It seems, sometimes, to MARC's critics, that the standard is overly efficient about things with which it needs to be flexible, and too explicit about things that would be better expressed efficiently.
For example, every monographic MARC record states whether it is a Festschrift, but only a very small percentage of monographic records actually are. Routinely having to code data values that should be optional gives a disproportionate prominence to the fields with fixed byte values.
Another example of the tension between using coded, control fields and the, optionally occurring, variable length fields is the discrepancy in the way each are handled.
For instance, byte 28 of the 008 field indicates that gazetteers should be coded as dictionaries; the LC subject headings, however, treat them differently.
On the whole, I would question whether using cataloging codes is really necessary. With disk space and bandwidth not the problem they were when MARC was created, should we still live with the limitation of not being able to search for the values represented by these cryptic codes?