Thanks David for the elaborate explanation of the concept and the precise requirements. As i mentioned, i am currently using a commercially labelled fluorescent probes for the extension. The length of extension i am interested is indeed short (between 10-30 bases), that is why the 12%. Summing up my problem and the "impossibility" of pulling this off highlighted by you, i think i might have to look for a way around to size my RNA.

I agree almost fully with David's excellent answer and would add that the dye labels are differently charged so modify the speed at which dna fragment run which is sorted out by some fancy computing in the sequencer conventionally. addressing your problem ,however if the samples were loaded at time intervals some of the sequence will be near the end of the gel so will be separated as was done routinely with radioactive sequencing. I wonder if you had a biotin labelled primer and loaded 4 tracks each with the different terminator then  blotted and added a common avidin labelled chromophore or even an antigen that could be visualised with labelled antibody techniques like ELISA if there would be enough signal to visualise the individual bands. Very labour intensive and time consuming. You could discuss the problem with a commercial sequencing company...if you had a sequencing primer with a long 5' tail to make the sequence fit into the analysable part of the sequencing run and adjusted the sequencing kit for sequencing short sequences there might be some scope. Does pyrosequencing ( probably again commercial ) work...it is very good for short sequences close to the primer