(In reply to comment #0)

Something similar is needed for
http://en.wikipedia.org/wiki/MediaWiki:Exif-software-value (associated with
http://en.wikipedia.org/wiki/MediaWiki:Exif-software ) but here the problem
is
that if $1 merely states a bare version number like 6.0, that might be
ambiguous - in the context of an Apple iPhone, it should link to
http://en.wikipedia.org/wiki/IOS_6 but if the camera is not an iPhone, that's
clearly the wrong link. Either the value passed into $1 needs the camera make
adding, or a second parameter needs to be added which has the camera make.

See http://en.wikipedia.org/wiki/File_talk:Torre_dei_Becci,_June_2013.jpg for
background

Not that it really helps, but software is not supposed to really have anything to do with the camera make and model. Software is supposed to be what software processed the photo after the camera took the picture. /me blames camera manufactures...

I agree with you Bawolff that the camera manufacturers are to blame here, I think that adding a $2 parameter to [[MediaWiki:Exif-software-value]] that contains the same value as $1 does in [[MediaWiki:Exif-make-value]] then it would be easy to use parser functions or whatnot to say if $2 is Apple and $1 is 6.0 then link to iOS_6.

I had a patch that added a parser function which could access other metadata fields (gerrit change 67047. Its been in limbo for a while now). That is one possible solution to this problem.

(In reply to comment #2)

I agree with you Bawolff that the camera manufacturers are to blame here, I
think that adding a $2 parameter to [[MediaWiki:Exif-software-value]] that
contains the same value as $1 does in [[MediaWiki:Exif-make-value]] then it
would be easy to use parser functions or whatnot to say if $2 is Apple and $1
is 6.0 then link to iOS_6.

Hmm. It seems just really arbitrary to me, given the fields are supposed to be independant.

What about special casing it so if the make is apple, and software is 6.0, then we display exif-software-value-6.0 rather then the normal msg. Somehow this feels like it would be less arbitrary to me, although it would still be pretty arbitrary.

Is 6.0 the only value this happens for?

According to the discussion redrose linked, it's not the only one. There is also at least a 7.2.x with a similar issue. Being able to get a wider set of important values and using parser functions is most logical to me. How those values are obtained is unimportant whether it's done with a $2 variable or the new parser function you mentioned above.