My 2007 predictions

I'm so sorry! My prediction for the third millennium is the Informational Individual! See www.geocities.com/gmatei2000

The browser back button is a sticky issue for designers and developers. Quite a bit of User Experience is compromised due to the 'back button'. I wish there was a way to programaticaly disable the browser back button when needed.

Rauf

"1. Open sourcing Java won't matter - it's a non-event." I disagree. GPL'd Java will encourage both more community participation (rather than the very corporate oriented JCP), and community innovation, and it will encourage Linux distros to distribute it by default. The latter will help accelerate adoption of both growing, awesome technologies. Linux and Java will feed off each other's momentum.

"2. Ruby and Ruby on Rails won't make it in 2007 either. I still do not see a compelling reason to switch." You don't see a compelling reason to switch, but obviously many people do, considering the large amount of enthusiasm RoR generates (sure a lot of it is hype, but a lot is real and genuine). And the compelling reason many see for switching? RAD/ease of development. RoR certainly won't replace JEE, PHP, .Net, or whatever else. But it will continue to grow, and generate enthusiasm. After all the hype has died down, we'll get a clearer picture of actual usage and actual growth, which will be much less than the hype, but still very healthy.

"3. Ajax hype is stronger than I thought mainly because of the life support offered by frameworks like GWT. But still, I'm not going to recommend enterprise IT shops make any serious investments in AJAX." Yup, Ajax is definitely way over-hyped, and arguably over-used. But it is an easy, cheap (completely free), and open way of making web interfaces more interactive. The hype will die down, but Ajax usage will continue to grow.

"4. We are going to watch some interesting competition in the RIA arena between Adobe's Flex and Microsoft's WPF/E. Adobe has more mature technology, while Microsoft is an established player among enterprise developers. I won't be surprised if Adobe will dramatically drop the licensing fees for their Flex Data Services." IMHO, RIA will go more towards Ajax, DHTML, OpenLazslo, or other open, free fully cross platform solutions. When one is doing something that is web based, why lock oneself into a proprietary, not fully cross platform technology? Why use Flash based Flex, no matter how slick it might be, when the majority of people absolutely detest Flash, and it's a massive CPU hog? Oh and by the way, FlexBuilder is horribly expensive. Why use MS WPF/E, and get locked in to MS? When there are totally free and open, and easy and viable alternatives, there's no compelling reason to use Flex or WPF/E, no mater how slick those might be. The free alternatives are slick enough for most people.

"5. Java remains the best choice for the server side enterprise development," Agreed, especially with EJB 3.0 and Spring and Hibernate around, JEE is as strong as ever.

" but it won't be able to compete on the desktop." True in the past, but that is steadily changing. With NetBeans' Matisse, Eclipse RCP (which, ironically, FlexBuilder is built on), SWT, Swing optimization, and now Java 6, which is significantly faster and Swing looks better, Java on the desktop is looking more viable than ever. And the steady entries into "Swing Sightings", and the listings at Eclipse RCP, is very strong evidence that the uptake of Java on the desktop is very very healthy indeed.

"6. IT outsourcing remains a part of our lives despite (or because of) the poor management by American corporate IT staff, and the reason is not the lower labor cost of overseas programmers, but the abcence of programmers in the USA." All true, and the foreign outsourcing phenomenon has come full circle.

"7. Switch from plain stateless text-based to rich internet applications will slowly continue. But this won't be an easy process - it's not that easy to get rid of these annoying but familiar habits of dealing with one-page-at-a-time applications. The fight for the Back button on the Web browser will continue." Yup, that's true. And the majority of it will go to Ajax/DHTML, not Flex or WPF/E

"8. I'm not going to afford an early retirement. Let me go and buy this lottery ticket for tonight's mega millions..." Join the club. :-)

Jeff, I'm glad I did not do too bad - you agree with me on 5.5 out of 8 predictions :) So these are my quick answers to the first three of your comments.

1. Java source code was available from the day 1. Sun was not too responsive with bug fixes, amd this is the only advantage I can think of in opensourcing Java. But this comes at a price, namely forking. If I will fix a bug in my version of Java, how fast are you going to get it in yours? Do not forget that approval of any (even minor release) of Java is a process. Community participation was always excellent in Java even without GPL availability.

2. I'm not sure I made myself clear. I always talk about enterprise development. I have no dobt that Ruby and RoR are very excited products to work with, but I'm having hard time understanding why any decent size enterprise would stop doing development, say in Java and start working with RoR.

3. Is locking yourself to a proprietary technology supported by a strong company a bad thing? Does everything have to be open sourced? Ajax will find its use in prettyfing Web applicatoins, but again, I would not start development of the serious enterprise application in Ajax just because it's free. Do not forget, that beside licensing fees someone has to pay salaries of Ajax programmers, and given the fact that Ajax development is slow, the savings on the licensing will be eaten up pretty fast by additional labor cost.

"Ruby and Ruby on Rails won't make it in 2007 either"

hahahahahahaha

you don't have a clue what you talking about!!!

go here and see your self

http://wiki.rubyonrails.org/rails/pages/RealWorldUsagePage1

"Jeff, I'm glad I did not do too bad - you agree with me on 5.5 out of 8 predictions :) So these are my quick answers to the first three of your comments."

I like reacting to your columns from time to time. It's fun.

"Community participation was always excellent in Java even without GPL availability."

True, but the biggest influences were always Sun, Oracle, IBM, etc, thus the existence of the EJB 2.x monstrosity. Then along come Spring and Hibernate, completely independent, open source alternatives, and driven by developers in the trenches, rather than corporate big wigs, and the term "POJO" becomes common in Java development vernacular, and then the EJB 3.0 specification becomes an imitation of Spring/Hibernate POJO frameworks. In other words, community involvement drove a very positive change in the EJB, and the JEE overall specifications. The GPL will further encourage "grass roots" involvement, and that can only be a good thing.

"2. I'm not sure I made myself clear. I always talk about enterprise development. I have no dobt that Ruby and RoR are very excited products to work with, but I'm having hard time understanding why any decent size enterprise would stop doing development, say in Java and start working with RoR."

RoR won't make any kind of significant impact in the large enterprise. It's probably ill-equipped, unproven, and is definitely unsupported (on a corporate scale). But in smaller, independent situations, RoR will continue to grow. So, even though it's not as interesting ;-), I think we're in agreement here.

"3. Is locking yourself to a proprietary technology supported by a strong company a bad thing?"

It most certainly can be. Being locked into one vendor, no matter how good that vendor might be, makes it so you can no longer choose "best of breed", because you're stuck with only one breed. It also means that the vendor no longer has to really work to retain your patronage. Just look at what happened with Internet Explorer, before Firefox really took off. It was essentially a dead product from a development standpoint. Then Firefox made a big impact, and MS woke up and actually tried to improve IE (and IE7 is an improvement). In short, when vendors compete, we (developers and customers) win.

"Does everything have to be open sourced?"

Not at all. In fact in many cases I hold the opinion that proprietary solutions are superior. However, software that is based on open standards, but is not necessarily open source (but it can be), is the main thing that avoids proprietary lock-in. Also, IMHO, fully cross platform solutions should always be preferred. This gives the customer/end user the ability to use/deploy their technology investment across a variety of other technology investments, i.e. it provides flexibility, and, in turn, helps yet again avoid vendor lock in. All of this increases overall value, and minimizes overall costs.

"Do not forget, that beside licensing fees someone has to pay salaries of Ajax programmers"

True, licensing fees are only part of the picture of TCO. However, being that Ajax is simply built on JavaScript and the DOM, with HTML, I think it's probably easier to find Ajax programmers than Flex programmers or WFP/E programmers. Also, the fact that Ajax technology is obtainable for free, assures that there will me more developers playing with it. The fact that FlexBuilder is very expensive (yes, the sdk is free, but involves more a time investment), ensures that fewer developers will "play with" the technology. No, just becaue Ajax is free, does not mean it will be the best solution. But it does ensure that it is accessible.

"and given the fact that Ajax development is slow"

Well, if someone just starts coding raw JavaScript, with the DOM and the XMLHTTPRequest object, yup, it would be pretty slow and tedious. But with the plethora of both free/open source and proprietary frameworks out there, Ajax development becomes much much easier. Heck, there's even GWT, which allows Java devs to compile Java source to Ajax Javascript and HTML. Simple.

Now, don't get me wrong. I don't think Ajax is the be all to end all, by any means. It's just a nice set of tools to add some RIA qualities, with minimal investment, to otherwise static web pages.

Open Source Java: Having access to source code like Java Web Start helped us understand how it handles certificates.

AJAX: AJAX is useful for those IT tools but for medium to large enterprise development, it just doesn't cut it.

From what I understand, please correct me if I'm wrong you are not a big fan of Ajax because you have to program in JavaScript. This is a quote from another post:

"..Having the need to execute ten pounds of code of any interpreted/platform dependent language in a non-standardized environment such as a Web browser is bad"

I can agree with this.

But if you use GWT or JSF enabled AJAX components, you actually don't see any JavaScript at all. But you are still not a fan of Ajax. Why?

>you actually don't see JavaScript at all

I do not see, but it's deployed and is running in prod. It's being delivered over the wire to the user, it has to be debugged, compressed, obfuscated... GWT makes the life of developer easier, and improves the quality of programming, but in the end, it's still a JavaScript app.

One more note, if JavaScript is such a good language, why do you think each of these great AJAX frameworks start their marketing materials with the same phrase: "With our framework you won't need to program in JavaScript"?

>>it has to be debugged, compressed, obfuscated... GWT makes the life of developer easier, and improves the >>quality of programming, but in the end, it's still a JavaScript app.

This is very true, but you have a few experts that know how to do this, they are good at this and they create these components.

>>One more note, if JavaScript is such a good language, why do you think each of these great AJAX frameworks >>start their marketing materials with the same phrase: "With our framework you won't need to program in >>JavaScript"?

I agree with you that JavaScript (in AJAX context) is complicated but JSF components hide the complexity. You don't need to worry or even care what is being rendered to the client.

Here's an interesting blog touching on the usability of RoR in the enterprise development: http://blog.citizenduck.com/2006/12/18/java-ee-vs-ror-vs-net/

One of the reasons we've moved to RoR for my current project is the framework's built-in AJAX support. There is some synergy there--RoR should further AJAX use and vice-versa.

"Does everything have to be open sourced?"

No.

But it helps.

"The browser back button is a sticky issue for designers and developers. Quite a bit of User Experience is compromised due to the 'back button'. I wish there was a way to programaticaly disable the browser back button when needed."

The back button is a good thing for users and it isn't going away anytime soon (nor should it). Frameworks like the Wicket web framework, which is stateful and OO, have been designed to handle the back button seamlessly. In Wicket components are versioned as they change from request to request so they can be reverted to their render-time state automatically when the back button is used. The result is that the back button "just works" transparently with Wicket.

Prediction 4. Agreed there will be a battle for RIA by Adobe and MS. However, expect other options, including those based on Java to come up.

Agreed, Eclipse RCP + Swing on top to fill in where JFace is weak is the way to go. Flex can't compete w/ all the benefits of Eclipse RCP, and your development time is much faster if you build on Eclipse RCP and do Swing for effects later, versus get free effects from Swing and try to build the infrastructure that RCP provides(we did an extended proof of concept to validate this). It's almost laughable that Flex tries to say their a java desktop replacement space yet their IDE is built on Eclipse RCP!!! I'm not passing judgement on Apollo yet...will be interested to see what it brings to the RIA competition.