Just keep in mind that the key to any storage solution is backups. Also, make sure they're automatic and/or easy enough so they actually get done regularly.
The needs of the two groups that use TTS software are different enough that the state of the software in Linux for the vision-challenged is not nearly as bad as it is for the speech-challenged.
For the vision-challenged a command line interface is a plus; for the speech challenged, it's a minus. The vision-challenged are quite willing to sacrifice naturalness of voice for speed of reading; the speech-challenged are just the opposite, wanting naturalness and not speed.
It seems that a lot of the current state of TTS software on Linux is because the developers have been targeting the vision-challenged and not paying any attention to the speech challenged at all. Let's hope that Ken's efforts lead to this shortcoming being properly addressed.
Just keep in mind that the key to any storage solution is backups. Also, make sure they're automatic and/or easy enough so they actually get done regularly.