Re: Smoothing the edges by example (was pilot case?)

On Wednesday, May 8, 2013 at 6:07 PM, Clint Hill wrote:

> I'd like to add a few things that I feel are important for getting more
> adoption. You could call these goals too I suppose.
> 
> 1. We need a better story for the average developer/engineer who is having
> trouble due to browser limitations but has a great idea. We need to
> educate this person that they should prollyfill.

Sure.  
> 2. For these average developers/engineers we need a very clear and very
> easy project template for their great idea. This template includes
> fixtures for tests, examples for documentation and all of the normally
> expected OSS artifacts. Yeoman generators are a good idea here. I believe
> that Chef has a great pattern for cookbooks and that we could provide a
> similar framework for developers to use in building their prollyfill.

Depends. If we want prollyfills to serve standards bodies, then we should probably look at using the same testing tools. This was kinda the point of using WebIDL - you get the stuff you need for a spec + you get tests generated for free (through webidlharness.js).  

Of course, one could side step all that and just focus on creating a great library. Remember that this library needs to work with other libraries, so putting in too much cruft can be harmful (remember how we got a bit screwed by using Grunt when they decided to break backwards compat). 
> 3. These developers/engineers should get this great idea published into
> the prollyfill registry for free. The said project template should
> automate the registration for them (I'm thinking of Ruby gems like
> Jeweler).

That would be great.  
> 
> 4. There needs to be an "engine" that runs the prollyfills that developers
> need not worry about. The responsibility of the engine is to execute the
> prollyfill for which ever target the developer has (JavaScript, CSS or
> HTML).

This would be nice, but it's quite complicated.  
> Now I've assumed many things in those steps [engine, registry, project
> template, code generators] but I think those are the things we fill in.
> Having these things in place makes the work of getting it in front of the
> right WG much easier and with some organization. I would assume most WG
> would be displeased if all the great ideas came at them in varying forms
> but are all called prollyfills.

Agree. That would kinda suck. But the main thing is to for a prollyfill to show the use cases in the wild.  
> It's my opinion that we could collect tons of experience and we can do the
> research to smooth out the edges - but it would all be for not if the
> above things don't get better and easier.

I'm open to trying stuff.

Received on Wednesday, 8 May 2013 19:19:24 UTC