i have to strongly disagree with your thoughts on photographs as a research tool. if read correctly and with the biases inherent in the act of taking a photograph factored in to the reading, photographs can form an invaluable primary source.
i will say that my main experience with using them as source materiel is for another hobby, prototype based model railways, not military history. with regards to modelling in the form that interest me, being creating an accurate 4 dimensional representation of what was happening at a particular time and place, photographs and oral/written histories are the only way to know how things actually were as supposed to how things were "supposed" to be. all of the plans and paper documentation still leave massive gaps in the information needed to accurately portray what really happened.
as we all know, in the real world, the way we are "told" to do something isn't always the same as the way it gets done.
the purpose of the taking of the photographs really sets the to fields apart, i'm not sure about military laws relating to private photography in the US services in WW2 but i believe it was illegal for australian servicemen to take photographs on duty. of course it still happened (the family still has some photographs taken in singapore by my grandfather) so the majority of ww2 photographs we see were taken by the services themselves or news agencies.
pre digital when we as normal people, took a photograph, it is generally for our own usage, to capture capture a moment, the people we are with or the scene we are seeing. for a service or press photographer it was meant for the consumption of others an as such that end use (with its intended message) was considered during the act of taking the photograph.
as we know official ww2 photographs were often staged to get the desired end result and most railway photographs were taken by interested people to capture a scene or moment in front of them, making the railway photographs far easier to read.
knowing who took a photograph originally makes the choice of how to read it a lot easier, but you can make a educated guess of who took a photograph just by looking at it's mechanical quality, subject matter and it's art/design values.
another thing to remember is that photographs can not only be read individually but also as part of a set (or sets) that together can give a far greater overview and extra information than each photo separately. by reading photographs (or any document for that matter) we also add our bias to the interpenetration but as long as you know you are doing this it can be partially mitigated.