Re: inline declarative manifest, was Re: New manifest spec - ready for FPWD?

More comments inline, but I’ve started running a developer survey here about the proposed solutions:
https://gist.github.com/marcoscaceres/7783977

See also:  
https://twitter.com/marcosc/status/408150324629630976

Some really great feedback from the dev community on twitter! Please take a look.  


On Wednesday, December 4, 2013 at 8:10 PM, Charles McCathie Nevile wrote:

> > Having said that, there are issues also with navigating installed web  
> > apps. The phonegap guys have a wealth of experience to share here,  
> > though they are proponents of single page apps to overcome limitations  
> > in the Web platform (e.g., avoiding the flash of white when loading  
> > another web page when navigating). Anyway, we can deal with those issues  
> > later - but just want to be clear about what we’ve seen in the dataset  
> > we’ve been looking at in WebMob and what the forthcoming issues are  
> > going to be if this goes mainstream.
>  
>  
>  
> Looking forward to you publishing that data. Is there a simple pointer for  
> those of us who haven't been following webmob closely so we can start from  
> somewhere better than "all of webmob"?

The report is here:  
http://w3c-webmob.github.io/installable-webapps/

The data is from webdevdata.org (a W3C community group a few of us set up to gather this kind of data - you should join!:))

> > I really don’t like this - specially messy with the single quote/double  
> > quote thing which is one screws it up is a huge pain in the as to debug.  
> > Structured content really feels like it should be in an element.
>  
>  
> Yes, but I wonder how big a real problem that is. I happen to use a  
> bare-bones version of vi that doesn't balance parentheses, quotes, etc,  
> but I believe that shows I am a very strange person. As far as I know,  
> debugging this kind of error semi-automatically is actually pretty  
> trivial, and common.


It might not be a huge problem because I’m trying my hardest to keep the structure flat. It becomes a nightmare pretty quickly once things start nesting. The lack of commenting and not being forgiving about trailing commas doesn’t help, etc. It also feels crapy that one has to special case this one attribute to use single quotes.  
  
> >  
> > I’ll bounce it to HTMLs people and see what they say.
>  
> From an authoring perspective I don't think the verbosity is a big issue,  
> and the two options *seem* about equivalent in cognitive load - switching  
> elements is odd but people are clearly used to doing it for style, and  
> setting a type attribute in a script element compared to using link/meta  
> is the same thing.


True. That’s a good point.  
  
> Which makes the big question about taste in syntax (unless there is some  
> real implementation or web-compat issue). So I expect it to be the hardest  
> one of all to solve :)
>  

I’m being told in IRC that there might be CSP implications with either of the solutions that I need to investigate.

Received on Wednesday, 4 December 2013 13:09:28 UTC