Honestly, that line betrays my own apathy towards the feature. I hardly ever use them. They probably should be dead.
However! There is currently a plan (with associated scary-looking diagrams) to change how tabs behave[1].
Essentially, you'd be able to use a tab for each logical grouping of open files.
For example, in a django project, you might create a tab for each separate app you're working on. Or maybe a tab for your views, and another for your templates.
I understand that, even with that change, tabs won't fit many people's workflow. So they will always remain out of the way, unless you actually create more than one.
[1]: https://github.com/fileability/chocolat-public/issues/159
For this reason, I've pretty much forgone tabs in favor of minibufexpl.vim (https://github.com/fholgado/minibufexpl.vim), which gives me a pretty robust buffer-display window on the top of the editor and easy navigation keybindings ([shift+]ctrl+tab).
You should take a look at FuzzyFinder as an alternative to minibufexpl: http://www.vim.org/scripts/script.php?script_id=1984 I mapped <leader>b to fuzzy search my open buffers. Much quicker and less cognitive overhead for me than managing a list of buffers manually and scanning it visually for what I need.
20-25 buffers open at once? Easy!
I get that many in a day of hacking. With emacs open for a couple of weeks, I can hit a hundred pretty easily.
So the only time I accidentally :q the last buffer is when I close all files, override my search results, don't run a new search, then close the only file. Needless to say, that is pretty rare.
In text editors? Me.
I don't like thinking about which individual files are open. Any file under the root of my project (i.e. the editor's current working directory) is fair game for an open buffer. I treat open file as "go to class".