Sequencing with silent install

Question

When sequencing one could run through the GUI's of a setup. After that some adjustments are likely needed like deleting the uninstall/desktop shortcut and so on.
The other possibility is to use a silent setup as input for the sequencer. A lot off applications come with a vendor .MSI nowadays. This gives the opportunity to create a transform file which in many occasions can prevent the creation of uninstall/desktop shortcuts
and other customization's as needed.
Now, I'm writing guidelines on how external parties should deliver app-v packages to us. I stated that, if possible, they should deliver the command line of the silent setup used as input for the sequencer. However, they see this as the creation of two packages,
MSI/silent setup and App-V.
I wanted to check how other people work. Do they first create a silent setup and use that as input or do they launch the setup GUI and repeat what is documented in the manual installation document.
In my opinion the advantage of using a silent setup as input is that you can easily replicate a local install in case there is a problem with the appv package. It also customizes the setup during install so adjustments afterwards are not needed, resulting in
a somewhat cleaner package. The 3th advantage I see is that often you can reuse the .mst/command line when a new version is needed.
What are your thoughts on requesting a silent install from the sequencers?

Answer 1

In my opinion the advantage of using a silent setup as input is that you can easily replicate a local install in case there is a problem with the appv package. It also customizes the setup during install so adjustments afterwards are not needed,
resulting in a somewhat cleaner package. The 3th advantage I see is that often you can reuse the .mst/command line when a new version is needed.
What are your thoughts on requesting a silent install from the sequencers?
This is a perfect case of there is no wrong way to do it, but what you said above is very true.  In fact, packaging software like Flexera Admin Studio encourage you to do it this way.  They don't want you to think of a package as a single deployment
type, but rather however you want to deploy the software (MSI, App-V, etc).
The investment in time is likely the biggest hurdle.  When you have to do something simple, creating an MSI/silent command first and then creating an App-V package is 2x the work.  Its not a bad standard to go with though.  And as Tibervis
said, it does come in handy if you need to perform an upgrade or otherwise need to start sequencing from scratch again.  It is in my opinion 2 packages, just as it would be 2 packages to create an App-V 4 and App-V 5 package even if you used powershell
to convert the 4x package to 5.  This gets very into your process though. 
Just watch out you aren't making you own life 10x harder to protect yourself from something that only might happen a small percentage of the time.  Keep in mind that there are also going to be deployments where the App-V package requires work the MSI
doesn't - a driver, a component that needs to be outside the bubble, etc.  Having an MSI won't help you create the App-V in that case, and if you ask packagers to create an MSI, an App-V, and an App-V instruction guide (for packagers eyes only) you might
be shot on sight.