特性管理

Software powers the world. LaunchDarkly empowers all teams to deliver and control their software. Theoretically, teams can engage in all four pillars of feature management without relying on a dedicated feature management tool. In many cases, engineers will construct a patchwork solution for feature flag management in-house. Some organizations, however, do build and use world-class homegrown systems. But in our experience, the latter is the exception. Unsurprisingly, LaunchDarkly contends that, in the long-run, a commercial feature management platform is more affordable, sustainable, and painless than the alternative. You can explore the pros and cons of building vs. buying a feature management solution here . ...

What is Learn? The Learn pillar of feature management focuses on enabling all teams to better understand how software changes affect users and systems. It is designed for the entire product delivery team: developers , DevOps and site reliability engineers (SREs) , and product managers (PMs). Of all the pillars, Learn is especially concerned with helping software orgs become more data-driven. As a part of this, it provides mechanisms for engaging in hypothesis-driven development, an approach that embeds the scientific method and experimentation into the process of building software. In this way, Learn also supports the “relentless pursuit of continuous improvement,” a key principle of Continuous Delivery . ...

What is feature management? Feature management is a new class of software development tools and techniques powered by feature flags. A feature flag is a lever of control within your code (an if-else statement) that decouples code deployments from feature releases. Developers have used some variation of feature flags for years. But when it comes to enjoying the full benefits of feature flags, many have only scratched the surface. At most organizations, the art of feature flagging is confined to just a few teams across a few use cases. Such limits stem, in part, from the fact that managing feature flags at scale is quite challenging. For one thing, technical debt can rapidly accrue if you start using flags in large volumes. As a result, only a handful of developers end up using them, thus limiting the value an organization can capture from those flags (see what we did there?). Another drawback of conventional flags is they only support basic true-false, or Boolean, use cases. This, too, prevents organizations from realizing the full benefits of feature flags. ...