[dojo-contributors] what is the raison detre for dojo 2.0.
mike at mikewilcox.net
Thu Oct 11 12:12:04 EDT 2012
I've never heard a dev say they don't use Dojo because the APIs aren't perfect. The fact is, Dojo has a much larger barrier of entry than just about any other library. That's the problem.
Here's the elephant in the room: I don't see a majority of devs picking up Dojo and doing a build. How many web devs are on Windows? 70% maybe? How many of them even touch the command line, much less embrace it?
The new build tool was sorely needed, is powerful, extendable, and generally a huge success. But it should only be considered the first step. The next step is an engine that creates the profile and package based on set criteria. The next step is to put this on the web so devs can download Dojo a la carte. Many other libraries have already surpassed us here (and there was a web builder tool n the works - did it get outdated by AMD?). A lot of devs still write all their JS in one file using functional programming, and this would fit their needs - and they wouldn't have to do a build.
The next step after that is a GUI, probably built with Adobe AIR, that leverages the build tool and, the dojo foundation package management, and CPM. The GUI tool can help the user get started on their project with a Dojo scaffold. It could auto-update the style sheets, and could even stub out modules with pre-selected dependencies.
Finally, I've mentioned this before, and Eugene and I talked about it again a few weeks ago - Dojo could help initiate users with a "learning mode". This would essentially be a certain type of instrumentation that gives hints to devs when they put pieces together wrong. And it wouldn't be any harder than the experimental or deprecated functions.
In the same vein, Dojo has a severe lack of feedback. Everything else has logging. The build tool has logging! Why does Dojo not have logging?
More information about the dojo-contributors