On Mon, Jul 29, 2013 at 1:28 PM, Gijs Kruitbosch <gijskruitbo...@gmail.com> wrote: > On 29/07/13 21:53 , Anne van Kesteren wrote: >> The current thinking is >> that offering developers the primitives will give us a better higher >> level API longer term. > > Isn't that reasoning part of why we are now in the position where we have > indexeddb and nobody uses it, instead preferring localstorage which is a > worse implementation but a better (that is, easier) API? I'm not convinced > that's the right way to go about solving the offline usecase.
Indeed. Somewhat off-topic for this thread, but I think this "let's provide primitives and let other people build higher-level libraries" trend for Web platform features is pretty dangerous. Having to "use a library" to solve fundamental Web platform use cases (like "work offline" or "store data") increases the barrier to entry for Web developers pretty significantly. I understand that designing high-level APIs can be hard and we won't always make the right decisions, but I don't think we get to make that someone else's problem if we really care about improving the Web platform in the near term. (There's a near-term vs. long-term distinction here too - "wait until the libraries sort it out and then put that in the platform" approach is costly if it means N additional months/years before the feature can actually be relied on by Web developers. I understand the desire to avoid backing ourselves into a bad compatibility corner and having to maintain a "bad" API forever, but I think that's sometimes a tradeoff worth making in order to drive progress.) Gavin _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
