A plugin API is a double edged sword, and if I was in Apple's shoes right now, it's not something I would be willing to embrace at the moment. Xcode needs to be fairly dynamic while the toolchain evolves and grows. While Xcode 4 goes through the growing pains, an API that has to remain stable and documented becomes a boat anchor to a point in time, and in essence handcuffs the development team to a model that has seen radical change in the last 36 months.
When your release cycle is incremental upgrades every 2-4 years, you can have a public API and you can 'tweak' it with each release.
That is not the Xcode cycle, and hasn't been since Xcode replaced Project Builder. Xcode is still maturing, and as such tossing out that anchor would be a horrible idea, that in my opinion should be rejected by both Apple and the community at large. Would it be nice to have? sure. But at this point, we aren't to that level of maturity and stability in the product.
While there are things I see that Xcode could do better, the plugin API is so far down the list, that tackling it now would, IMO a complete waste of resources, that could be better spent improving the core product. Add to that, there are already facilities in place that enable a broad spectrum of automation if the developer simply takes the time to discover them. The amount of functionality that can be leveraged just through the template & snippet libraries is not trivial, and it is functionality that required 3rd party add-ons in other prominent IDE's.
Feel free to flame me, but do so off list.
--
Andy 'Dru' Satori
On Friday, March 23, 2012 at 3:22 PM, lbland wrote: