One of the most under-appreciated skills that programmers (and members of other professions as well) have; and a key reason why programmers aren't commodities. Domain knowledge is expertise in a particular problem area - especially problems which aren't the primary concern of a computer science curriculum.
Examples of important areas of domain knowledge that can help land you a job:
Graphics and multimedia
In many cases, DomainKnowledge can be very company- or site-specific, i.e.:
Knowledge of the accounting systems and procedures in place at GeneralElectric
Knowledge of how to write plugins for Adobe PhotoShop
Expertise in designing codecs supporting the MPEG standard.
Generally, skills in a particular programming language (or programming skills in general) aren't considered a problem domain. Some computer topics are legitimate problem domains of their own, which one might acquire DomainExpertise? in:
Programming language design/compiler design. (A programmer who is skilled at writing code in Java wouldn't be considered a "Java domain expert". One who is skilled at implementing JavaVirtualMachines, on the other hand, might be.)
In general, domain expertise is hard to come by (which is why it is so valuable). True domain expertise requires both a considerable amount of study in the topic (formal or informal), and a considerable amount of experience in design and implementation. One doesn't read the RedDragon? and become an expert in compilers; but if one reads the RedDragon? (and a few others) and works on GCC or a commercial compiler for a few years, one could start to claim domain expertise.
DomainKnowledge can be modelled with SemanticNets.