Genealogists researching Maine families sometimes run into a problem using the International Genealogical Index (IGI). 
When searching for a marriage record, they will sometimes find two records for the same couple, each bearing a different date, each recorded in a different town, and each transcribed from an original record. How can this be?
We in New England are spoiled with excellent town records of births, deaths, and marriages, dating back to the 17th century. Genealogists living elsewhere in the country, where jurisdiction over vital records lay with the county, and where recording requirements were lax in early years, should be envious. But, there is a price for this. Let's call it matrimonial overdetermination (or, on the other hand, let's not).
Here's an example. Charles B. Brooks of Oxford, Maine, and Roxana A. Cordwell of Greenwood, Maine, were betrothed in the spring of 1842. They were required by law to publish their intentions to marry in their towns of residence. So, in Greenwood a publishment—a publication of intentions—was entered on the books:
Intentions of Marriage between Mr Charles B Brooks of Oxford and Miss Roxana A Cordwell of Greenwood were Published in Greenwood April 26th 1842.
A similar notice was entered on the town books in Oxford, on April 30.
A couple of weeks passed, and no one objected to the marriage, so the town clerk in each town certified the couple's intentions—i.e. gave them a marriage certificate. In Greenwood, this happened on May 14, in Oxford on May 18.
The couple was married in Greenwood on May 22, 1842.
Here we have five different dates associated with the same event, only one of which is a marriage date. And it gets worse. Ministers were required to return a record of each marriage they performed to the clerk of the town where the ceremony took place. The marriage return was often dated (a sixth date) before it was recorded, at which time the clerk could affix another date (a seventh).
The Brooks marriage is also recorded in county marriage returns. Records of hundreds of other Maine marriages were returned to the state, and are now available on microfilm. And let's not forget contemporary newspaper accounts, Bible records, the records of ministers and Justices of the Peace, divorce records, pension files, etc.—all of them possible sources of marriage dates. Further, some marriages were recorded in more than one town, and could be returned to the county or state by any or all of those towns.
So, when different marriage dates appear in the IGI, recorded in different towns, often one of the records is a date of publishment or certification of intentions. One great flaw in the IGI is its failure to distinguish between actual marriage dates and dates preliminary to this (also a fault of the contributors, I suppose, who substituted an intention date wherever a marriage date couldn't be found).
There is one great advantage to finding and jotting down an ancestral couple's marriage intentions: intentions were almost always recorded chronologically, while marriage records were recorded as they filtered in from ministers and JPs in the community. Intentions are truly primary records—written down at the time of the event, probably with one or both of the marrying parties standing in the room. Marriage records, though legal and official, were copies of records kept by the officiators. (Charles and Roxana's marriage record is perhaps more authoritative than some others; they were married by the Greenwood town clerk.)
As genealogists, which should we prefer: many, possibly conflicting, records of a marriage; or one record, perhaps inaccurate, but also irrefutable? There should be no question. Despite the possibility of conflict, it's always better to have more data than less.
Then there was Columbo. He was a fumbler interested in the quirkiest of things about unsuspecting persons of interest, sidestepping his way to what he really was after. I also think he was a closet genealogist. He was forever talking about his cousins and revealing things about his family - things such as his father having been a tail gunner on a beer truck during Prohibition and his grandfather being 40 years old when he began wearing dentures.
Get out the camcorders and film them making those recipes in person. Have them show you just how to do it, so even if they don't have measurements, you can zoom in and see that it's a "jelly glass full of water" added to the pot. Or you fold the dough in thirds, then flip it "like this."
There are a few ways to monitor these search results and be alerted whenever an obituary is published that meets your criteria. Over in the left sidebar, click on Blogs Alerts to be notified by email. Or click on Atom or RSS to monitor this search with your favorite feed reader. If you use a personalized iGoogle homepage, you can also add a blog search gadget for your keywords by clicking the appropriate link at the bottom of any search results page.
. Those wanting to accurately emulate the "Register style" should seek out 
The praise was for those errors of inclusion census enumerators made while filling out their schedules, including more information than their job required—like recording town of birth, instead of just the state or country; or, in the 1900 census, giving the full date of birth, instead of just the month and year. The author listed several examples of the errors; below are a few more I've run across over the years:
