knows if that is going to be the dominant model for MS, but from this particular model, it's far from enterprise class service. I a person wanted to integrate mapping into a site that gets 500K map hits a day, retrieving the map every time is going to be expensive bandwidth wise and require a lot more work to build a scalable website. Others may have different experiences, but that's what I've seen first hand.
I don't proof read cuz I'm lazy
If you go on MSDN and look at the MapPoint.NET docs, examples and the problems people have been having you'll concede there's mixed message coming from microsoft. What do I mean? First off, the way MapPoint.NET SOAP works, it requires authentication using NTLM. Some will recognize it as IIS Challenge/response. Problem is when you try to connect to MapPoint.NET soap servers, it may take multiple connections. In MS MapPoint SOAP press release they claim to be better than Mapquest. I've used both. With Mapquest it takes one connection because the authentication and request happen with the same connection. Not only that, you can get a remote URL to a map. With MapPoint.NET you get map encoded in base64. This means any developer integrating MapPoint SOAP has save the image locally and manage those resources. If MS went to a stub model where the request generates a stub to an image, the response would be faster and reduce the traffic to the client's server. The application can print the URL to the browser and the browser would get the image from mappoint servers. You might think the MS model is better, but having worked with several mapping products, the MS model of a "web service" sucks. Following the MapPoint SOAP model would mean more work and more hardware to support what would be a simple application with Mapquest. Now I don't that this is how MS sees web services, but MapPoint.NET soap is the first official .NET service from MS that I know of.
Now lets look at this closer. MS recommends using UTF16 encoding for SOAP messages, but digest authentication with IIS requires base64 encoding. Having the ability to use different encoding provides great flexibility, but doing that over HTTP seems lame to me. Sure SOAP is not dependent on HTTP, but according to MapPoint SOAP in its current form it is tied HTTP. MS could have easily used encryption to hash request and make req/resp one single connection, but instead they chose to turn on challenge/response.
There were several questions on msdn about using other languages with MapPoint SOAP. If someone is trying to integrate mappoint soap with something written in C or C++ on unix, aix, irx, or vms it would be a pain. when developers asked about other platforms and languages, the response was "it's not supported, you're on your own." How's mappoint soap suppose to make a developer's life simplier? If MS is really interested in making loosely coupled web services, then it should support other languages and platforms, so that someone on a vms mainframe who wants to add mapping to an application can do it without writing their own SOAP, UDDI and WSDL drivers. Technologies like castor and IBM's WSIF appear more balanced and flexible. Even though MS uses standard protocols, their implementations and use is different.
Lets not confuse implementation with protocol specification please.
this article on ibm is honest about web services and it's short comings.
No where in the current WSDL, or SOAP specification mention stateful management or framework. If you look at IBM's perspective on web services, it includes both stateless and stateful applications. If you look at MS's application center and application server, it explicitly says:
I still a lot of conflicting/confusing/contrasting marketing bs about what web services really means. Until everyone agrees and makes all the necessary protocols and frameworks free, web services is going to have a hard time finding wide acceptance.
Here is an article that is more balanced and realistic about web services.
IBM has been steadily putting lots of good code and ideas about how to approach web service. Having read through most of the white papers on microsoft's site and worked a little on .NET application which uses SOAP and WSDL, microsoft's model is not ready for prime time. IBM recently release a Web Services Innvocation framework and web services toolkit.
there are also other efforts like castor, which are designed with web services model. IBM has a protocol called Web Services Flow Language, which borrows ideas and techniques from pi calculus. Web services as a technology will mature, but not at the rate microsoft claims or in the way microsoft thinks.
Having standard ways to do remote communication is good, but all the bloated hype from all the various camps does a lot of dis-service. As far as I'm concerned, all the big tech companies are guilty of it, so it's not like MS or Sun are worse than others. The real benefit for businesses with technologies like C#, Java, .NET and web services is standard protocols and frameworks. The chances of a company finding skilled replacement when staff leave is more likely than a non-standard proprietary system. For me, the hassle of marketoid gibberish is worth it to move towards more standard and flexible ways of developing software.
Big business owns the government, so getting tough laws passed to measureably improve software security is a very tough task. The key here is measurable. Not some bs statistics that politicians can throw around. I want results.
In my own benchmarks of web applications, I try to include a super simple test to establish a baseline, so that there is a point of reference for comparing the real application.
Not revive a dead horse, but what would the good doctor say about this. Could his approach do any better at Go than current attempts?
How many companies are going to tell a new programmer to go ahead and spend 6 months reading through all the code? How many companies encourage all the programmers to look at old code, check every line every couple weeks and perform extensive regression testing? From my own experience, few companies look at old code and the regression testing is typically a narrow focus on the functional aspects. Things like a trojan aren't going to be caught by the typical regression testing procedures.
On my free time, I do read through open source code for fun. From my own biased experience, open source code tends to be much cleaner and better documented than closed source projects. This isn't including all the PERL code I've seen written in creative ways to make visual art. I've also seen clean PERL code, but that's another story. Back to the point. Persistence is what wins in the end when it comes to security. The minute a person get lazy is when an attack will happen. But I seriously doubt this will change in the near future, since it's really a matter of culture. Businesses can't afford to have a team of programmers to sit around and audit their security every couple months. So unless our culture changes and realizes quality is more important than convienance, things like this will continue to increase in frequency. Of course everyone living in a modern techno society is guilty of it. But that's not to say technology is the cause of it, though they are related.
Reguardless of whether there is global warming, (which no real scientist can prove beyond reasonable doubt) does not negate the fact humans are majorly F_ _ _ing up this planet. Using resources too quickly, poluting the air/water we breathe/drink and basically killing every freakin thing in site. Any body stupid enough to buy into this whole global warming argument is falling into a political brain trap. Instead of arguing about how to actually change society in measurable ways, the debate ends being about global warming, which typically results in more fossil fuels/resources being used to try to prove something unprovable.
This whole pattern of corporate america dictating was is news worthy needs to swing back towards objective reporting. I think many people are sick and tired of the drivel on the news stands today.
I hate to see people loose their jobs, but zdnet and pcmag have been useless for over 5 years now. PCMag in particular stop reporting back in 95 and only rehashed PR junk. It is time for new crop of magazines to take the place of all the dead weight in the tech reporting industry. Zd used to have solid in depth articles that were fairly objective. But soon after win95 came out, that all changed. Of course these are my own biased opinions, but I know other techies share similar perspectives.
IBM announced a deal a few months ago that would see the NYSE move to DB2. Take that, Sybase! I stand corrected. Why wasn't that article posted on /.? That's real news, not this other junk.
Microsoft has to become a major player in enterprises services. If they don't they won't be able to sustain their past growth rates. The OS is rapidly becoming a commodity, now that win 2K is stable enough that that most people don't need to upgrade ever again. Win 2K already does what most people want and more, so the only way to continue to grow is new markets.
That everyone already knows. Microsoft is doing all these interviews to paint a picture that .NET really is ready for the enterprise world of 24/7 computing. Back in 98 MS commissioned some company to prove SQL Server was good enough to run the NY EX, but everyone in the RDBMS business knows Sybase ASE run the stock market. Is it possible that if MS can't get it's act together with .NET, that they have reached their peek?
Maybe I'm an optimist, but I tend to think people aren't that dumb. Perhaps people get lost in the moment, go along with the flow and forget there's a real consequence. But I'm sure if you ask Clinton, Bush or Reagan, they'll tell you the "true" reason :)
Are people being stupid, or simply letting themselves get caught up in the excitement?
I'm not saying every freakin app needs to be stateful, since most of the current web sites don't. News, e-commerce and search engines don't really need full blown stateful sessions or transactional processing. It's just lame to excuse Microsoft's poor performance on JDBC drivers when it's their own fault. What Oracle and Sun do to smear MS is also not excusable. Those results only show the person doing the benchmarks were biased and purposely handicapped each set of tests for their own agenda.
I think you're interpreting the data erroneously. The first two charts clearly state Oracle and Mysql beat SQL Server in throughput and response time. The 3rd and 4th charts show a .NET based application. If you compare the first two charts with the last two, it's obvious some caching is occuring on both SQL Server and .NET Application server, or the request buffer is set high enough to pace the incoming requests. Anyone who has done n-tier application development will say "no duh!". Having worked with ASP for E-Commerce, it's nice to see ASP.NET performing better than the older non-scaling ASP. Microsoft still has a long way to go to reach full n-tier architecture with a full fledged persistence engine and generalized stateful session framework.
Covalent developers have contributed tremendously to apache, but let's make one thing clear. Covalent working with apache != apache.org working with microsoft. There's nothing wrong with apache.org working with micrsoft as long as MS agrees to apache's license agreements.
Also, if you look at the clustering described on .NET application server, you will see it is basically round robin load balancing. What Microsoft calls clustering isn't application clustering. It's load balancing.
My take on it is this. For a simple bookstore app, you don't really need stateful transactions. Not unless you have a very complex system that absolutely needs to be transactional and have to be totally fail safe. If someone wanted to do the opposite and show how bad .NET application server is, they could write a stateful application and make .NET crawl.
Not to defend microsoft, but their test case is absolutely on target because they're going for medium sized businesses that want something which scales better than IIS with Microsoft Transaction Server. And by far, .NET does that without a doubt. But comparing a full fledged application server that is mature and industry tested to a glorified stateless transaction server is just plain old wrong and stupid. Like anyone with real experience with transactional systems is going to believe that one. Now I could be wrong, but from the examples files the page provided, it's pointless to setup clustering in BEA. Anyone with BEA experience will tell you this particular example it's better to throw a local director infront and leave clustering off.
Given the huge population of China and potential market, Microsoft's chance of continuing the same level of growth just got worse. Perhaps that's why microsoft is pushing into the service world. If they don't, they won't be able to continue the same level of growth over the next few years and will no longer be the darling of the stock market.