I don't see anything nostalgic about that, there's value in the benefits you list and rightfully associate to application-layer protocols.
A few months ago I stumbled upon a good article that praised them too, over today's "HTTP for everything", but I'm sadly unable to find it in my bookmarks. Anyway; more than anything, it lauded the interoperability that comes with those protocols.
I'm sending email from Thunderbird, and can reply minutes later from my iPad, and follow-up/search my archive at work from Outlook or Mail.app. Transmission merrily talks to uTorrent or Deluge or rTorrent. My windows box reads video files from a Samba share.
What now? Apps interfaces are at worst totally obscure (Skype), or at best exchange readable but undocumented JSON/XML over HTTP(S). Put differently: had email been designed & implemented today, we would hardly enjoy the same interoperability. There seems to be little interest around designing new protocols (and maybe little help coming from languages/libraries too?). IRC v3 ( http://ircv3.net/ ) comes to mind, but it's pretty niche. Anyway, I'm ranting. Can anyone complement from experience trying to do such protocol design/implementation today and the challenges associated with it?