Fields are generally extracted at search-time, so any configuration related to that will be done on the instance you use for searching (typically the indexer in a simple deployment). So, you shouldn't need to worry about forwarder configurations.
I'm not sure I understand your first question - what do you mean by "data types" and where would the pre-definition of this be done (and, to what end)?
Splunk configuration, particularly on forwarders where the web interface is generally disabled, is most often done using its configuration files. These reside in various directories under etc in the directory where Splunk is installed.
I will read that article, but let me expand my question so that we are all on the same page.
Is there no way to pre-define what data types will be coming from a specific host via the Universal Forwarder? I did read in the documentation that the forwarder is just a dumb process, it does not parse, just sends the data blindly, so I understand that it will not do the parsing.
Secondly, how do I reconfigure the forwarder that is installed on my client? Once it was setup and running, I found no way to modify it's configuration.