the chemist in me says to change just one variable at a time
Well, yes and no... And here is why 'no' is sometimes the best answer:
First, it is impossible to change just one thing. There are always at least two things interacting -the sound of one hand clapping is dead silence: alone in a solitary vacuum the hand ceases to exist. Changing any one thing changes it's interaction with all the other things in the universe.
Really thorny problems are the result of the interaction of many independent causes and removing just one of the causes may not have any noticeable effect. As the causes are independent they tend to combine as the root-mean-square, and the removal of one item causes only a very small change in the final result - often the change is so small it can't be told from normal experimental variation.
As an example of a root-mean-square addition take a system with 5 equally contributing causes: fixing any one doesn't have the expected 20% effect on the outcome but only a 10% effect.
An experimental approach, starting from a non-working system, will not find the causes.
For an experimental approach to find out what is happening you have to start with a
working system - as you can then see the effect of adding each possible cause. If you have the same system with 5 equally contributing causes, then a working system that has only one of the causes added to it will show 45% of the effect of the combination of the 5.
In the uneven development problem under consideration there seem to be many candidate causes:
- Agitation is too gentle
- Agitation isn't random enough
- Agitation isn't frequent enough
- The film is sensitive to agitation
- The developer is sensitive to agitation
- The development time is too short
If these all have an equal contribution, changing just one of them will only have an 8% reduction in the unevenness of development - the result of the change would be barely noticeable.
In this example you would start with a protocol that gives even negatives and then make the agitation gentler and gentler until the system fails. Then go back to adequate vigor in the agitation start to reduce the frequency until the system fails. And so on.
A good starting system would be one where the film, the developer and the development protocol are all from the same source (say Tri-X, D-76 and slavishly following Kodak's directions) - hopefully this will yield even negatives. Then change the film to Acros, see if the negatives are still even. Then go back to Tri-X and try Ilfosol. Then try the combination of Acros and Ilfosol - assuming there is some magic expected from this combination and getting it to work is the desired goal. If the results are still OK, then experiment with agitation. Somewhere between a bog-standard Tri-X/D-76 protocol and the present method things will break down.
The chances are the combination you are using is possibly the worst combination that can be had: either more or less agitation (stand developing) would improve things.
* * *
Problem solving is something they don't teach in schools. Professor's don't get enough real-world problem solving experience to be able to teach it. Which is OK by me, my firm has made a great deal of money from clients who have spent years trying to solve their problems with a self-congratulatory "we only change one thing at a time" approach.