Users corrections to the released version of the code

[arango]

It has come to my attention that some users make some modifications to the internal structure of ROMS code like changing modules, globaldefs.h, and the model kernel when trying to solve problems with their particular application. In most cases, such changes are not needed and are usually related to the lack of familiarity with the model.

The model is quite stable nowadays and it used in a wide variaty of applications running in different types of computers. So I highly recommend to examine what are you doing first before start making such modifications. Perhaps, it will help to ask the question: Why is this problem has not been reported by any of the ROMS users worldwide?

This does not mean that everything in the code is written on stone and cannot be changed. If some change is required, please let us know about the problem and we study the situation. Please be very specific when posting such problems and provide enough information. Simple statements as the the model doesn't compile or the model blows-up will ignored by lack of information.

However, this does not mean that we have the time to debug everybody's application. It is our experience that new users suffer when starting using ROMS. But this is kind of healthy because the user learns more about the code when trying to solve their problems. So please be patient and do not jump to hastly conclusions

[kate]

I've got one for you. I have a test problem called MEDDY, using the rho=T equation of state. It used to run without modification to ROMS, but then ROMS changed and the MEDDY problem blew up. It turns out ROMS no longer allows a pure rho=T equation of state, insisting that rho=-alpha*T and that alpha is always positive. Suddenly rho = -T and things are statically unstable!