Roumen's Weblog

NetBeans 4.1 Release Candidate 1 is Out!

You can download NetBeans 4.1 RC1 here.

Final release will differ only slightly from Release Candidate 1. I recommend to upgrade to everybody who is using NetBeans 4.1 EA, EA2, beta, q-build or daily build, RC1 is definitely more stable. For users who prefer final releases: you probably want to wait till 4.1 final which will come in May. On the other hand if you have some issues with 4.0, you could try 4.1 RC1 if they were solved, there was a huge amount of bugfixes - see table below.

NetBeans Quality in Numbers

I've played with Issuezilla, our bug tracking software and got interesting and quite surprising numbers for bug fixes in NetBeans:

Version	Bugs fixed	Release date	No. of months	Avg. bugs/month fixed
NetBeans 4.1	2932	May 2005 (planned)	5 months	586 bugs/month
NetBeans 4.0	5001	Dec 2004	8 months	625 bugs/month
NetBeans 3.6	2662	Apr 2004	9 months	296 bugs/month
NetBeans 3.5	1157	Jun 2003	6 months	193 bugs/month

The numbers for 4.1 are for current status but there won't be many changes till fcs.

Important are the numbers on the right - average number of bugs fixed per month for each release. Wow, they are increasing, aren't they?

How to explain these numbers? If you would ask our competitors, they would probably say that the numbers of our bugs are increasing so the quality is going down. Ok, guys, thank you for your opinion, but this is clearly not the case. I am confident that quality of 4.1 is high, 4.1 is largely a stabilization release - except for new J2EE modules. We were testing the daily builds for months and they were very stable.

If the quality is good, how come that the number of fixed bugs is increasing? The remaining options are that we are adding lots of new features (we are, look at J2EE and J2ME support) and for other modules we are stabilizing the IDE. The other way to explain increasing number of bugfixes is increased focus on development of the IDE. My take is that it's a mix of these reasons - which is good.

I recently read in several blogs and articles why people should switch from Eclipse to NetBeans. I won't be telling anyone to switch, we are in a opensource space where everybody can hopefully make his own free decision. So I only recommend to take a look at NetBeans, it is improving and if you'll still miss anything, let us know through our mailling lists. We are listening. The competition in Java tools space is good, because the one who benefits from the competition the most is you - the developer.

UI Design for Programmers

I've just finished reading a book from Joel Spolsky called UI Design for Programmers. I can recommend this book to anybody working in software development, no matter what you are doing (as long as you are doing something). It is about designing software for real users - it won't teach you how to create good dialogs, but it will help you to think as a user. Why is that important? People working on software forget one important difference - you already know a lot about the program you create and care about it, while normal people know almost nothing about your program and don't care so much about it.

By reading the book you realize some hidden facts, like most people just don't read manuals - you knew that, didn't you? But it goes further, most people don't read descriptions in dialogs either - unless really necessary. So they just click on the error message you gave them, ingoring the content. Very bad users. But how many times did you click the button without actually reading what's in the dialog? Let's admit it, we all do it from time to time.

Joel is great in explaining simple principles - to understand them you don't need to have a degree in nuclear engineering. Still these principles are a bit surprising. By showing some of the annoyances of software we are using every day (web browsers, office suites, e-mail clients, etc.) you get a lot of knowledge of what *not* to do when designing software. Every programmer thinks that users like many options, because then the software they are making is greatly configurable. Wrong! Most users hate many options, because they don't understand them - they didn't spend all those months by designing the software as you did. They don't care if your program's database should be small, medium or extra large with double chips and hot mexican sauce. All they care about is that the program does the job they want from it.

What I also like on the book is it's size, it has 134 pages with a lot of pictures. Joel is always to the point, you read the book with one breath - it's very easy to read and amusing. Probably inspired by opensource, Joel published big part of the book online on his webpage here. I recommend checking out the other articles he wrote, there's a lot of truth about software development. I have to admit that Joel inspired me with his writing style and his posts were one of the impulses for me to start to blog. Thanks, Joel!