allbib does not know anything about macros, consistency check and other such wizardry one sees mentioned in other programs (Bibview etc.). Nor do I, by the way. I guess one can enter almost anything in the form.
allbib usually crashes when parsing files that do not resemble those that allbib creates, so don't use it with other files: you might loose time and data. It does what I ask it to do, but I did not ask too much.
allbib only checks for duplicates in citation keys, and notifies you if such duplicates were found by replacing, in the citation keys' index, the duplicates with ``duplicate''.$record_number. It does not change the buffer or the file unless you go to those records and replace the citation keys by hand.
If the citation key of the first record in the file is repeated in another record later on, allbib won't complain: the first citation key will be replaced in the index by the second one. Still, a second duplicate will be detected and pointed at. This happens because in the array that holds the buffer the first record has the number 0. I could have fixed it, but it would have made the code even less readable for me and somebody else than it already is. Until I think of a way to make it more intelligible, I will not change this.
allbib handles only one file at a time. It would be easy to have two or more files opened in the same time, but I think that a better BibTeX parser should come first on the priorities list.
The order of the fields in a record is mixed up, not respecting the order ``author, title ...''. BibTeX does not seem to care, but it's confusing when reading the file outside allbib.