Why choose ASP.NET vs. J2EE ?

Philip,

I'm trying to understand why MIT would have chosen .NET over J2EE as the toolkit for developing Internet applications. MIT comes from the open-source and Unix world (and a lot of people there do anything they can to avoid anything Microsoft-related), so it would make sense for them to choose a more vendor-neutral platform (e.g., Apache/Tomcat/Java) and all the other open development tools available.

Would that be a fashion statement as well? What are the particular advantages of .NET over J2EE that made it the superior choice? In my mind, I would look at a solution that didn't require paying Microsoft's excessive licensing fees, and that used as much open- source software as possible.

And what about security? A poorly-configured Apache server can be as bad as a poorly configured IIS server, but isn't it apparent that .NET servers will be less secure than their open-source counterparts (given the security track record of Microsoft products -- at least in the short term)?

-- Marcelo P. Lima, April 10, 2002

Answers

MIT as a whole is pretty wealthy but individual IT budgets tend to be modest. With only $1 or $2 million to spend per system MIT needs to put its money into capabilities that are valued by end-users. A true J2EE system would involve EJB and container-managed persistence. All of that automatically generated SQL code from the application server can be 100X slower than hand-authored SQL, which means MIT would have to buy 100 times as much computing hardware to support the same application. Not to mention that projects that built on top of J2EE are famously unproductive, expensive, behind-schedule, and inflexible. If you had a $20 million, 3-year budget to build something simple like photo.net, you could certainly do it with J2EE (or raw C with the Oracle C library for that matter). But MIT needs to produce things sort of like photo.net with a few undergrads hacking away over a summer. And they need to be able to tear down 20 percent of it and add another 50 percent the next summer as ideas evolve.

J2EE also looks like a classical IT dead end. There are dozens of different application servers and execution environments for Java. All have subtle differences so that an application built for, say, WebLogic won't run under Tomcat or Websphere. Maybe the application be ported by changing only 1 percent of the code but it isn't obvious which 1 percent. This is the same situation that Unix presented in the early 1990s. An AIX application wouldn't just work on SunOS or HP-UX. Close but not compatible. Faced with all of these choices, most application developers chose to develop only for Windows. In the amount of time it would take MIT to evaluate and select the right Java application server, V1.0 of a system could be built and launched using the Microsoft tools.

Anyway, it wasn't my decision. I'm a humble volunteer teaching one class out of nearly 1000 classes. The MIT bigwigs don't come down through the ranks to ask for my advice and I don't offer it unsolicited. But certainly if the decision makers came to my 6.171 classroom they would not find any support for using J2EE. About half the class elected to use the Microsoft .NET tools. The other half of the class is using simple Java Server Pages. Not a single graduate of the course will be experienced with programming in any J2EE environment. So if MIT wants something that can be maintained and extended by its students and graduates, Microsoft .NET is a logical choice for that reason as well.

(Your security question is beyond my area of expertise. I guess I'd be prepared to believe that both WinXP and Unix/Linux will require substantial sysadmin investments to be kept secure and that neither can be left unattended.)

You raise some good points about Microsoft licensing fees but they don't apply to MIT. We get most things for free or cheap. We don't need too many CPUs because we have fewer than 10,000 students. It is not as though we need to rack and stack thousands of machines. If I were building something that I wanted to give away to every organization in a Third World country I'd ask myself if a Microsoft world in which the basic software costs more than the hardware made sense. But if I were building something to sell to big companies, I'd build it on Microsoft .NET because I could be sure that they would have Windows servers and even if I liked J2EE I would never be able to know in advance which variant of J2EE they'd adopted.

-- Philip Greenspun, April 11, 2002

Ben:

In response to your message below... The J2EE zealots seem to think that a two-tier straightforward and understandable JSP site is not "true J2EE". Jin Choi and cohorts did a nice JSP version of ACS 3.4 and people complained that it wasn't "sufficiently J2EE" because you could still recognize some fragments of SQL and HTML in the code.

The vast array of choices that you see as an advantage for J2EE I see as a disadvantage. If I were paying a group of programmers, I'd want them solving my users' problems, not evaluating different "operating systems, development environments, databases, web servers, servlet engines, and application servers". The end user can perceive only data model and page flow. I would want all of my programmers' energy focussed on these challenges and not on the tools. My students who elected to use Microsoft .NET tools spent an average of two hours setting up their systems. Those students who elected to go the J2EE route wasted two weeks and ended up having to drop the course because they couldn't get their application servers up and running (despite the fact that all of these students had at least one semester of Java programming experience).

As for your long run argument, Ben... I noted that J2EE did not seem viable for the long run due to the plethora of incompatible systems. But I forgot to mention that the most serious problem for J2EE's long-run viability: the Java language. The idea that every programmer in a large organization will want to use the same computer language seems fanciful. IBM tried, in the late 1960s, to browbeat everyone into using PL/I. It was more widely adopted than Java but never crushed Fortran and Cobol like it was supposed to. Today, 20+ years after PL/I's peak, it is tough to find anyone who even remembers the language.

Why won't everyone want to program in Java? The novices will want scripting languages and declarative languages. The experts will want to program with features from modern computer languages (Eiffel, Lisp, ML). The speed freaks and hardware nerds will want languages like C. The Microsoft tools (IIS, Active Server Pages, .NET runtime) aren't tied to any particular computer language and recognize that an organization will have a diversity of programmers and therefore a diversity of programming languages.

[All of this ignores the fact that Sun greatly hastened the demise of the Java language by picking a fight with Microsoft (Sun's 1997 lawsuit, for example). They're now in the position of saying "Java is the world standard computer language except that you can't really use it to develop an application for computers running Microsoft Windows, i.e., for 85 percent of the world's computers."]

A theoretical discussion about programming tools can be entertaining for a few minutes. But here we are in the 8th year of Java and it should be easy to judge the tool by results. Where are the great sites that have been built with Java? On time and under budget? You see a lot of nice Perl sites, e.g., Slashdot. You see a lot of nice VB/ASP sites, e.g., Netflix. You see some weird things, such as Amazon (C plus proprietary scripting language) and Orbitz (Common Lisp back end; see http://www.paulgraham.com/paulgraham/carl.html). Even Sun's marketing staff can't come up with anything interesting built with Java: see http://java.sun.com/products/jsp/success.html and http://java.sun.com/products/servlet/success.html .

Oh well, that came out sort of long and negative. I guess I should close on a positive note. From a brief scan of their marketing PowerPoints, it looks as though the Java/J2EE folks have realized that interfacing to other languages will be necessary. They want everyone to learn and use CORBA! That should provide us all with many years of entertainment.

-- Philip Greenspun, April 20, 2002

Philip,

I think you are throwing out the baby with the bathwater. Many companies use J2EE but don't use EJB persistence. EJB persistence isn't the solution to everything in the Java world. It makes sense when you have a piece of functionality that you would like to package and sell to many customers or developers without dictating their choice of database. MIT would not need that, and they could create SQL manually and either create a data access layer or just embed the SQL directly into the JSP. So the performance would be comparable to .NET.

Java allows you to have many options in building an application. You have choices for operating system, development environments, databases, web servers, servlet engines, and full-blown J2EE Applicaton Servers. It's possible to choose a set of these that performs badly as you described, or efficiently like the original ACS. Using .NET limits the choices considerably.

Java also has excellent security. .NET supports multiple languages and allows developers to make mistakes that are not possible with Java. You can run Java applications with open source Apache and Tomcat, which have a better track record for security than Microsoft.

In the long run, the competition (both closed and open source) in the Java arena will lead to better products than Microsoft will be able to offer.

-- Ben Ballard, April 16, 2002

As an example of a nice Perl site, Philip gives Slashdot. This is a rather arguable example, though I personally like it, many people don't.

As the local mod_perl zealot, I'd point out http://www.imdb.com/ as the ultimate Perl site.

I also agree that Java/J2EE going to CORBA should provide us with plenty of entertainment.

-- Pierre Phaneuf, April 20, 2002

Philip,

I see your point, and I think there is a lot of sense in it. My company's product is written with Java, and since working here I have become most proficient in that language. Aside from my own preferences, I have a lot of respect for Bill Joy, James Gosling, and the other folks who created Java, and from what I understand, they solved a lot of the problems of C and C++ with the Java language. For many companies, the lower performance is worth it to have more reliable software that was developed in less time. Also one of these days they may fulfill the original purpose of Oak, the predecessor to Java, and there will be many little embedded devices that people will be able to develop software for in Java.

That being said, I'm not crazy for J2EE, and would prefer to create 2-tier or 1-tier applications around Oracle. I like to write my own SQL, and keep the interface simple. I want both speed and reliability.

Sun has made some serious mistakes. In addition to starting an unproductive war with Microsoft, they've pissed off a lot of their best users by not giving it away as open-source, and not cooperating fully with groups like Apache Jakarta. It will be interesting to see what happens in the next few years. I wouldn't mind if .NET and Java become acceptable options that can interoperate.

-- Ben Ballard, April 23, 2002

In my opinion, Microsoft programming tools are a lot like fast food on a road trip. You are in a rush and it is easy to pick up and you seem to initially save time. Near the end you feel bloated and sick(that not to say proprietary solutions from Sun are _that_ much better). I hope that Microsoft is unable to destroy the World Wide Web via its .NET initiative. .NET is simply a proprietary client server application that wraps around web standards (xml, html, soap etc).
I wrote a little rant on what I think about IE only sites.
Anthony

-- Anthony Barker, July 10, 2002

Philip mentioned in his response that half the class chooses ASP.NET and the other half chooses JSPs. I wonder what the percentage is that choose alternatives like Perl, Python, PHP, Tcl, etc. However, in the Internet Application Workbook Philip suggested not to dismiss these scripting languages. I wonder what percentage of the course chooses such an alternative route and why they tend to shy away from it so much. Are there any overwhelming reasons?

-- Kendrick Hang, September 15, 2003

choosing asp.net has many advantages some of them I describe here. In asp.net we can reduce the code. Asp.net -ide has importent tools for creating server side controls and easy for validation control. Developer don't need to learn so difficult coding as in J2EE. Asp.net support event-driven programming in which developer can easy understand and debug the coding in less time. One thing of using asp.net is that asp.net is not so case-sensitive as java. But there is one major disadvantage of using asp.net is that asp.net tied you up in Micrisoft Environment.

-- jitendra kumar prajapati, March 10, 2004

I've had the opportunity to develop applications in both .NET and J2EE. The main advantage I observed about .NET was that a shorter development time could be accomplished. For example:

With EJBs, if any change were made to a class file or EJB in particular, it would require that the entire application be repackaged and redeployed. With some application servers, you are able to skip the step of repacking by working in an exploded environment, but redeployment or restarting the app server was always necessary. At least this was the case with JBOSS, WebLogic and Oracle 9iAS. Depending on hardware the entire process can take up to 3 minutes. Considering that 1 line of code change will cost you 3 minutes to deploy, it all adds up to much wasted time.

Another advantage of ASP.NET to J2EE is that everything is tightly knit. With J2EE, the benefit of being able to pick and choose your components comes with a cost. I would not be surprised if students were unable to setup the J2EE environment in 2 weeks. We had oracle consultants come in for 2 days and leave with no clue of why the environment was so unstable. Application servers are often tested with outdated specifications. Very few application servers are ready to support the latest technologies that IT managers and developers want to use. It is true that MS does not offer the flexibility that J2EE offers. But at the same time it lends it self to fewer headaches.

Performance is another advantage of the .NET platform.

The biggest advantage I see in J2EE is cost. Because of its nature, and the fact that we are able to pick and choose, we are able to pick the environment we see fit. It is possible to develop a full blown J2EE application without ever spending a penny on application server software, operating systems, and databases. Open source does come with its costs in terms of “no support or service”. But very often, you will find that the open source community can help you resolve most if not all the issues you may run into. I’m not going to argue for or against open source as it would be out of the scope of this thread.

J2EE is not susceptible, or at least not as, to vendor lock-in. It was mentioned that only 1% of code needs to be changed, but that 1% is not always obvious. Although IÂ’ve never actually ported an application to one vendorÂ’s server to another, we did come very close. We sat down and evaluated what would be involved, and found that no code changes were necessary. The only files that needed to be altered were 2 xml files. The ability to move an application from a $100k server to an open source server is a huge deal.

In conclusion, I canÂ’t argue that one platform dominates the other. I think the two technologies cater to different but similiar group of consumers. They should be able to co-exist for quite some time. For purposes of marketing yourself in your career, IÂ’d say that it doesnÂ’t hurt to know both. Using either of these technologies in a project will require careful evaluation,

-- Vincent Wong, March 12, 2004