Re: math: URI scheme and protocol handler

On 2014-04-29 10:42, Paul Prescod wrote:
> Thank you, that video clarified.
>
> The way the browsers are implemented, it is more seamless to launch an
> external application with some data using a URL protocol rather than a
> content-type.
>
> http://www.chromeplugins.org/google/chrome-plugins/how-start-external-applications-9717.html
>
> I have done this myself in the past.
>
> I don't know what to say: the implementation defects of the browsers are
> at odds with how web architecture is supposed to work. URL handlers are
> not supposed to be application-launchers. They are supposed to describe
> the actual protocol for downloading or manipulating data. Think of the
> mess that will arise if every file format also needs a standardized URI
> format as a way of working around browser behaviours.
>
> If you do not worry about a formal standardization process at the W3C
> then nobody will care that you are using URIs in this way.
>
> The systemic fix is that HTML should have a way of stating that the
> target file is designed to be "transient" and the user should not be
> harassed about a filename for storage. If the user wants to save the
> data then they could do a "Save As" from the viewer app.
> ....

Not sure what you're trying to fix here. If the browser knows the 
handler for the media type, it's supposed to download to a temporary 
file and pass that to the media type handler.

Best regards, Julian