Skip to main content

RFEs vs. Bugs

2 replies [Last post]
Joined: 2003-06-28

Where do you think the most of the efforts in Mustang must be concentrated - on bugs or features? Fixed bugs make the platform more stable, but slow the progress of Java. We hear numerous requests for features, but how often do people start using new features when new release comes? What is the perfect combination of the amount of fixed bugs and features , for Mustang?

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Joined: 2003-08-24

Here is my wish-list from highest to lowest priority:

1) Fix the *social* problem: better communication between Sun engineers and end-users. discussion forums are an example of what we want more of. Ability of end-users to commit patches (for Bugs or RFEs) for Sun to include in the next release. Ability of end-users to include screenshots and other useful attachments when filing issues on BugParade.

2) JVM on-the-fly: Availability of a JVM that consists of Java Webstart and its dependencies and *nothing else*. The target size is under 2MB (10 mins of download time for dial-up users). If you have to make Webstart a native application instead of Java based, so be it. Just make this happen! As the user makes use of applications, Java Webstart downloads core classes "on the fly". The key here is that it should be possible to do the following:

I want to deploy a Java Webstart application, but not all end-users have Java installed. When someone clicks on the "Try now!" icon on the website it should do the following:

[ user clicks on icon ]
a) Check if user has Java Webstart installed, if so jump to step d
b) If not, pop up a dialog that reads: "The application you have selected requires Java Webstart to run. Please confirm that you wish to install Java Webstart and the application [OK/CANCEL]".
c) If the user clicks on OK, download and install Java Webstart
d) Download and install the JNLP application.

The majority of the world's users are using dial-up connections. The JVM userbase is very small amongst desktop users. Getting a user to install a desktop application almost always requires him to install the JVM first (because of the small userbase) and for dial-up users this is an hour-long download which is not acceptance. Targeting desktop users is impossible without this new JVM.

3) Focus on "big picture" solutions which will allow us to reduce the turnaround time between an issue being filed and being fixed. For open-source projects the average time is approximately a month for most issues. For Sun it is 2+ years.

Joined: 2003-08-24

BTW: I want to note, Sun wouldn't need to write any special code to make this happen. There are already numerous mature issue-tracking packages out there. JIRA is one such example. I am sure you could negotiate a very nice contract with them which would benefit both your organizations. They would get amazing advertising of their product in exchange for your ability to use it, etc.

Management decisions aside, you could have this up and running within a few days and spend another week or so migrating the preexisting BugParade issues into it.