Short-term vs long-term memory
Working memory is a relatively recent term, a refinement of an older concept - that of short-term memory. Short-term memory was called thus to distinguish it from "long-term memory" - your memory store.
One important difference between the idea of short-term memory and working memory, is that short-term memory was conceived of as a thing. Different from long-term memory (variously analogized as a library, a filing system, a computer) chiefly in the duration of the records it held. But working memory, as its name suggests, is now conceived more as a process than a thing. A state of mind. A pattern of activation.
Working memory contains the information of which you are immediately aware.
To put information into our memory store, it must ... be worked on - i.e., be held in working memory. To get information out of the memory store - to “remember” something - it must again be in an active state - be in working memory. How can we know what we remember if we're not conscious of it?
However, you can only keep something "active" for a very short time without your conscious attention. It is this which so limits working memory capacity.
The magic number seven
Probably the most widely known fact about working memory is that it can only hold around seven chunks of information (between 5 and 9). However, this tells us little about the limits of working memory because the size of a chunk is indeterminate.
1 2 3 4 5 6 7 are seven different chunks - if you remember each digit separately (as you would, for example, if you were not familiar with the digits - as a young child isn't). But for those of us who are only too well-versed in our numbers, 1 through to 7 could be a single chunk.
Recent research suggests however, that it is not so much the number of chunks that is important. What may be important may be how long it takes you to say the words (information is usually held in working memory in the form of an acoustic - sound-based - code). It appears that you can only hold in working memory what you can say in 1.5 — 2 seconds. Slow speakers are therefore penalized.
What we term "working memory" contains several functions, including the "central executive" which coordinates and manages the various tasks needed. The extent to which working memory is domain-specific (different "working memories", if you like, for different sensory and cognitive systems, such as language, spatial memory, number) is still very much debated. However, at a practical level, we may think of working memory as containing several different components, for which you have different "capacities". Thus, your capacity for numbers may well be quite different from your capacity for words, and both from your capacity for visual images.