I bought Dr. Strahl's Internet Applications book (great!) and have also read through some of the MTS white paper he wrote as well as several MS VFP MTS IIS texts. I think the reason I'm getting confused is because some things seem to contradict each other.

1. What I thought I could do is use SQL-Server as a data store and have web browser users or VFP client users (using CreateObject in each case) get their data through an out-of-process VFP .exe server. This middle-tier application would be able to authenticate users (based on local group membership in a local NT SAM) and fetch remote views of data and pass them to the calling application. But I thought this might lead to problems because I am not using MTS to actually broker anything (just-in-time activation, ODBC connection pooling) in this case. I also don't know what "universal data format" I would use to pass data back to the caller. I also thought that the mid-tier software would be essentially running in the client's memory space (CreateObject()) and not be truly detached.

2. Then I thought I would just do an in-proc VFP .dll instead because then I could make it run inside MTS, but I am still instantiating an object inside of the client-tier's memory space. I think I also have to put some client-side logic there to talk to MTS in order to support the rollback/commit logic that the mid-tier is supposed to be doing.

Can just using (Strahl's updated) FOXISAPI.DLL be the 'holy grail' of VFP-mid-tier services, or do I just live with option 1?