Everyone I've encountered in China seems to do it just fine. It is only hopelessly slow if you type one character at a time.
> "adopting an input method based on the character radical composition or handwriting recognition is much faster"
What character radical input method are you talking about? I can't imagine any input method using radical composition being faster, it would be comparable to using Latin roots to type out English words; there are just too many to be practical. You can actually do this on Pleco, but it takes forever, even longer than typing one character at a time in pinyin.
Handwriting could work assuming you can handwrite Chinese quickly (I would be deathly slow as I can't read a typical handwritten cursive note, much less write one). I think handwriting recognition is already pretty good, people just use pinyin because it's the easiest.
Over the years I've tried to teach myself, and I can sort of do it, but I never got fast enough to really switch away from pinyin.
Also, pinyin IMEs can be fairly fast even with only minimal practice as the matching is done at the word level (and possibly some IMEs look at preceeding characters) rather than at the character level. For example, when entering 南方人, my tying sequence was "nanfangren<space>", and the IME was already suggesting the correct answer at "r", so I could have actually typed "nanfangr<space>".
I dont recall seeing any of my friends or colleagues here using this feature.
I am a foreigner living in China and this was never an issue for me, so I doubt the native people here find that a problem.
I rarely see any Chinese people use the handwriting character recognition on their phones. I use it sometimes just for practice, but just writing pinyin is much faster.
[0] - https://en.wikipedia.org/wiki/Kodansha_Kanji_Learner%27s_Dic...