back to main

• Teen Grid
• Tips & Tricks
• Civic Center
• Developer Focus
• Fab Five
• As We See It
• Your Letters

As We See It

New Features vs. Bug Fixing: A defense of features

Recently, a Resident approached me about the Features vs. Fixes debate. And although I’m prodded about this issue on a regular basis, something about the way in which said Resident framed the topic caught my attention. He asked, “What’s the right balance?” Thinking about features and fixes in this light helped me draw certain conclusions and clarify why I support introducing new features despite the fact that Second Life still has bugs, many of them major.

New Features vs. Bug Fixing Juxtaposition

It is my view that the New Features vs. Bug Fixing juxtaposition misunderstands the issue and unfairly totalizes the detriment of features to the platform, in part because many new features are designed to optimize overarching performance. WindLight is an excellent example of feature balance. On the surface, WindLight appears to be largely an aesthetically oriented feature. While WindLight inhabited the First Look client, Residents noted that performance was slightly but never-the-less adversely affected. Linden Lab concurred, and WindLight was taken off the test grid for substantial “under-the-hood” maintenance. Since then, we’ve made great strides in improving WindLight’s stability, and to demonstrate further our commitment to quality, Linden Lab is not planning to rush WindLight into the production viewer. Furthermore, a critical view of WindLight indicates that the feature also embodies advanced and specific rendering optimizations. Check out Steve Linden’s Render Roadmap on the wiki for more information.

WindLight illuminates how features are not binary, not good or bad, performance cripplers or enhancers. Instead, it is my view that Linden Lab strives for a carefully considered balance when developing and releasing new features. Ideally, each new feature is characterized by not only an overt addition to the client, but also an improvement in performance and stability.

What about the bugs?

Bugs are important. Performance and stability is essential. Partly in response to your concerns, Linden Lab has created and continues to host weekly general and specialized bug triages. Don’t let the name fool you though. Resident feature requests are also investigated. Benjamin Linden hosts the weekly User Experience Bug Triage. Bridie Linden hosts the weekly Viewer Crashes Triage. Iridium Linden hosts the monthly Internationalization Triage. Rob Linden, out of whom these triages and JIRA were born, hosts weekly General Bug Triages. Bug triages ensure that both features and bugs that are crucial to you come to Linden attention.

Open Source plays a role in our ability to bring bug fixes to the Second Life Community. Your contributions of not only patches, but also reliable bug reproductions have enabled us to tackle bugs faster and more effectively. It bears repeating that Cory Linden held a town hall in May to discuss stability and performance. During that town hall, Cory noted that nearly 70 percent of Linden Lab developers focus exclusively on bugs and scaling. Those developers are working on both internal and public JIRA input (privileged by vote, but not exclusively) to improve stability and adhere to your bug-fixing demands.

It is my claim that the binary of New Features vs. Bug Fixes fails to provide a rich account of features, as our example of WindLight has shown. Furthermore, Linden Lab is committed to fixing bugs and will continue to dedicate the vast majority of developer resources to such fixes. However, we believe that releasing new features is crucial to Second Life, if we can ensure that they will both enhance the platform itself and increase performance and stability.

Agree? Disagree? Let us know - send us your Letter to the Editor - with your Second Life username - and we will publish a representative set in the next edition of The Second Opinion.