W3C home > Mailing lists > Public > public-web-mobile@w3.org > February 2014

Re: Feedback on installable webapps

From: Marcos Caceres <w3c@marcosc.com>
Date: Mon, 3 Feb 2014 12:37:53 +0000
To: David Bruant <bruant.d@gmail.com>
Cc: public-web-mobile@w3.org
Message-ID: <3D68447B59FC4DD5803C82B9DA8590C6@marcosc.com>
Hi David, 

On Sunday, February 2, 2014 at 1:23 AM, David Bruant wrote:

> Hi,
> Sending feedback regarding http://w3c-webmob.github.io/installable-webapps/
> First off, thanks a lot for this document, I learned a lot!

I'm glad you found it useful.  
> I'd like to comment on the last part about requirements for a 
> standardized solution:
> "MUST provide a mechanism that allows a developer to indicate that the 
> application is able to function as standalone. If no such indication is 
> given by the developer, then the application MUST launch in the Web 
> browser as normal. That is, the solution MUST NOT conflate bookmarking 
> with allowing a web application to be launched as standalone. "
> => I wonder why that is. It's been the admitted practice, but it's not 
> clear why it MUST be that way. A browser could provide an "install" 
> button akin to the "bookmark button" and decide on safe default 
> semantics (different than "launch web browser") when the developer 
> didn't opt-in.

Yeah, when initially wrote the document I had not thought that through enough: it's actually completely a user decision. In the manifest spec itself, you will notice that  "mode" is now gone as an option.  
So, I need to change this requirement. Filed bug:
> A while ago, Mozilla did some research on how people use apps and mobile 
> browser [1][2] (took me ~30 minutes to find that again #SEOFail). One 
> conclusion is that people use apps when they know what they want to do 
> and focus on it and the mobile browser when they're searching for things 
> (this is actually a great insight on why people confuse web browser and 
> search engine).
> It looks like if people click the "install" button, what they want is 
> "this web page without the browser UI" (which is associated with 
> searching/navigation, not focusing on one task). It's possible to offer 
> them that even if the developer didn't provide opt-in.

Yes, exactly. Thanks for tracking down that research (though a sample size of 12 is a wee bit small to conclude much from - but it's certainly indicative, I think, of a lot of people's experience). I've added the research to our wiki so we don't lose them: 

> "MUST provide a mechanism for the application to check (e.g., via 
> script) if it's running as a standalone application. "
> => I imagine a media query could make sense too.

Yes - that was the idea that I've had for a few years (see view modes [3])... but I've not managed to convince browser vendors of that (apart from old Opera, which supported view modes). 
> "MUST clarify the security model with regards to how permissions and 
> data are shared with the web browser from which the web application was 
> bookmarked or installed."
> => Yes. Among other things, if I could "install" a website and have 
> separation of credentials between my standalone site and the browser, 
> I'd probably have Facebook in its own "standalone app" and disconnect 
> from Facebook from my browser (so they're less aware of my browser 
> history via Like buttons ;-) )

That's the idea... but there are a lot of issues, like having to log into Facebook again etc.  
> David
> [1] https://blog.mozilla.org/ux/2012/03/browsers-and-apps/
> [2] https://air.mozilla.org/what-has-the-ur-team-been-up-to/

[3] http://www.w3.org/TR/view-mode/ 
Received on Monday, 3 February 2014 12:38:27 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:59:03 UTC