I have Sonar 8.5, but I don't think it matters as this should apply to any version. I seem to do reverb differently every time I start a project. Sometimes I use the native reverb on a sample library, plus Spaces; sometimes just Spaces. Sometimes I put a send on each synth in Console View, and adjust the amount of verb going to the Spaces bus; sometimes I put Spaces on each synth. Is there a best-practice way to do this, or is this another case of "There's a lot of ways to do this?" 

Michael, 
There ARE indeed a lot of ways to do this, but it's more about what you want to accomplish by adding reverb.
 
Are you trying to give some space or airiness to a track???  Using reverb as an insert can help there (not the only solution, but A solution)
 
Do you need to put multiple instruments in a room or space?  Try a verb bus and "send" each instrument to the bus.  
 
Reverb as an effect?  Try a bus (or Aux track) and automate... 
 
FWIW, find one (or two) ways to do it that you can call up quickly.  Make them part of your workflow and then experiment with automation.  It's a lot easier to work with a few ways then to try to learn all the ways... 

arlen2133 pretty much nailed it. I'd just add that we've been conditioned on what to expect when we hear sound in a acoustic space. I prefer the "bus to reverb" concept to create a cohesive space. I feel that too many reverbs create a sound that, consciously or not, we may hear as "phony."

Anderton
arlen2133 pretty much nailed it. 

Aww thanks Craig.. It's just Arlen though.. Plain ole Arlen. 

Thanks Arlen and Craig. To be more specific in what I'm trying to do, I compose for orchestra. (nothing special, just a hobbyist). but I need to blend different libraries, currently GPO4, EWSO Gold, Dim Pro, Cinematic Strings, and soon to be, GPO5, which will be out sometime this month. I was using Vienna Special Edition, but gave up on it because it just doesn't blend well and causes other problems. Each of these has its own native reverb. I usually try to get all these native verbs equal as possible, then put them all in Spaces on a bus. EWSO is a wet library, GPO is dry, but has very nice convolution reverb, so I try to get it equal to EWSO. But sometimes I will not use the native verb, and instead use just Spaces, with sends. sometimes I will group all the woodwinds on one synth; all the brass on another, etc. Or I will just use one send for each synth, as all instruments on that synth pretty much have the same verb or lack of it. I know I seem to be making it more complicated than it really is, but when dealing with the orchestra, it really is more complicated. But there must be a way of reducing all this complexity to one basic approach that will work on all my projects. 

I found this from one of the other threads listed above. It's from IMF in Sweden. I hope it's OK to quote him:
 
"Let each string library track have it's own sends - since you need individual control.These sends go to a reverb bus for strings.
The return of this reverb bus  go to strings submix bus.
 So what you have then is ability to raise submix bus fader and still get the same reverb mix - what is usually handled through VCA groups in some daws. In this manner you don't need that.
 And if you want to raise reverb individually on a single library you do that as well.
Or group them if you want to move them all.
There are very advanced abilities in Sonar for grouping - if to move by dB or movement etc."
 
This makes sense to me. One of the problems I am having is with volumes getting out of control. This seems like a good way to deal with that. 

I think that can work and is pretty much on the basis of what Craig and I were describing.  The only added feature would  be the submix bus.  
Remember you can do all of this via an Aux Track and simply put that Aux/Reverb submix track in the folder with the library.  Keeps things nice and neat.

I work a lot like Arlen and Craig.  I never use the patch reverbs, I keep them dry.  I send them all to a reverb bus with what ever reverb I select set to 100% wet.  Then in the console view, I find it best to do this there, I select how much reverb I want to add to each instrument(s), i.e. usually more on the high strings and less on the bass strings.  This makes it sound more realistic IMHO.  YMMV.

One other advantage of using a centralized resource for reverb is processing overhead. Reverbs will eat up resources so having one instance on a bus or aux track is preferable to several instances scattered around.

Reverb can also benefit HUGELY if you EQ it, either before the plugin or after, sometimes even both.
 
You can also compress the output of the reverb to enhance/suppress the tail, increase the overall level, anything really.