Feasibility Study for Potential fP Software - Part #3 (final for the day)

[Walter Vaughan]

Jay R. Ashworth wrote:

> No, you're missing my point: *they* don't *have* to add clustering.
> PostGres has that already.
> They just need to weld the right shaped plug onto the bottom.

Exactly what I have been preaching. There *is* a vacuum that exists for the 
things that filePro has always been good at like reports that are designed to be 
printed on dot matrix printers and green screen apps.

I've spent the last few months learning a framework (ofBiz) that has a huge 
learning curve, but once the light bulb goes off it's scary powerful. Everything 
communicates via JDBC and as long as the database supports foreign keys and 
transactions it just plain works great. Heck I don't have to even keep 
everything in one database server. I could keep the order header in MySQL 5.0, 
the  line items in Slony-I (pgSQL cluster), the customer information in an 
Oracle grid, the product prices in Cloudscape, and everything else in MS-SQL. If 
I was crazy enough to do that I am guessing that it would be at worse 24 line 
changes to a xml configuration file (host, login, password, and three table 
directives per DBMS) and accounts on all those systems. But that would be it... 
the framework would do all the rest of the lifting. It will build necessary 
tables auto magically, add or delete fields as needed and handle things just 
like filepro's @cd/@cb/@ud/@ub at the row level for all tables. Each of those 
servers could be on a different continent, and it would just work.