All those things could be used as metrics, but I don't think you could ever say one number is absolute and definitive, unless it's what you could compress a person's bitstream down to for a Star Trek transporter beam. Nucleotides of DNA in the genome, each corresponding to two binary bits, is a good rough measure, but there are a lot of caveats. For example, DNA never builds a cell or an organism by itself, so there is some information in the machinery of the cell, which isn't as easily quantifiable. OTOH, you could probably throw out much if not most of the DNA of a human being and it wouldn't make any significant difference to the development of the organism (provided you knew which bits to throw out!)
Here's an example of an increased number of coding sequences (the gene + later the disabling RNAi snippet) that cause a decrease in the number of proteins created.
Well, the use of an endogenous "siRNA" transcript to down-regulate another gene is conceptually no different from the use of a protein transcription factor to down-regulate another gene. It only really buys you something if the expression of the silencing agent -- the siRNA or transcription factor -- is conditional and regulated. If it were just cancelling out the effect of the other gene all the time, the whole thing would just represent a waste of bits, and would be lost over time, or never develop in the first place. And generally there also wouldn't be much point if the regulator has only one target, because then you just get a regress, since the target gene could just be regulated directly without the complication of the intervening regulator. But regulators are few and controlled genes are many: gene networks feature hierarchies of control, just like armies, nation-states and business corporations. One regulator controls many downstream genes, just as one foreman gives orders to several workers. These considerations hold regardless of whether we are talking about traditional protein-mediated control, or these novel siRNA-mediated mechanisms. Protein versus RNA is just a question of instrumentality, like JAVA versus C++.
In the process of creating proteins, the DNA of course requires the use of raw materials from outside the nucleus located in the cell itself. These raw materials of course got there because when the cell was formed as a certain kind of cell its purpose was to get just those raw materials. That is why different cells have different functions and produce different proteins even though theoretically they could produce any protein at all since they all have the same DNA code.
What all this leads to however is a sort of chicken and egg problem - which came first the DNA or the cell? This is a really big problem for those who believe abiogenesis is possible.
One regulator controls many downstream genes, just as one foreman gives orders to several workers. These considerations hold regardless of whether we are talking about traditional protein-mediated control, or these novel siRNA-mediated mechanisms. Protein versus RNA is just a question of instrumentality, like JAVA versus C++.
Interesting that you mention programming languages in the discussion of gene regulation. You sound like an anti-evolutionist insisting that the genome is a program and was thus intelligently designed. Surely you do not wish to assert that programs are written and modified at random do you?