As explained elsewhere, the basis for the CMGG data is held in three rather large Word documents. To enable the user to search more flexibly (for example, on English words and concepts in the translation or the notes) and across all entries, PDF versions of all three documents are provided here, as well as the bibliography and some appendices explaining format and other details:
CMGG1 and CMGG2 are identical in format, while CMGG0 is slightly different.
In CMGG0, each table row represents one variant of a syllabogram - i.e. different variants of the same syllabogram are spread over several rows. Only sub-variants of a single variant are listed in one particular row - the row for that variant. Furthermore, there aren’t very many other columns besides the rightmost one: there’s one to say which syllabogram is being described, and another column to just give that variant a nickname. These nicknames may not have anything to do with the iconographic origin of the syllabogram - for example, there’s a u-variant designated as "two eyeballs" which probably has nothing to do with eyeballs.
In contrast, in CMGG1 & 2, each row is basically about a single logogram, with all variants of that logogram in the same row. Furthermore, in addition to rows for logograms, if a logogram has a syllabogram-only equivalent, then that follows immediately after the logogram row, in its own row. The "Type" (fourth) column records whether a row is for a logogram or for its syllabogram-only equivalent. Note that the latter is fundamentally different from rows in CMGG0. In CMGG0, each row is a single variant of a syllabogram, whereas in CMGG1, a "Type=S" row consists of combinations of multiple syllabograms, each of which is the syllabogram-only equivalent of a logogram. Or it can stand by itself as a "Type=S" row if there is no logogram equivalent of a word, where the word exists only in a syllabogram-only spelling (e.g. words like yitaj or jaay). In a few instances, such a "Type=S" row holds combinations of syllabograms and "logograms used as a rebus". Such usage makes the logogram behave "almost like a syllabogram", with one difference being that the "rebus-logogram" can have a final consonant, which syllabograms of course never have.
In CMGG1 & 2, Type=P indicates that the row holds information about a whole phrase, and Type=M indicates that it is a "mixed" entry. These entries appear in the Home page "Non-Logogram Words and Phrases" grid but will never appear in the Concordance as they represent multiple glyphs / glyph-compounds, while the Concordance shows only single glyphs (logograms, syllabograms, or logograms used as rebuses).
In CMGG2, there is also Type=D which indicates a day name and Type=B which indicates a month name. The other columns are the "CAT" (second) column which indicates the part of speech and "SUBCAT" which is a more granular classification.
CMGG0 & 1 & 2 collectively form "conceptually" the CMGG. And it’s the rightmost column of CMGG0 & 1 & 2 which is used to form the main bulk of the CMGG popups linked from the LMGGC. In the conversion of the CMGG Word column into the LMGGC popups, colour-coding information is suppressed, to keep the popups simpler and cleaner for the website user. The colour-coding however exists in the PDF versions of CMGG0 & 1 & 2 as they hold important "administrative" information. The colour-coding scheme is: