At some point soon the entire idea of "genes" will have to be replaced with a more general concept. One sign that it is close to the end is that we can't precisely count genes in a genome. Small variations in the definition of a gene result in large differences between number of genes in the genome.
The number of genes in the human genome have varied between 10^4 to 10^6. Right now, the number seems to have settled down to 3×10^4. However, the definition chosen for this number is somewhat arbitrary.
One thing that complicates the definition of gene is that some contiguous sequences of RNA are transcripted from discontiguous segments of DNA in the chromosome. Another thing that complicates the definition of gene is that in eukaryotes, proteins are often synthesized in a two or three step process after the genes are translated into proteins. So enzymes do not always have one on one mapping to the corresponding sequence of DNA.
Developmental biology still has a revolution or two in its future. Although a "review" of the definition of gene would be helpful, there is no way to fix the concept of gene. I suspect that the theory of gene networks may be due for some breakthroughs. Furthermore, we have to find out more about epigenetic inheritance. By epigenetics, I mean inheritance through molecules other than DNA.
Don't get me wrong. The "gene centric" models of developmental biology will still continue to help evolutionary biology for a considerable amount of time in the future. Even while a new "synthesis" of biology is being developed, the gene centric approximation will be extremely useful for simplifying and clarifying evolution. However, there are some fundamental inconsistencies in the theory of genetic inheritance as we know it today.
I predict that in the future, "the gene" will have the same level of acceptance in biology as "the fluid element" has in classical mechanics. "The gene" will be recognized as a mathematical concept based on an approximation rather than a strictly physical concept based on the general theory.