Have some proof for that? I mean the comparison of calling and caching a webfont once vs. calling all that CSS + JavaScript + single SVG graphics. What are the loading times and execution times for both cases when used on an IPS installation? I am happy to be convinced of improvements, but I am a little sceptical. Newer doesn’t automatically equal more efficient—which is quite obvious when sticking with your car example, where tons of new features in a car might actually raise the fuel consumption. 😉
In any case, as you know, the suggested native implementation of 5 and 6 are not downwards compatible, so a large amount of theme templates needs to be touched, which creates conflicts for modified templates during upgrade. And of course every theme designer would also need to update every template in each of his or her themes that uses FA4. And then there are other 3rd-party templates, like my Pages templates, which also make use of FA4, which would all break and require updating through me and the users, which entered FA4 codes in their settings …. Considering all this, it’s no surprise should IPS decide to postpone an upgrade until a major releases (like 5.0) that might restructure and touch all templates anyway.
To be clear: Not that I have anything against an upgrade of Font Awesome. I might like it when it’s there. But looking at the bigger picture—as explained—I tend to believe that the cost–benefit analysis might not swing so easily to “overwhelming benefits” as your first post or that entertaining FontAwesome ad want to make it appear.