DSD and DSD+ (two different programs, different authors - that we suspect, that is, but very similar in operation) are just command line executable programs i.e. they are both just a <program>.exe file and nothing more. There's nothing to really install save for needing to have the lame_enc.dll file included in the same directory as DSD+ - the original DSD has no such requirement, and the lame.enc.dll is readily available at Rarewares.org as part of the LAME bundle or at other locations online (it can't be distributed directly with DSD+ hence you need to get it from another location).
Aside from that, you would need a discriminator tapped scanner to provide the signal that DSD and DSD+ require for decoding purposes, or you can use one of these newfangled "cheap USB TV tuner sticks" aka RTL sticks most of us are really getting involved with for SDR purposes - SDR meaning Software Defined Radio. If you have the tapped scanner or RTL stick then it's a question of the audio system on the Windows computer in question and getting the audio path right so the signal goes from the source through the sound system inputs to DSD or DSD+ and then back out from DSD or DSD+ to the main audio out, typically the speakers on most people's computers, etc.
DSD will use typical inputs/outputs on a computer and can't really be configured much at all whereas DSD+ is far more configurable in that respect by allowing you to specify the input and output devices or mixers. In Windows you'd need to make sure either the Stereo Mix or the Recording Mix or the "What You Hear" mix is where DSD or DSD+ get their source from, and then it'll get piped out to the speakers (again, typically) or in DSD+'s case, another output you can choose from the ones available.
It sounds difficult and I suppose it can be in some respects but, think of using DSD or DSD+ as a simple pathway: you have to get the source (the baseband audio required to decode digital comms or control channel data) through the audio subsystem to DSD or DSD+ and then back out again to a destination so you can actually hear what those digital data streams (for the comms) represent in terms of actual voice comms. Control channel data isn't really decoded into voice output so that won't be an issue, an app like Unitrunker in particular is outstanding for such purposes.
Hope this helps get you started, and by all means read over the past few weeks of threads here in this subforum because DSD has been around for about 2 years now and DSD+ just "hit the market" so to speak in early December 2013, it's the new kid on the block but it's gaining in popularity very quickly due to it being so configurable. I still use DSD from time to time for testing decoding on some systems, but DSD+ is my go-to digital speech decoder overall.
If you have questions (and you do, or you will), just ask...