We are often asked why the wide-field imaging and deconvolution steps are more difficult than their equivalent for single-field. The main answer is that doing wide-field observations with an interferometer is kind of paradoxical. Indeed, (sub)millimeter interferometers are before all tuned to get the best possible spatial resolution. A natural consequence is the lack of measurement of the low spatial frequencies which are extremely important in wide-field observations. Hence the paradox. In addition, the first goal of the current generation of (sub)millimeter interferometer was to demonstrate the feasibility of routine interferometric observations in the (sub)millimeter range. The need of doing wide-field imaging at high spatial resolution has thus been recognized only after the success of high spatial resolution single-field imaging. This makes the implementation of wide-field imaging more difficult. The situation will improve with ALMA because wide-field imaging has been included from scratch in its design. It remains that wide-field imaging pushes (sub)millimeter interferometers to their limits.
From a software point-of-view, we always try to hide the technical complexity in more user-friendly tools as soon as our experience increases. However, because of its paradoxical nature, wide-field imaging with an interferometer implies a knowledgeable use of those tools.