There is a difference in the way STELLA & VENSIM on the one hand, and SPHINX treat a CONNECTOR from a FLOW to a CONVERTER. The difference becomes a problem when we try to implement a STELLA model on SPHiNX and you have DT smaller than 1. I've suggested a workaround.

Throughout SPHINX, quantities asociated with flows are in units per period. In the case here, the flow is presented in quantity per 10th (or 100th or 1000th) of a period.

The SPHINX CONNECTOR reports the value of the flow at the prior step, which is a quantity per 10th of a period when DT=0.1, whereas STELLA & VENSIM apparently report the value for the prior period, or perhaps the prior step value /DT. When DT is 1, the prior step and the prior period are the same. To work around this, at the CONVERTER you need to divide value by DT. This is not exactly the prior period value, but whether it should be is a philosophical matter. Perhaps it should be the sum of all the step values that constitute the prior period? Anyway ...

The following is from the fish banks model in file D-4543-2.pdf located at http://www.atmosedu.com/WSU/esrp310-550/FishBanks/D-4543/D-4543-2.pdf

In their STELLA model, there is a CONVERTER named “Revenues” which obtains the value of “Total catch per year” via a CONNECTOR from a FLOW named “Total catch per year”.

The STELLA equation in “revenues” is

“revenues” = “Total catch per year”*”Fish price”.

The above works in SPHINX only when DT is 1.

In SPHINX we need

“revenues” = “Total catch per year”*”Fish price” /DT.

While the program is reporting values by period, it is calculating each period in 10 steps (when DT is 1/10 ).

Alternatively one could rewrite the model so that profits are handled via FLOWs - I'll submit a follow up to explain.

I have also noted a condition where sphynx puts java in a loop – will report in a second message.

It would be desirable in my view if SPHINX were changed so that flows are always quanty per period. I've downloaded the source, but it will take me some time to locate the applicable code.

BUT

Ignoring these 2 minor issues, this is one heck of a good program.

Thanks so much - I love it

DIck