Why are there so many different executable names used under Windows?
The reason the executable names vary so much under Windows is that Oracle tries as much as possible to support the installation of multiple
releases of their products on a single machine. Under UNIX that's done by creating separate Oracle Home directories for
each Oracle release. For some reason, when Oracle first ported their product to Windows, multiple Oracle homes weren't
technically feasible. The compromise
solution, which only allowed multiple releases of the client software, was to embed release numbers in the executable names.
The format used for Server Manager looks like this:
XX in the name represents the first two digits of the product's release number.
This works to keep the executables for different versions separate, but it became a bit confusing as the individual product
release numbers did not always match the database release numbers.
The various Oracle utility release numbers were usually off by a value of five in the first digit. Thus, you had:
Server Manager 2.2.x shipping with Oracle Server 7.2.x
Server Manager 2.3.x shipping with Oracle Server 7.3.x
When Oracle8 was first shipped, Oracle synchronized many of the product release numbers to match the database server release number.
One of the exceptions was Server Manager. While products like SQL*Plus jumped from release 2.3 to release 8.0,
Server Manager was simply advanced to release 3.0.
Beginning with the Oracle8i release, Oracle has figured out how to support multiple Oracle homes on Windows NT.
Because of that, there is no need to embed the release number in the executable name, and the Server Manager now uses the UNIX naming convention.
So if you are running Oracle release 8.1.5 (the 8i release) under Windows NT, your Server Manager executable will be named svrmgrl.