Relying on transcriptions is always a dicey proposition*, and around the web and throughout the Irish genealogy community, there has always been annoyance that you had to pay so much to access just a transcript. But I have never been as upset about it as I am now that I have access to the originals.
The RootsIreland website clearly states that "The records have been transcribed directly from the original Parish registers and Civil records in Ireland." Somehow it never occurred to me that it was strange that of the dozens of "records" I had paid to access, and hundreds of search results I had seen, all were in English.
Now that I've seen the originals, I can tell you that it's very strange that all of the "transcriptions" on RootsIreland are in English, because a substantial number of those are so-called transcriptions of Latin records. Those aren't transcriptions at all, but translations. I can't find any mention on the RootsIreland website of the fact that many of the records they provide are no longer in their original language. In fact, on their page about first names, they list the Latin names among "common variants," along with nicknames and abbreviations. As in, "Even if you knew your ancestor as Charles, don't be surprised if he shows up as Carolus instead." But never, in all my searching, did I find a Carolus, even though I'm now discovering that at least some of the results I was looking at came from registers written in Latin.
Take, for example, the sister of my 3x great-grandfather Richard Toner. She was born in 1828 and the IFHF had told me that she was Mary, daughter William Toner and Margaret Walsh. But here is an image of the parish register that recorded her baptism:
 |
| 20 April 1828. Baptism of Maria Toner. Maynooth Parish, Co. Kildare, Ireland. registers.nli.ie |
She is clearly recorded as Maria, and every other first name is in Latin, too. This is a particularly problematic example because there were plenty of women in Ireland named Maria. Searching the 1901 Irish Census returns over 21,000 women who gave their first name as Maria. 639 of those women were born within 5 years on either side of 1828. So while I'm fairly certain that my 4x great-grandfather was not called Gulielmi in his daily life, I can't actually assume, on the basis of this record, that his daughter was named Mary and not Maria.
I don't know why the Irish Family History Foundation thought they could make that assumption, and try to sell me the record without being open about the assumptions - and changes - they were making.
*And I should note that responsible researchers didn't "rely" on them, of course, but for many areas accessing the originals or microfilmed copies was difficult or impossible unless you were in Ireland.