so the "rt implementation of retinex could not give us the total removal of haze, like fattal algo or true algo.
ok i’ve to install the local lab version a step ahead 
Jdc in French in spanish in japanese google translate help always.
Apart from jokes thanks, even for the i would look into doing that, i understand that would be an effort do that, and that u could not guarantee anything.
then, with james and shreedar we are are almost there, shadows a bit uplift but not “bitmap effect”,i’ve got an idea of workflow thanks to all you tomorrow i would try out and report…
Afre if u give a look on rawpedia u will discover that retinex have a lot more to offer.afre i u could share you result,
only one thing, the scope the help i asked is in this case is understand how to remove haze with the aid of retinex mantaining the photo usable, taste is a bit secondary, cause once remove dthe haze we could think about of how make the photo pretty, if we manage to get the rid of the mist we would have a common starting point,
the trasmission is the key, how to manipulate that mask, correctly… for the result…
where would the light be? for that photo to be honest is right were is i choosed that photo cause could be corrected, not cause really need to be cured, and could be prettier in various way, as shreedar have demonstrated using only lab curves.
I thinked this photo would be decent cause is not too badly exposed, the fog is more on the second plane, so a “standard” condition for a “dehaze”.
Retinex by definiton of the thing could work for a lot of different scenarios not only with mist or fog, with lights… recovery of part of a photo from deep shadow. think about concert or other much more harsh condition where this tool could be really useful once truly grasped.
ps thanks for the like of the photo, is one of a row, for a pano,. when finished i would post it here