Is there a method to identify and measure truncated transcripts by RNA seq? How will the system distinguish between the full length and short transcripts?
Yes, In RNA seq mRNAs are fragmented in short fragments before library formation and sequencing. So it is not a problem. The only problem is mRNA enrichment and of good quality.
If this transcript is abundant in comparison to the normal form, such as that resulting from a mutation, then there should be no problem with detecting that even with standard RNA-seq protocol. You might want to do transcript assembly option instead of mapping to a reference when analyzing your RNA-seq data.
However, this method will not work if your transcript is not abundant (say, less than 50% of transcription events result in this transcript), because you will not be able to pick it up against normal background. In this case you might want to use a protocol that specifically captures 3'-ends of intact RNAs. A version of this may be found in this paper. http://www.nature.com/nmeth/journal/v7/n7/abs/nmeth.1464.html
or in this one
http://www.ncbi.nlm.nih.gov/pubmed/21248844
This of course is more involved than the regular RNA-seq.
Is the truncated transcript polyadenylated? If not, it is likely to be unstable and will be degraded so that levels are much lower than a full length version from the same locus. If it is polyadenylated then it is likely to be detected and the RNA seq reads will map to the portion that is transcribed but not to the deleted/truncated part of the gene.