In a World of Web Framework Choices, Some Developers Still Build Their Own

There's a ton of web frameworks in Java, but only a few stand out. If you're making your own project alone, you can just use whatever you wish, but if you're doing an enterprise solution or working in a team, especially one where there will be a cycling of ressources/developers, you only want to use something standard, mainstream, well known, etc.

Reason behing, even if you can save an extra 3-5% of time by using what you think is a better framework, it just takes 1 dev going on vacation or quitting to make you lose days or even months on training a new ressource. And even once trained, they'll never be better than what they have 7 years experience with. In the end, you'll lose out.

So stick to the mainstream, unless you have carefully evaluated the consequences of not doing so.

I love America (US), because America is a free place where people can do whatever they feel like doing as long as they don't hurt other people's freedom. This is as important to me as the air I breath. So if I want to create a web server in Java or a email client, or a todo list application, I WILL. And nothing or nobody will be able to stop me.

The god news is: I will not put a gun in your head to use it or like it. You are also free to ignore it or to take a look and not like it. Also feel free to throw your comments, to say it sucks.

I recommend a little read about Darwin and the Natural Selection theory. I also suggest checking out Cuba as a place to live for people that don't like competition or diversity.

By the way, I am very happy I did not listen to people when I created the Mentawai Web Framework 3 years ago. Today many people and companies around the world use and like it. Feel free to check our online book: book.mentaframework.org/, our cookbook: recipes.mentaframework.org and our website: www.mentaframework.org/

And keep in mind that if you don't like it, you can just ignore it. There is not need to feel bad or say sorry.

Sergio++

Research says that some choice is better than none, and that informed customers take advantage of more choice. And guess what, most of these frameworks are open source and their communities work in the public. All the transparency one would wish for!

It is fantastic to see all those frameworks that are started and take on the established ones. It means that people aren't settling for the status quo, but instead are constantly looking to improve things. Sure, you might have to spend more time keeping up (that's probably why you read this site) and figuring out what the best option is for you/ your project. But a defeatist attitude (oh, there's too much choice, let's just choose for the one that everyone agrees with is outdated) isn't helping anyone.

At this point it's just hurting. I really wish Spring and DI would really get into mainstream Java. At this point people still think of it as a web framework which is why I think people find themselves choosing Struts.

Frankly working in the average Java shop is depressing: Java 1.4x, Struts, and people coding this is all in vi. Of course php or something similar seems like an easy choice.

Less is definitely more. The hard and complex thing about Java is not the language, but the fact that there are so many choices to be made and no obvious ways thanks to Sun's devotion to complexity. It requires a lot of knowledge, skill and research to be successful (bet on the right horse). Every framework usually has a strong point but all in all, I find adopting a KISS route and avoiding frameworks alltogether (or at least, stick with very small and/or de-facto standards) is a win in the long term. The Oracle ADF stuff I wrote 3 years ago has now reached end of life for its tool support (JDeveloper & OAS) and is now effectively unmaintainable. The servlet micro-framework I am using now is so lightweight and simple that any Java developer can maintain it, and requires just a simple Tomcat. As much as I love the Java community, I am sure my job would be far less stressful and more productive if I was provided with fewer but stronger options, i.e. as the .NET world has it.

For the newcomer, too many choices may be overwhelming. But worse than having too many choices is to use too many things at once. This is called a Rube Goldberg Machine. EJB is the classic example. JSF also comes close.

To do a web project in Java is very "easy". All you have to do is use Struts, Spring, C3P0, Hibernate, JSTL, JAAS, Commons Email, Commons File Upload, Tiles, OSCache, etc. This is non-sense, we need FULL-STACK-HIGH-LEVEL solutions.

This is exactly what Mentawai, RoR and SEAM are trying to do.

So too many choices is something inescapable in a free world. I don't need to know all fruits to choose what I want to eat today. I don't need to know all grocery stores in my neighborhood to choose where I want to go today.

Take a look into the Mentawai philosophy. You may get impressed, but if you don't like it, just consider another grocery store in another corner.