Re: should we take it up?

I think the below is a good start, but raises a lot of questions to a TAG dork like me. The below is not criticisms - it's just questions I personally like answered if I was to read the awww section about poly/prollyfills.  

We can probably work together to address these questions… would also be nice just to start this on Wikipedia, as it would benifit more people than awww (http://en.wikipedia.org/wiki/Polyfill).    

On Monday, 1 July 2013 at 21:03, François REMY wrote:

>  
> The World Wide Web is evolving constantly,
"constantly evolving"  

Would need to clarify what "constantly evolving" means, in the context of awww - are you saying browsers (i.e., in concrete terms of new features, security fixes, etc.). If so, would be nice to give concrete examples, like "the Foo Web browser added X number of features from 200X-2013" and so on.  

However, what does it mean for the "Web" to be "evolving"? Are the documents all evolving? What percentage of them are changing? Are many remaining static? how many? i.e., what is the change differentiation in content over time, etc.    
> but some devices may be stuck
> with older User Agents for some time.

Stuck sounds kinda passive aggressive. Citing sources of the rage of browser updates would be useful here. And showing the problem for users (e.g., IE users in China use X browser, and haven't updated much since XXX… while users in South Korea can't update IE because of banking sector using Direct X plugins etc. - or so the urban legend goes … is it true? citation needed.)    
> With the new ECMAScript capabilities,

It's not just new ECMAScript capabilities - but the platform as a whole and the role ECMAScript plays in the architecture. It's not ECMAScript functionality that is being poly/prollyfilled most of the time (we can't prollyfill syntax, though we can pollyfill functionality a lot of the time… like with new array methods) - it's platform level features (e.g., new HTML tags, or DOM APIs)… or just stuff that people need that standards bodies didn't think of - or just never got around to doing properly (e.g., templates).   
> it's becoming an accepted practice

citation needed: Accepted by who? What number of websites use polyfills? Is jQuery a polyfill (yes, it polyfills stuff)? etc. How can it be shown that it's becoming more accepted over time?   
> to provide as fallback an author
> implementation of features implemented natively in newer user agents to
> older user agents by the mean of scripts.

Sentence could be improved a bit. Also, it's not just scripts - the prollyfills for picture I've seen also relied on markup (e.g, using <noscript> to prevent certain things from happening … so sometimes it's a mix of abuse of markup and script … or even abuse of CSS to achieve certain desired effects all working in tandem).  
> Those scripts are commonly
> refereed to as 'polyfills'.

Citation needed. Who coined the term? Was it:
http://remysharp.com/2010/10/08/what-is-a-polyfill/
> Sometimes there's a need to add to browsers features before they're ironed
> out

Avoid metaphors like "ironed out". Maybe "before participants in a standards-setting body reach consensus"… though some standards setting bodies don't use consensus.   
> by standards committee. The scripts implementing snapshots of unstable

"snapshots" is problematic - "unstable" even more so (see my oh so angry ramblings on the TAG list). Prollyfills may even be more ahead than standards (e.g., URL.js). My Web MIDI prollyfill follows the Editor's draft (living standard, not snapshot) and wilfully deviates where I don't agree with the spec in the hope to show the Editors that it would be better done another way, etc.    
> specifications are often referred to as 'prollyfills'.

Citation needed. Who coined the term?  
> It's important for Web APIs to provide the right amount of extensibility

What does this mean in practice (i.e., right amout)? Can you point to an API that exhibits these "extensible" or "prolly/polyfillable" properties/amounts? Can you show some that don't (and express why).    
> hooks to allow high-quality polyfills and help to move the web platform
> forward.

I what way? What examples can be cited where a poly/prollyfill has directly helped moved the platform forward? What does it mean to "move forward" for the platform? What is a measure of success (or progression)? Who benefits from this move and at what cost?   
  
--  
Marcos Caceres

Received on Monday, 1 July 2013 21:15:26 UTC