The ISPF product running under TSO is filled with panels containing the three fields PROJECT, GROUP and TYPE. These three fields constitutes the name of a dataset with three qualifiers. The contens of these three fields are remembered by ISPF across different panels and across different sessions and that is good. The irritating part is that you can't reference datasets with more than three qualifiers.

But is this completely true? ISPF just concatenates the contens of the three fields and put a dot between the contens of each field. This gives you the ability to cheat a little, but only as long as the qualifiers in your dataset names are quite short. Think of a situation where you want to use a dataset called MY.FINE.NEW.DATASET. You can reference this dataset with PROJECT equal to MY.FINE, GROUP equal to NEW and TYPE equal to DATASET or you can try with PROJECT equal to MY, GROUP equal to FINE.NEW and TYPE equal to DATASET. Both of these methods will work equally well. The general rule is that as long as two or more qualifiers including dots are shorter than or equal to 8 characters they can be put together in one or more of the three fields.

When it comes to datasets with just two qualifiers or datasets with more than 26 characters the decribed method dosn't work anymore. Then you have to use the "normal" ways of referencing those types of datasets.