• Lossiness
  Thomas Barthel's transliteration system is lossy. The C.E.I.P.P.'s extended system remains lossy as it provides no convenient way of distinguishing between the various "adornment" appendages which are all expressed by the same suffix, ^o or s. (see the list of affixes).
• Internal inconsistency
  • Signs that may logically be analyzed as linked are assigned unique code numbers.
    E.g. instead of 400.405
  • Likewise signs that may logically be analyzed as stacked.
    E.g. 518, analyzable as 513a linked-stacked on top of 8, hence logically 513a:8 in Barthel's system, 513a;008 in the C.E.I.P.P.'s extended system.
  • Some components are assigned a code number of their own when others are not. E.g. sign 678 is the head found in signs 660 to 676 and 680 to 684, but the type-2 head of signs 200 to 299 and the type-3 head of signs 300 to 499 have no code number of their own.

Despite those shortcomings Barthel's system remains the only coherent, practical transliteration system ever proposed, without reference to which it is impossible to discuss the rongorongo.

Nevertheless, there is a need for a better system, as modular, as lossless, and as unambiguous as possible. Such a system is better built on the bases of Barthel's. Whoever is familiar with the rongorongo will have become familiar with Barthel's system, and will automatically associate a leading 2 with the head shape of the 200-299 series of signs, a leading 6 with the bird's head typical of the 600-699 series, a lone 8 with the star-like sign. Preserving and abiding by those implicit standards makes the system easier to learn and remember; breaking or ignoring them introduces confusion where Barthel had introduced a beginning of order.