Ciao Stefano,
I'll switch to English now.
The main difference between DBF and SQL data access is a completely different approach: when using a DBF file, your application opens the file physically and steps trough it. That means that if you skip to a new record, this records data is read from the disk, and if another user/application has modified the data, you will received the updated data (but it could be some delay for client or server side caching).
Using the bBrowser, only the records that are displayed in your grid are read. That is really memory efficient and, if not for a slow network, really fast. But moving through the data is causing data reading and network traffic (if the datasource is not on the local disk).
When using SQL data access, your application is requesting data from the server. This data is fetched to the memory (on the server or on the client). That makes it very efficient for small subsets of large databases, and, once the data is in the memory of the client, no more network traffic or data read is needed and you are working on disconnected data, so any change in the database will not be reflected on your local dataset.
Therefore I would always emphasize the fact that the change from DBF to SQL needs also always a change in the mind of the programmer because it is a completely different approach.
Wolfgang
P.S. in the last years, I have seen a change in the mind of many customers: about 20 years ago, customers requested fast incremental searchs, but now the are asking more and more for filters and not more for seeks. Therefore I'm changing many browser windows to listview windows even with underlying DBF data to permit better filters.