I looked into the C++ version of cURL. While it appears productive, I decided the C-based implementation is suitably abstracted from the rest of the program through the http class I have defined. Since C++ cURL relies on C-based cURL, I decided that sticking with the latter would be more efficient from a build/compile/link/deploy standpoint. That also keeps dependencies to a minimum.
I also eliminated all references to the POCO C++ library in the areas of HTTP, networking, and file system interaction. The POCO C++ library now just handles XML parsing (which I may switch back to Apache Xerces that I used in a 2015 version of the RSS program) and minor string value tasks. That cuts the dependency on a large library I found does not handle everything well in favor of specific libraries better specialized to tasks.
The compiled size of the debug version of the program did not change. The size still holds at 1.2MB. However, observed efficiency is better. The latest update for 7/14/2018 on github represents everything stated so far. You can build it on Linux using the build script if you have development versions of libcurl, POCO, and GTKmm 3 installed. Other than tweaking which XML library I use, the plan is to work on a more detailed description of the program at michaelgautiertechnology.wordpress.com.