Hi List, I actually responded to LARC's original gedcom ? to her directly on FRI thinking I had sent the post to the List but somehow I missed sending it and so I will now make my case to the list and then promise to stop posting,,,for today. Matt, I see that you have worked out the merge/append process in a way to "best avoid" duplicates. Others have likely not developed such a process but should try to do so. However, even a system which "at best avoids" some duplicates just may not be "good" enough. Here is the dilemma as I see it...and why I never went beyond the 500 people in my tree whom I Knew belonged in their respective families. And the reason why I don't feel comfortable with the compilation of data which gets passed on, and passed on and passed on...then posted on the internet and yet when you contact "the contact" person for a correction you get "No Reply at All." In my family line (like I suspect all) the practice of repeating names and naming children after one's siblings leads to a string of first cousins with the same first name born in the same locality all within the same decade. If you do not know for sure the "Maiden Name" of the spouse they married the chance that you will be able to determine who the parents or children are is nearly impossible. Here is my (pseudo-real) example....this made up dates for example 3 Bros. T, N, & W all born late 1760s, all marry late 1780's all have 3 sons each born in 1790s, all named for the original 3 Bros. So you have T(1760) with T(1791) N (1793) W (1795) N(1760) with T(1792) N (179?) W (1796) W (1760) with T (1790) N (179?) W (1795) If all survive by 1815 you have 12 adult males all with the same surname; 4 groups of 3 with the same first and last name all capable or marrying/remarrying/reproducing. so in 1815 if the 9 first cousins (let's assume the Dad's don't widow, remarry & reproduce) marry and each have 3 sons by 1820 you now have 27 second cousins. 9 T's (born bet 1815-1820) 2nd cous 9 N's (born bet 1815-1820) " 9 W's (born bet 1815-1820) " plus their fathers 3 T's (born abt 1790) 1st cousin 3 N's " 3 W's " plus the grandfathers 1 T 1 N 1 W So here you have the possibility of 40 related males with the same surname in the same locality and of those 40 there is the possiblity that there are 3 groups of a dozen each sharing the same first and last name... How without knowing the Mother's Maiden names of generation # 2 born bet. 1815-1820 can we (with any certainty) place these guys in the correct family without "ASSUMING" they are duplicates when in reality they are likely cousins. As a hobby I think it is okay to create the database, tracking folks, tracing lines, gathering clues and especially for your own line where "your memory" and "care with data" is likely better. I also feel that share and share alike is the basis for this hobby and whose to say whose line is whose line when you see the surnames marry and blend overtime. But I know there were perhaps a Minimum of 3 adult Nicholas Kennedy's in Crocker's Cove between 1783 and 1807...one married to a Elizabeth King, one to a Mary, and one to Grace Young and without an Obituary or a Gravemarker I cannot assume it was the Same Nicholas whose wives happened to have poor luck with longevity. By assuming Nicholas (abt. 1765) perhaps was married and widowed 3X and thus blend the families we run the risk of eliminating 2 other families, likely those of his first cousins. This is where I have a problem with merging, duplicating, appending, blending the data because overtime people will accept it as "record" rather than contrived data based on what little information is available. Now you may think we'll never really know with the time frame being so early...and this is even more reason Not to blend families without the confirmation of a Maiden name or exact birth and death date. I am not opposed to them being entered into a database...for it would be impossible to keep track of them any other way but Verify, verify, verify ..... X-Message: #2 Date: Sun, 27 Jul 2003 16:50:42 -0400 From: "Matt Mullaly" <2matt@rogers.com> To: NFLD-LAB-L@rootsweb.com Message-ID: <013c01c35480$becc1a20$34847018@rchrd.phub.net.cable.rogers.com> Subject: [NFLD-LAB] APPENDING/MERGING GEDCOMS Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: 7bit Brendan, See below for a few comments re how I prevent duplicates when I MERGE incoming GEDCOM data into my main file. ----- Original Message ----- From: "Brendan Hynes Jr." <mclaclan@sympatico.ca> To: <NFLD-LAB-L@rootsweb.com> Sent: Sunday, July 27, 2003 1:27 PM Subject: [NFLD-LAB] Gedcoms....pros and cons.... > I used to be a firm believer in using them as a means of data sharing, > in fact I recall once even suggesting that perhaps we should create a > page on the sites just for them... > > Having however worked with a couple in the past I find myself of > differing opinion now.... > > While it's true enough that they make things easier in the entering of > multiple names in a few mere clicks of the mouse even the finest of > merging operations won't nescessarily catch all the possible duplicates > and one can quite easily discover after the fact that X number of names > and branches are now duplicated in one's database. Brendan, here's what I do: When I receive a new GEDCOM, I load it into my Family Tree Maker (FTM) thus creating an FTM file and examine it. Then I: 1 - Use the FIND and REPLACE function in FTM to massage certain incoming data to fit MY format. e.g., I change all the NEWFOUNDLANDs, NF, NL, etc. to simply Nfld., the abbreviation I use. 2 - I assign a birth DATE (an ABOUT birth date) to EVERY person in the incoming file that dosen't already have one. MORE ON GUESSED DATES BELOW. To do this I use a little formula that has worked well. If you are interested let me know. 3 - If the incoming SURNAMES are in the format I use (Joseph Roberts SMALLWOOD), I again use the FIND and REPLACE to change the format to Joseph Roberts Smallwood. By the way, Joey is in my file as grandmother was a COOPER from Lower Island Cove, in my neighborhood and his wife was a OATES from up the road in Carbonear. MORE ON SURNAMES BELOW 4 - Of course, I back up my main file before I allow anything to be added to it. 4- When I actually get to the APPEND/MERGE operation , I use the APPEND ONLY option. This prevents FTM from MERGING anything on its own. All I've done is add the incoming data to my file without MERGING anything. 5 - Then I go through each incoming person - one-by-one - beginning with the first person in a family branch. Before I add each one, I look for possible duplicates based on either real dates or the ones that I've assigned. SEE # 2 above. As I do that, I mentally add and subtract about 10 years to/from those dates that I've guessed at. And I also look at spouses and anything else that may indicate a duplicate. 6 - As I add each person permanently to my file, I change their surname (now in lower case - SEE # 3 above) to upper case to fit my format. 7 - Before I finish, I look at my main index to make certain that there are no lower case surnames in there. What I've actually done is to use FTM to massage and APPEND the data and perform the manual MERGE function as described. By the way, it's more difficult to describe this process than to actually do it. I've been using (and refining) this method for years now and it works for me. So endth the lesson - hey it is Sunday. <big grin> Cheers. Matt