Re: IndexedDB, what were the issues? How do we stop it from happening again?

On Thursday, 7 February 2013 at 15:43, Karl Dubost wrote:

>  
> Le 7 févr. 2013 à 10:28, Marcos Caceres a écrit :
> > I don't understand "made by the branding on the spec more than by the reality of interactions and sources"?  
>  
>  
>  
> People say: "Oh it's an ugly API! Yeah for sure, it is designed by $ORG. This is a proof that design by committee is bad."
>  
> What I'm saying is that a spec published at $ORG is orthogonal to the "design by committee" rant. Some will be edited by one person, some by a group, some will be good, some will be bad.
Sure - but I don't think anyone eluded that it was the W3C's fault (just that design by committee tends to create stuff that sucks - as opposed to beta/usability-testing the API with real devs - and yes, Mikeal dislikes standards organisations with a certain passion [1]). But the fact remains that people think the API sucks - perception matters. It's not a symptom of the W3C (as not all W3C APIs suck, I'm sure). There are things in the API that make it hard to use. Consider, this is supposed to be a "simple" example of how to use the API:

http://www.html5rocks.com/en/tutorials/indexeddb/todo/#toc-step2

I think that is the example Alex was referring to in his blog post.    
>  
> In the case of Web Alarm API, it seems it is a DRAFT (important to not) which has been inspired by a spec from one specific vendor without modifications. And… the 1st DRAFT. So why should it be perfect from the start?
No one is saying it should be. However, it's important to start addressing potential issues early (and especially if _every API in the WG is using the same pattern_). If those APIs get into the wild, then it's party over.      
> but I digress.  

>  
> Agreed with the rest of the email about indexedDB and identifying the patterns.

  
[1] http://marcosc.com/2012/10/dont-fuck-the-standards-bodies/  

Received on Thursday, 7 February 2013 16:04:57 UTC