I am always looking for the simplest route to getting packages of files, custom web applications and other data/software to clients. What I try NOT to do is to send them a compicated set of instructions for deployment if I am not handling installs and implementations myself.
Thus after years of using and abusing the rpm format for my own needs, I am finally getting around to using it as a distribution method. Primarily I spent a great deal of time exploring the rpm.org How To, the man files and experimenting with spec files.
The spec file is the heart of building an rpm file and controls the destinations for your data being shipped out. While not everything I package is necessarily being ‘installed’ as in Apache or MySQL — more that there are solutions I am assembling that like any web application have various destination binaries, configuration files and so on.
Ultimately the scenario is to send an rpm to a client and let their technicians know to simply ‘rpm -i myfile.rpm’ after they review a readme file detailing what is included and where the files will end up.
Some developers may find this convenient when they prefer not to be in the install and setup business. There is also an obvious advantage to this when remote clients cannot provide behind-the-firewall access to systems receiving your solutions.