Thanks to everyone for their comments. I will to address the questions from Ghostwheel and Maiden Stone in some detail, as an example of the "reverse engineering" I am trying to do, as it may illustrate the trials and tribulations that most everyone doing Irish research will face.
I started out my family history research, about 23 years ago, already knowing the name of my great-great-grandfather, Edward Roche, a cordwainer. We also knew the name of a sister, Anne, while their parents were believed to be Patrick, with a wife named Sinnott. [We were that sort of family, great for genealogy
].
Some family members looked in the parish records (the originals), and without much difficulty found baptisms of both Edward and Anne, with parents Patrick Roche and Eleanor or Ellen Sinnott, in 1821 and 1816, respectively. So far, so good!
The issues arise at the next step. The Wexford records extend back to 1671, so obviously we wanted to trace the families further back. We found likely looking candidates for the baptisms of both Patrick Roche and Eleanor Sinnott ca. 1790, and further generations on back. However, the parish records, as is often the case in Ireland, don't include addresses (for either baptisms, or marriages), or parents names (marriages), so these identifications were never 100% certain. [But very little in Irish genealogy is truly 100% certain, anyone who thinks otherwise is kidding themselves]. In any case, we entered the names in the family tree database and life was good.
Fast forward 20 years to September last year, when I became aware that the records from a missing "hole" in the Wexford baptism records as filmed by the NLI, from 1787-1815, were actually available on RootsIreland! [Why are there so many such "holes" or missing registers in the NLI microfilm set, when the originals clearly exist?]. So, I signed up and immediately found baptismal records for an additional four children for Patrick and Eleanor, stretching back to 1803. With these additional discoveries, our previous identifications for their births (around 1790) were clearly impossible. So, back to square one.
The position now is that we have the names of the two parents, 6 children and 12 godparents, from births spanning 1803-1821, with which to try and reverse engineer if we can get a match to possible births to the parents and siblings, and their parents, back in the 1770-80s. The results of this matching/identification exercise might change if the godparents were grandparents as opposed to uncles/aunts. There is no "perfect" match, it will be a question of deciding if there are any, or one or more reasonable matches