Neal Ford about Software and learn from history

There are already wild speculations about the what, why, and how on the Internet, but facts are pretty rare. Facts are that the OS will be launched next year, 2010, it will be Open Source (yeah!), and the community will be involved soon. Google Chrome OS will focus on mini netbooks and desktops. So there is an overlap with Android systems, but as stated in the Google blog post, they have different goals. Google Chrome OS (the name is a bit long I think) will focus on Web applications and performance. One thing is clear, when Google talks about performance, they are quite serious about it.

So it is longer only OS X vs Windows, odds are good will leverage Linux to worthy rival.
TimeWhale Time Tracking

I start with an overview of the new features and then with my personal opinion about the new version. Disclaimer: I am only writing my personal and subjective opinion, I didn’t do any performance tests, measurements, or statistics analysis. So if you have some objections, just leave a comment.

Before we start, I will give you an overview about the new functionality in NB. The slogan for the new release is “Connects Developers” (the slogan for NB 6.5 was “The only IDE you need”). I am not sure if the slogan is a one because only one new feature, the Kenai integration, has something to do with connecting developers:

Connected Developer Tools

• Create Kenai-hosted open source projects directly within the IDE
• Manage your Kenai project and collaborate with like-minded developers
• File issue tracking reports using both Jira and Bugzilla

Java

• Hudson services integration for Ant and Maven projects
• Generated source roots now displayed in the Projects tab
• Struts library migrated to version 1.3.8

Maven

• Library dependency graph viewer
• Improved code completion for Maven plugin parameters
• Support for Java EE, Web Services creation and consumption
• POM Editor and Navigator enhancements

Groovy and Grails

• Out-of-the-box support for Grails 1.1
• Code completion
• Disable browser opening on run and configure platforms to use with Grails projects
• Invoke Grails command or custom script from context menu item

Ruby and Rails

• Remote debugging
• Improvements to Ruby constants support
• Run and debug actions for test cases and suites included in context menu
• Support for Shoulda tests

These are only the most obvious features of NB 6.7, for a full list have a look at the release notes.

So what’s my view on NB 6.7? One thing that stand out, the NB team concentrated on a lot of different things, but not on Java. Of course most technologies are somewhat in the Java realm, but not specific Java. I tried a couple of times to switch my Java development to NB, after all using only one IDE would be nice. However, the Java code editor in NB always felt kind of awkward. In Eclipse I have the feeling the code editor ‘knows’ what I want to do. I mean the NB code editor is good in general, but has a lot of rough edges. Here is an example of what I mean:

NetBeans strikes you down with suggestions, and Eclipse just shows suggestions which are valid to the context.

Another case against Java in NB are missing plugins, e.g. FindBugs. I know there is NB plugin for FindBugs, but is seems fairly outdated. Also tool integration from other companies, e.g. Google GWT and Google APP, doesn’t exist in NB. So I guess NB 6.7 for Java is still a not for me.

Project Kenai is Sun’s Google Code with some extra features for issue tracking and collaboration. Although, it hosts some interesting projects, like JRuby, it is fairly unknown. The new tools allow you to create Kenai hosted open source projects, to collaborate with other developers, and track issues. I played with the tools a little bit and found them ok, but not really exciting. Equal tools are already around.

NetBeans focuses more and more on dynamic languages, like Ruby, PHP, or Groovy. Groovy and Grails are getting really pushed in NetBeans right now. I did some work with both and like them. That said, I am confused about Sun’s strategy. On one hand they push Ruby with JRuby, a lot, on the other hand their only IDE switched focus to Groovy. This is not a problem per se if the Ruby support is already that good, but it still has issues and sadly performance got worse in NB 6.7. My guess is because Groovy is popular in Europe and the core NB team is also in Europe, there is maybe some sympathy towards Groovy.
Supporting a lot of languages is great, but I am not sure if the NetBeans team has enough ressources to fully support all the languages they want. With all the downcuts and the still uncertain future of the Oracle deal, it would better to support less and let the community/partner do the rest. However, getting the community and partner on board was always Sun’s (NetBean’s) weak spot. This is were Eclipse clearly won. A note on the NetBeans RCP, Matt Thompson, Senior Director of Sun’s Developer Cloud Tools, stated the the RCP is now less of an goal. This could backfire when e4 is around.

Like I said I use NetBeans primarily for Ruby on Rails development. NetBeans was and still is the best Ruby on Rails IDE, but it lost its momentum and other IDE aren’t sleeping. The only useful new features for me is the support of the Shoulda framework. Should is an awesome Unit test framework for Rails. Useful, but for me not a feature, is the improved constant support. The rest is like NetBeans 6.5, only slower (and sometimes buggier). A couple of weeks ago NetBeans 6.7 had some serious performance problems (we discussed this in the CAT program). The NB team invested a lot of hard work and fixed many of the problems. But NB 6.7 still feels sluggish compared to NB 6.5 and doesn’t do good on scanning large projects.
Another NB Ruby/Rails caveat, NetBeans includes an outdated version of JRuby, 1.2. The actual version is 1.3.1. So it is probably best to install your own JRuby version and don’t rely on the provided one.

Conclusio

Is NetBeans 6.7 bad? No. Is NetBeans 6.7 good? Maybe, but more like ‘ok’. NetBeans 6.7 gives me the feeling that it includes a lot of features the majority doesn’t need or use (e.g. features on demand which was at the beginning a big deal and now seems just a side feature), and doesn’t care enough about improving existing ones and performance. The results of the NetBeans community survey support my thoughts. For the next version I would wish something like Snow Leopard or Windows 7, a release just to clean and speed up the base.

After this rather neutral/negative review, I want to thank the NetBeans team for their hard work and community involvement. The CAT program allowed me and others to speak up their minds and the NetBeans team listened. As a developer I know that there are a lot decisions and considerations (or politics) the user, which is in this case me, simply doesn’t see. Thanks for NetBeans 6.7 and I am looking forward to the next version.TimeWhale Time Tracking

We are currently busy with testing and tweaking the application, but we are already thinking of future features. So please, if you have any ideas and suggestions, drop us your thought at http://feedback.timewhale.com.

Last, but definitely not least, a big thank you to all our Beta participants. We hope you are having a whale of a time with TimeWhale.TimeWhale Time Tracking

With this release, the Java 1.4 support has been dropped. The highlights of the release are:

• Fixes so that JRuby runs in restricted environments better like GAE/J
• Accessing primitive Java arrays are about 10x faster
• timeout.rb is now ~40% faster
• Method cache performance improvements
• irb works in –1.9 mode now
• Additional Miscellaneous 1.9 fixes
• rubygems 1.3.3, rake 0.8.7, and rspec 1.2.6 upgrades
• 66 bugs fixed since 1.2.0

More infos hereTimeWhale Time Tracking

Jim Weaver, a JavaFX enthusiast, already blogged about the new version, as did Stephen Chin. Especially Stephen’s post a good overview what to expect from the new version. JavaFX 1.2 is a big update to JavaFX 1.1. Here Stephen’s top ten feature list:

1. Skinnable UI Controls
2. New Layout Classes
3. Charting Support: Area, Bar, Bubble, Line, Pie, Scatter, X/Y
4. RSS/Atom Feed Support
5. Local Data Storage
6. Stage Infrastructure: Screen and Alerts
7. Improved Asynchronous Processing Model
8. BoxBlur effect
9. Math and Properties Classes
10. Direct Referencing of Java Arrays via nativearray

TimeWhale Time Tracking

It took us a little longer as expected, but we are proud to announce that TimeWhale is going to arrive soon. More specifically at the beginning of June (rumors are June 1st).  This first release will a be a beta version. This means that TimeWhale can still have some minor problems, e.g the layout is not perfect, translations may sound a little bumpy, or it simply has a “real” bug. That said, we are currently using TimeWhale at a daily basis and we haven’t encountered any serious problem yet. Further, we take security and safety very serious. Data backups and data encryption are just some of our precautions.

You are probably thinking now: Why should I use TimeWhale Beta and not simply wait for the official release? The (short) answer:

• Be there from the beginning. Help us to shape TimeWhale to a time tracking system we all love. Add your own ideas at TimeWhale feedback page and decide what should be next.
• Test it with no risk. In less than 5 minutes you have your own TimeWhale account. Try it and start tracking your time. If you don’t like TimeWhale (it won’t happen), just close the account. No hassle and no obligations.
• A service without users is useless. You, as a beta participants, are giving us your trust and believe in TimeWhale. We are really appreciating this and are planning additional benefits in the future.
• Last, but definitely not least: register for TimeWhale Beta and get TimeWhale the rest of the year for free! You can save up to 300,– USD.

More information about TimeWhale will be available when launched. We are looking forward to hear from you.TimeWhale Time Tracking

Disclaimer: The results are as they are. I didn’t validate them nor did I check their integrity. Please also consider that performance is just one criteria for hosting a web app.TimeWhale Time Tracking

Jonathan Schwarz, Sun’s CEO,writes that at JavaOne a new project, probably called Java Store (working title “Project Vector”), will be revealed. He writes about including the Google (although Google is not directly mentioned) toolbar in the Java installation process and how Sun made a bunch of money with it. Sun now wants to extend this concept and  to distribute  all kinds of third party applications via the Java update mechanism. Java is widely spread and installed on over a billion devices, hence, a huge potential customer base already exists.  At the moment facts are rare, as far as I know Jonathan’s blog post is the only source. So I guess we have to wait for JavaOne for more concrete information.

At the beginning I wrote about Apple’s and Google’s application stores. However, these I stores are for mobile devices and Java store aims at the desktop (at least at the moment, JavaFX mobile can change the picture). So my guess is Java Store will be (in the long run) more like a package management system, like APT or Ruby Gems. This would be great. Not only is Windows missing this from the beginning, if Java Store follows the write once run everywhere concept, Java Store could be expanded to Mac, Ubuntu, and so forth. Another huge advantage could be jar sharing. There are libraries which are used over and over again (e.g. Apache), with Java store the libraries could be shared between Java applications, similar to Linux application installations. Is this why Sun made its own Project Jigsaw instead of adopting OSGI?

This would be great news for users and developers. Will it happen? Probably not. I am still sceptical about the Sun/Oracle deal and Sun still has to convince developers to jump on the train (which is not a strength of Sun). Still, if it happens it could be just great.TimeWhale Time Tracking