Slashdot Mirror
James Gosling On .NET And The Anti-Trust Trial
gwernol writes: "There's a short but interesting interview with James Gosling over on ComputerWorld. He talks about the differences between J2EE and .NET and also about the Microsoft anti-trust trial. Some interesting perspectives from the founder of Java."
I read this comparison last November. The author, Dean Wampler, Ph.D., concludes the following:
.NET are most appropriate for large-scale, high availability applications. The documented tests say little about how well these frameworks support those applications.
.NET and Java Pet Stores support the same features, but they implement different "nonfunctional" requirements. The .NET version assumes a single hardware/OS/database combination and makes performance paramount. The Java version supports multiple hardware/OS/database combinations and ranks performance as less important. In fact, both frameworks can support either emphasis. Hence, comparing the two code bases is misleading.
.NET is a well-designed framework with good tools. J2EE provides greater freedom, but the J2EE community can't ignore the need for tools that create powerful and efficient applications in a timely manner."
"The experimental flaws of the Microsoft tests render the performance comparisons unusable. All tests must be run on the same test bed and a more suitable application must be chosen. J2EE and
The
For developers who are comfortable with limited choices,
The Ultimate Linux Bookmark
Microsoft released their shared-source CLI and C# implementation a couple days ago. I've been playing with it on my FreeBSD box, and while it's hardly the Java killer Microsoft's making it out to be, it's an interesting piece of work. If you've got a spare FreeBSD or Windows XP box lying around, download the source, compile it and play around with it some; it's always nice to know something about the platform before you start bashing it mercilessly :-)
"The problem with the French is that they don't have a word for 'entrepeneur'." -George W. Bush
"They certainly could have been more creative about the language."
This coming from someone who tries to pretend he was inspired by smalltalk (since its more OO pure) even though its plainly obvious to anyone who knows jack shit about languages that the Java object model is a strict subset of C++'s. I mean, in smalltalk, things like reflection and introspection fall out of the way the object model works. In Java, its a bag on the side, because Bjarne didn't design it into the C++ object model, which Gosling stole wholesale.
Then lying about it and criticising others... This man is obviously incapable of feeling shame.
The worst part of it is that there are millions of "developers" out there who only know Java (or more often: switched to Java from Visual Basic) who simply accept Sun's marketing as fact.
From this comment in the article: And from a personal point of view, I personally actually read the [Windows] XP license and decided I couldn't sign it. So I've been shifting over to Mac.
:-P) and there is even a movement to create a Trusted Darwin http://www.stosdarwin.org/ . This could be a real opportunity for university CS departments to adopt a platform that really does support Java instead of the Win boxes that so many universities seem to be purchasing for their CS programs.
This is very interesting and parallels what we and others have been experiencing. There is this slow but dramatic sea-change taking place in the community of scientific computation and programming communities. Folks that never before would even look at a Mac are moving to the platform for a variety of reasons including its UNIX core and ease of use. Additionaly it seems that Apple is actually listening to their users these days. They include features requested and the open source Darwin allows for significant development from the community (assuming you are old enough to sign the agreement
Visit Jonesblog and say hello.
This is Sun propaganda pure and simple. I can't wait for a headline on the front page telling us that Coca-Cola says new Pepsi is disappointing. When Microsoft have made less-than-favorable remarks about Java in the past it has instantly been flagged as FUD.
I suggest folks take Sun PR and Gosling's remarks with a grain of salt. Evaluate the technologies for yourselves and decide accordingly.
What a disappointing interview... Could you expect something as boring as this from, say, Larry Wall? Never. Or when Stroustrup criticizes Java, he has some valid (and interesting) arguments. But I have not found any insights in this Gosling interview. Microsoft sucks and J2EE rules. How interesting.
;)
The funny thing is that he says
1) They copied everything from Java
2) They could add clever things to their language, but they didn't
Well, at least, he's honest about Java
Linux is totally chewing up their low end. They don't want to admit it straight out - they have been playing nice with open source folks while quietly taking Cobalt off of the market and making it a bit player.
Meanwhile IBM is taking it apart at the high end with a proposition that focuses as much on services as hardware and software..because IBM knows billable hours are where the real renewable revenue is.
On the architecture side, Sun is pitting itself against an entire enconomy - Intel and Microsoft. Sun simply can't outresearch, outspend or outmarket either of these companies, let alone both of them and their attendent co-competitors (Dell, AMD, HP, etc). Once Microsoft gets Win2k up to par in every respect with Solaris (it will happen), they will start peeling high-price clients off of Sun with little contest (meanwhile linux will chew up Sun's low end more and more).
On top of all of this, they're playing mindshare catch-up with the half-hearted JavaOne. Sorry James, MS beat you to the punch on webservices by a year.
I just hope Java can't be opened up enough that it doesn't evaporate along with its owner.
Here is an interesting article, describing how Microsoft is trying to get Java developers to write J# (java) code for the .NET platform. Unfortunately .NET only runs on the Windows platform, which has Sun a bit upset.
--It's Pimptastic!--
Yeah. Yeah. I mean, abused and ripped off was also in there just to some extent.
I do find it obvious that there are some things in C# that are like Java. However, it seems to me that this is generally moot considering both languages where heavily inspired from C++. When you do your homework, you find that C# is actually quite different the Java:
+ C# is completely OO - even an Int32 is an Object. Java uses primitive types.
+ C# uses Delegates for Event Handling (think function pointers, but different).
+ C# supports the use of Properties instead of Getter and Setter methods.
+ C# supports Indexers which allow objects to be treated as Arrays.
+ C# forces explicit Method Overriding (via the virtual/override or new keywords).
+ C# supports namespaces. Unlike Java's packages, namespaces do not rely on a file/folder structure.
+ The C# Abstract or "Virtual Machine" (CLR) is not designed for C#, rather for language neutrality (to an extent). Java and the JVM, however, are closely tied.
I could go on. Whethor or not you think that these differences are Good Things(tm), the point is, they are definitely different langauges. Although there may have been some inspiration from Java, I'd be hard pressed to call it a "Java Ripoff".
There is no longer anything that can be done with computers that is nontrivial and clearly legal. -- Paul Phillips
"Imitation is the sincerest form of flattery"
While I have great respect for Mr. Gosling's prolific contributions, clearly this imitation goes both ways. For example:
Microsoft Transaction Server 1.0, shipped 12/96
* automatic transactions for objects, including Java objects
* ObjectContexts for automatic services on behalf of objects
* declarative transaction requirements e.g. Transaction Requires New
* declarative, automatic role-based security, and IObjectContext::IsCallerInRole()
* etc.
Enterprise Java Beans, 1.0 final spec shipped 1Q98(?)
* automatic transactions for Java objects
* SessionContexts for automatic services on behalf of objects
* declarative transaction requirements e.g. TX_REQUIRES_NEW
* declarative, automatic role-based security, and EJBContext.isCallerInRole()
* etc.
The provenance of the ideas behind EJB/J2EE, arguably Sun's most commercially important Java technology, would seem to be revealed in its choice of identifier names.
-- an ex-Microsoft software developer
Object wrappers for primitives is not the same as the primitives themselves being treated as objects. Anyone whose used a true OO language like Smalltalk cringes and the inconsistency in Java between primitives and objects. Even C++ tries to make them as interchangeable as possible especially with templates.
For instance in Java there's no way to pass just a primitive like "5" or 2.6 to a method that takes an object while in C# and Smalltalk you can.
If you want to talk about non-OOP features, C# is full of them. Like structs for example. Who came up with that idea? And how about pointers? WTF?
The above comments how that you've somehow confused object oriented with Java which unfortunately are not the same thing. An object oriented system has 3 main qualities i) encapsulation or information hiding ii) inheritance and iii) polymporhism. All three of which can be done with C# structs (or value types). Secondl, I am immensely confused what the existence of an explicit pointer type has to do with whether a language is OO or not.
As far as Indexers go (and pretty much all the differences between Java and C#), they are just syntactic sugar that really just makes code confusing to read compared to Java.
Really? So is easier to read than On what planet?
That's a hoot! The fact is that CLR doesn't support anything that can't be accessed from C#. That's why implementations of other languages have had to drop features like multiple inheritance before CLR implementations. All CLR does is provide a Procrustian cot for other languages to lie on. Head over the top? Lop it off!
The Java VM was designed to run Java while the CLR was designed to be language agnostic. The fact that C++ can run on the CLR is a testament to this fact.
- You're using J2EE. That implies you probably should be using JBuilder 6 Enterprise which has numerous J2EE features for automating configuration of EJB's. If you're not using the enterprise edition, then your comparison is lacking in credibility, as Visual Studio .NET enterprise edition is approximately the same price.
- You have to edit four files to add a field to an EJB? Let me assume for a minute that you're using container managed persistence, which is the only scenario that would require such changes. Most tools will allow you to define the new field in your local interface, and will then propagate that field to your implementation class and your ejb-jar XML file. The second XML file, I will assume, is a custom deployment descriptor. Again, I would hope you're using a vendor's tool to manage this thing. But even if not, I find your indignaton towards all of this "work" somewhat amusing.
To put this tremendous amount of work in context, how much work does it take to add a field to a regular database table wtih a SQL call in JDBC, or for that matter, ADO.NET? That would require:
- doing a DML statement on the table to add the column at easiest. In some environments this may require several DML statements to create the new table, re-populate it with old data, populate it with the new column's data, then drop the old table and rename the new one.
- changing 1-2 method call signatures to take in extra parameters for inserts and updates.
- changing the JDBC code for reading, updating, and inserting to take the new field into account
- possibly adding the field to a data object that holds the data in memory.
Phew! I'm glad there's alternatives to EJB, it's so much easier without it.
Now, on to the next cow:
Deploying a JAR to two different places (jBoss and Tomcat). Firstly, I question what the problem is. You would deploy an EJB JAR to a jBoss instance and a WAR to tomcat, or you could just put it into one big EAR file and fo'gettaboutit. If you have two servers, then the ant optional tasks package could very easily do this work for you with approximately 3 lines of XML configuration.
The best part about your post is that "there probably are better ways". Yes, there are. Hire a consultant for 3 or 4 hours to help you out, it will probably be worth the $1000+. If you're missing GUI tools for jBoss, that would be because it doesn't really have any. Use a commercial server if you're not willing to hand-craft your config & deployment.
-Stu