Currently, for ONC RPC and DCE RPC service response time statistics:
From the GUI, you select Statistics -> Service Response Time -> {ONC,DCE}-RPC..., and it pops up a dialog that lets you select the particular program and version of that program for which you want statistics.
From the CLI, you use -z {rpc,dcerpc} with, as arguments, the ONC RPC program number or DCE RPC program UUID and the version.
The first involves an extra step, selecting the program and version from the dialog.
The second involves looking up, or recalling from memory, the ONC RPC program number or DCE RPC program UUID for the program you want.
The only *advantages* I see to this scheme are that, in the GUI:
1) you could, if you wanted to, change the program for which you're getting statistics without opening a new window;
2) you don't have as many menu items.
An alternative would be to:
from the GUI, select either Statistics -> Service Response Time -> {ONC,DCE}-RPC -> {name of protocol} or just Statistics -> Service Response Time -> {name of protocol};
from the CLI, use -z {name of protocol}
e.g. -z nfs.
As for the version number, is there any advantage to *not* having service response time statistics for *all* versions being displayed or printed? If not, I could see either
1) separate tables for separate versions
or
2) a single table showing all procedures and versions
or
3) a single table, but with procedures with the same number and name being lumped together.
Any thoughts here?