slickdeals,
You will probably know my opinion seeing I write one of the programs mentioned.
If you look at the reason in the code, it is because of development of 1st crack and requiring sufficient reinforcement to ensure that the tension force required to crack the concrete can be carried by the reinforcement post cracking. So yes, it is logical.
If you look beyond the code and look at reinforcement strain, which none of the major codes limit for some illogical reason, this is a defacto limit on the maximum strain, albeit a not very efficient one for anything other than rectangular sections.
For reinforcement with a maximum strain of 5%, you will find that with minimum reinforcement the strain will still be slightly over the breaking strain at ultimate strength. If you have less ductile reinforcement (eg Australian Class L or Eurocode class A, USA WWF and probably some other reinforcement types) then the actual strain could be 3 - 4 times the breaking strain. So minimum reinforcement really should be much higher for the lower ductility reinforcements.
Where the logic breaks down for T sections is the minimum reinforcement rule gives higher requirements as Z of the section increases (flange in tension for a T beam) where a proper strain calculation has higher minimum requirement where the compression face is widest (flange in compression for a T beam). If you used this logic for the section across the beam in your diagram, minimum would be much higher than the codes currently say in the beam direction at +ve moment sections.