Since the main benefit of CS is to add short spacings where there were none before, and since both Braun's original (1993) suggestion and Holdaway's (1994) tests were based on moving a single antenna in to station N1, it seems likely that moving a single antenna would be sufficient to gain most of the benefits of the CS configuration. Compared to the 1997 (two antennas) CS, this significantly improves the uv-coverage at intermediate baselines, and makes the new configuration less vulnerable to the loss of random antennas for maintenance and the like. Of course, it leaves CS itself more vulnerable to such losses - if the central antenna goes out, one is back in standard C configuration, and all the benefits of CS are lost.
This loss of robustness is quite important, and is the main argument against the single-inner-antenna versions of CS configuration. The goal of CS was to allow single-configuration imaging, i.e. to avoid additional observations in D configuration; a configuration which effectively vanishes when a single antenna goes down is not a very good way to achieve this goal. Consider an HI observer who observes a galaxy with CS configuration, only to find after the observations that she lost the CS antenna. It would be too late by then to apply for time in the coming D configuration; instead she must wait for the next proposal cycle, some 15 months hence. The CS configuration because it avoids the hassle and delay of proposing for, observing, and reducing the data from two configurations. If CS itself is not guaranteed to provide the advertised short spacings, observers are likely to revert to applying for both CS and D configurations, just in case.
One possible approach would be for NRAO to guarantee D configuration time to any project specifically requesting CS configuration, if the CS antenna were inoperative during those observations. This would be annoying to those who come to the site for their observations, and require some effort on NRAO's part both in keeping track of the CS antenna's down-time and in requiring a somewhat floating time allocation in D configuration. In the worst case, a few projects might also be hurt by solar interference if the objects studied were not up at night in D configuration.
Observers wishing the standard C configuration are concerned not with the newly-occupied inner station(s), but with the newly-introduced gaps in the intermediate uv-coverage. Here again the worrisome case is when several more antennas disappear for maintenance, VLBI observing, or other reasons. One approach, suggested by Frazer Owen, would be to adjust the current `3-antenna' rule during CS, to allow only one (if two antennas are moved to inner stations) or two (if only one is so moved) antennas to go down before calling someone out to deal with the problem. Such a `single antenna' rule was in effect during the VLA surveys. The practical difficulties are, first, that on weekdays this might complicate standard maintenance activities; and second, that during off-hours, it might be hard to track down the appropriate people to deal with a given problem. An alternative might be to specify certain stations as `critical' in CS configuration, perhaps counting them as two antennas towards the 3-antenna limit - these would include the newly-occupied inner station(s), and those adjacent to the newly-vacated intermediate one(s). Such a critical antenna rule would be especially important for the single-inner-antenna version of CS; fortunately that central antenna would be one of the easiest to reach!
Obviously it would also be wise to avoid choosing single-dish VLBI antennas either from the CS antennas or from those adjacent to the pad(s) the CS antennas used to occupy. On this same theme of `special' antennas, we should probably not move a 7mm antenna in to the central position - in the current (spiral) 7mm configuration, this would argue against moving the antennas from N10 or E12.