Hi, all…

In the bake-off on Monday, when we compared file types, Black Duck’s 
application returned a file type of APPLICATION for quite many file types, 
while others did not. While at that point we identified this behavior to be a 
bug, this behavior is actually a consequence of treating the MIME type 
(application/*) as APPLICATION.

From the spec (Sesction 4.3, g)
APPLICATION if the file is associated with a specific application type (MIME 
type of
application/*);

This MIME type specification allows for anything of type 
application/octet-stream to be labeled an application.  For example, running…

file --mime-type -b ./cpio-2.10/po/id.gmo

yields a type of application/octet-stream, even though everyone but us 
specified the GMO files as other.

The language “if the file is associated with a specific application type” is 
more restrictive than the MIME type. However, I would then question what it 
means to be “associated”. An Adobe Flash movie has the MIME type 
“application/vnd.adobe.flash-movie” and may be “associated" with Adobe flash 
players. Does that mean a flash movie is an application? How about Microsoft 
Visio files (“application/vnd.visio”, 
http://www.iana.org/assignments/media-types/application/vnd.visio)?

I think we need a definition of an application that’s independent of MIME 
types, or, for that matter, file types. I would propose the following, just to 
get the ball rolling:

4.3… g) APPLICATION if the file can be executed (natively or with an 
interpreter) to provide functionality.

I mention interpreters, because the execution of an application may well 
consist of running  a bash script or for that matter, a javascript file, and in 
such cases, we should establish that they should be labeled as such.

What do you think?
_______________________________________________
Spdx-tech mailing list
[email protected]
https://lists.spdx.org/mailman/listinfo/spdx-tech

Reply via email to