The concept of backwards compatibility lends itself nicely to illustrating one of the possible problems of introducing spelling reform. When changing the system, you not only need to worry about teaching the new one to children learning to read, but to the adults who are already proficient, imperfectly or not, in the old one.
One of the main problems is one I call “collision”. Basically, it may happen that because of the introduction of a new rule, you end up with a sequence of letters which represented a certain sequence of sounds in the old system, but is intended to represent a different sequence of sounds in the new one. Someone proficient in the old system is likely to interpret the sequence of letters in the old way. This does potentially happen in three cases in my spelling reform proposal for Spanish.
1- Because of the elimination of “tildes diacríticos” (accents like áéíóú which are used in some monosyllables, that as a general rule do not take these accents), words like “se” and “sé” come to both be spelled “se”. There is no difference in pronunciation on the surface, but this can affect sentence stress. “Se” in the old system represents a pronoun, which can carry no sentence stress; “sé” represents a verb, which usually does. As a consequence, in the new system, it can take slightly longer to interpret “se” and thus to figure out how to read the sentence with the appropriate prosody.
2- Because of the change from GÜE GÜI to GUE GUI, which in the old system represented /ge gi/ instead, there might be cases where the sequence is interpreted in the old way.
3- Because of the change from GUE GUI to GE GI, which in the old system was a variant of JE JI and thus represented /xe xi/, the same problem as in 2 is liable to happen.
However, upon closer examination 1 and 2 are probably not as problematic as they seem to be. Regarding 1, it’s very often the case that “tildes diacríticos” are omitted, either to save time, because of technical problems when typing, or because of spelling mistakes. Native speakers seem to manage to read texts written in this way (and read them aloud) without many problems. At least, with no more problems regarding sentence stress than usually arise when reading texts aloud.
Regarding 2, words with the sequences GÜE GÜI are relatively rare in Spanish, and like with 1, many times the “diéresis” (umlaut) is omitted, for the same reasons. Literate speakers will likely have little trouble reading these words aloud.
3 is the only case where a real problem might present itself. This should be easy to test with an experiment where native speakers are asked to read aloud a text written in the new system.
(The possibility of 2 or 3 being a problem rises when the word being read is unfamiliar to the reader.)