I wonder, words like fear, dear, near and so on have long e in pronunciation, and it should be transcribed as /i:/, but I've found it transcribed as short /ɪ/. Why is that?
Vowel sounds exist along continuous spectra in several dimensions (front/back, high/low, rounded/unrounded, tense/lax, r-colored, etc.). The symbols we use to denote phonemes or even allophones are points within those spectra. Therefore interpretations will often diverge between transcribers. Furthermore, individual speakers offer a lot of variation in pronunciation due to dialect and idiolect. It is natural to expect some imprecision in a system of single symbols for what amounts to a statistical condensation of wide-ranging individual sounds.
In a comment, John Lawler wrote:
In American English, for instance, the /i/ ~/ɪ/ phonemic distinction, like most tense/lax vowel distinctions, is simply neutralized before /r/. There are no contrasts, so there is near-free variation. Which, of course, means that that variation is probably available for ingroup identification, though I don't know of any studies showing that. Yet.