@Joe Riel
I overlooked the fact that my suggestion only makes sense for conventional compression springs. It looks like we will have to live with the current implementation unless a better contact condition is proposed that can also handle an unstretched spring length s_rel0 of zero length or even of negative spring length. Some experts might find such length definition useful.
I am afraid that my main concern, of pointing out potential use errors of sign, is not well reflected in this post. Here is another attempt that hopefully better illustrates the consequences of an invalid configuration (red mass).

As indicated by the green mass, there should be no movement, but the red mass starts moving towards the contact point (at the x=0.1) because of an error of sign that is not obvious to avoid. It seems that the only way to deal with it is measuring s_rel and verifying the sign.
On a different case I was lucky enough to see that “something” was wrong with the results, but it took me considerable time to identify the cause. Unfortunately the text passage of the help page of the Elasto Gap component, that I consulted several times to make sure that I used the component correctly, is rather supporting the natural interpretation of a distance in terms of magnitude. I think a reminder of the importance of the sign would not harm (Or examples. The model gallery is full of it, but there is no way to filter for components used in a model.)
Thank you for your comment. I hope the discussion at least helps to draw attention to the fact that there can be errors of sign that are not obviously avoidable by either good practice or by component design.
Prismatic_Joint_wo_spring_with_Elasto_Gap.msim