[Ksummit-2012-discuss] [ATTEND] kernel core dump and "dying breath"

Will Deacon will.deacon at arm.com
Mon Jun 25 17:22:44 UTC 2012


On Mon, Jun 25, 2012 at 06:16:03PM +0100, Matthew Garrett wrote:
> On Mon, Jun 25, 2012 at 10:12:48AM -0700, H. Peter Anvin wrote:
> > I think it makes much more sense to have the decoder server-side.
> > 
> > There are two reasons for it:
> > 
> > 1. *one* code base!
> > 
> > I can tell you just how bad a lot of the QR decoder software running on
> > smartphones are -- because I have tried them.
> > 
> > 2. if the decode doesn't work, you have the picture server side, and you
> > can perhaps improve the software and collect it as "a hard to decode
> > picture" and use it as a regression test.
> 
> And the main reason not to is that a picture gives us a single data 
> point, while most QR apps will take continuous readings until they're 
> able to decode. The user also knows that the decode was successful 
> *without* having to work out how to get a multi-MB picture off their 
> phone and into the interwebs.

So why not try both? Assuming that this is hidden inside an `app' of some
description, we could try to decode the image on the phone first. If that
fails, we send the image instead and possibly some information about the
phone so we can name-and-shame the offending hardware. I guess it depends on
how reluctant we are to develop/support a client-side application.

Will


More information about the Ksummit-2012-discuss mailing list