• case study: transfer of GEDCOM to new structure
    • Notes on source code
      • partial download from: [1] -"Not found" in Feb 2017
0 @I130@ INDI
1 NAME Sheldon Riley /Ferree/
2 GIVN Sheldon Riley
2 SURN Ferree
1 _UID 632050E023F3BF4B8FDC306DCE60124C6783
2 DATE 28 Mar 2007
1 SOUR @S58@
2 DATE 13 OCT 1849
2 PLAC Mount Vernon, Knox Co., OH
2 DATE 2 FEB 1918
2 PLAC Peoria, Peoria County, IL
2 SOUR @S39@
3 QUAY 3
3 PAGE Certificate No. 0040025, Date Filed 3-Feb-1918
3 NOTE Personal recording of database record
1 OCCU Drayman and Rachel was Keeping house
2 DATE 1880
2 PLAC Benton, Benton Co., IA
2 SOUR @S60@
3 QUAY 3
3 PAGE Microfilm Roll T9_327, Page 216B, Image 0054, Family History Film 1254327
3 NOTE Personal recording of microfilm database record
1 OCCU Carpenter
2 DATE 1910
2 PLAC Belle Plaine City, Benton Co., IA
2 SOUR @S65@
3 QUAY 3
3 PAGE Microfilm Roll T623_417, 10B, Ward 2, Enum Dist. 2, Image Not Stated
1 OCCU Drayman delivering oil and Rachel was keeping house
2 DATE 1900
2 PLAC Belle Plaine City, Benton Co., IA
2 SOUR @S64@
3 QUAY 3
3 PAGE Microfilm Roll T623_417, Page 10B, Enum Dist. 2
3 NOTE Personal copy of record
1 OCCU Farm Laborer
2 DATE 18 JUN 1870
2 PLAC Warren, Poweshiek Co., IA
1 RESI with his sister Mary L. Ferree
2 DATE 18 JUN 1870
2 PLAC Warren, Poweshiek Co., IA
2 NOTE In this census record taken June 18th Sheldon 21 yrs and his sister Mar
3 CONC y 17 yrs are living together.  Mary is working as a domestic servant. S
3 CONC heldon lists his Value of Personal Estate to be $275.  Census taken by A
3 CONC . J. Wood
2 SOUR @S59@
3 QUAY 3
3 PAGE Microfilm Roll M593_396, Page 247, Image No. 502
3 NOTE Personal copy of record
1 EVEN Union side
2 TYPE Military
2 DATE 7 Mar 1865 to 20 Jan 1866
2 NOTE Name: Sheldon Ferree
3 CONT Enlistment Date: 07 March 1865
3 CONT Side Served: Union
3 CONT State Served: Iowa
3 CONT Unit Numbers: 184 184
3 CONT Service Record: Enlisted as a Private on 07 March 1865 at the age of 18
3 CONT Enlisted in <u>Company A, 12th Infantry Regiment Iowa</u> on 07 March 1
3 CONC 865.
3 CONT Mustered out <u>Company A, 12th Infantry Regiment Iowa</u> on 20 Januar
3 CONC y 1866 in Memphis, TN
3 CONT <b>Regimental History
3 CONT </b>
3 CONT Twelfth Infantry IOWA
3 CONT (3 years)
3 CONT Twelfth Infantry. Cols., Jackson J. Wood, John H. Stibbs;
3 CONT Lieut.-Cols., John P. Coulter, Samuel R. Edgington, John H.
3 CONT Stibbs; Majs., Samuel D. Brodtbeck, Samuel R. Edgington, John
3 CONT H. Stibbs, Edward M. Van Duzee, Samuel G. Knee.
3 CONT This regiment was organized at Dubuque in Oct. and Nov. 1861,
3 CONT and was mustered in at intervals during those two months. It
3 CONT left the state late in November, went into quarters at Benton
3 CONT barracks, St. Louis, for two months, and like its
3 CONT predecessors, suffered greatly from diseases that seemed to be
3 CONT a part of the experience necessary to prepare them for the
3 CONT field work. Some 75 died of measles, pneumonia and typhoid.
3 CONT At Smithland, Ky., it joined Grant for the movement upon Fort
3 CONT Henry, was present at the capture of that place, then moved to
3 CONT Fort Donelson, where it took part in the fight and assault
3 CONT which resulted in victory. It won immortal glory for itself
3 CONT at Shiloh by fighting in the advance until sundown and holding
3 CONT back the enemy while the demoralized army withdrew to a new
3 CONT point and waited the arrival of Buell. The 8th, 12th and 14th
3 CONT IA comprised four-fifths of that advance line and surrendered
3 CONT only when surrounded by ten times their numbers. (In the
3 CONT history of the 8th will be found an account of the disposition
3 CONT of the prisoners until their parole and exchange.) Those who
3 CONT escaped capture were assigned to the "Union Brigade" and
3 CONT served with it until disbanded, being sent to Davenport, Ia.,
3 CONT and remaining there during the winter. The paroled men were
3 CONT declared exchanged Jan. 1, 1863, and soon after went to Rolla,
3 CONT which was threatened by Marmaduke, but returned on the 15th to
3 CONT St. Louis, where they were stationed. Lieut.-Col. Coulter
3 CONT resigned and was succeeded by Maj Edgington, and the latter as
3 CONT major by Capt. John H. Stibbs of Co D.
3 CONT The regiment was reorganized about April 1st and became a part
3 CONT of Sherman's command, participating in the movements of that
3 CONT division during the Vicksburg campaign, though it was in
3 CONT reserve at the assault of May 22. After the surrender it was
3 CONT engaged at Jackson and was in the skirmish at and capture of
3 CONT Brandon. It went into camp near Bear Creek on July 23 and
3 CONT remained there until Oct. 10. Lieut.-Col. Edgington resigned,
3 CONT Maj. Stibbs became lieutenant-colonel, and was succeeded as
3 CONT major by Capt. Van Duzee.
3 CONT In October the regiment was in a skirmish at Brownsville;
3 CONT proceeded thence to Vicksburg, Memphis, Lagrange and Chewalla,
3 CONT where it remained on railroad guard duty until near the close
3 CONT of Jan. 1864. While here it broke up the guerrilla bands that
3 CONT were pillaging the country, and built a strong fort. It was
3 CONT ordered to join the forces for the Meridian raid, but reached
3 CONT Vicksburg too late to take part and went into camp.
3 CONT Having been mustered in as a veteran organization, the
3 CONT reenlisted men were sent home on a furlough in March. In
3 CONT their absence, the non-veterans, numbering about 70,
3 CONT accompanied the 35th IA on the Red River campaign and was in
3 CONT battle at Lake Chicot. On their return from home the men
3 CONT reached Memphis on May 2 and were joined by the detachment
3 CONT about the middle of June.
3 CONT In May six companies under Lieut.-Col. Stibbs, went to the
3 CONT mouth of the White River, established a military post and left
3 CONT Cos. A and F under Capt. Hunter. The command proceeded to
3 CONT Tupelo, where it was engaged in July. The regiment while
3 CONT acting as a train guard, was attacked by a brigade, but
3 CONT repelled it in a handsome manner, and in the subsequent
3 CONT fighting it occupied the most dangerous post and received
3 CONT special commendations of the general commanding.
3 CONT Returning to Memphis, the regiment moved to Lagrange, thence
3 CONT to Holly Springs, via Lumpkin's Mills, remaining on duty there
3 CONT for some time. The detachment at White River in the meantime
3 CONT had been busy, protecting the loyal people of that section and
3 CONT building a stockade. The little force of but 47 was attacked
3 CONT before daybreak on the morning of June 5 by a force of 400,
3 CONT the men being compelled to fight in their shirts only, so
3 CONT sudden was the attack. A number of the enemy gained the
3 CONT stockade at one side, but Sergt. Isaac Cottle and Corp. George
3 CONT Hunter, armed with revolvers, boldly attacked them and drove
3 CONT them out in confusion. Hunter was shot dead and Cottle was so
3 CONT severely wounded that he died soon after, but the entire
3 CONT besieging force was finally driven off with a loss of over 50
3 CONT in killed, wounded and prisoners, their commanding officer
3 CONT being among the slain.
3 CONT Joining the regiment at Holly Springs, this detachment
3 CONT accompanied it to Oxford, then to Memphis, whence it proceeded
3 CONT to Devall's Bluff and Brownsville in search of Price. With 10
3 CONT days' rations it made the 350 miles march to Cape Girardeau
3 CONT via Jacksonport, Ark., and Jackson, Mo., in 19 days. From St.
3 CONT Louis it proceeded to Jefferson City, Smithton, Sedalia,
3 CONT Lexington and Independence, into Kansas, and to Harrisonville,
3 CONT Mo., after Price but was unable to catch him and returned to
3 CONT St. Louis.
3 CONT The non-veterans and some of the officers were mustered out,
3 CONT Lieut.-Col. Stibbs remaining as commanding officer. Moving to
3 CONT Nashville the regiment aided in the defense of that city and
3 CONT in the battle in December captured 2 flags. It joined in the
3 CONT pursuit as far as Clinton, then proceeded to Eastport, Miss.,
3 CONT where it assisted in building quarters and fortifications.
3 CONT Lieut.-Col. Stibbs was called to Washington in Jan. 1865, to
3 CONT become a member of the military tribunal, the same, which
3 CONT later, tried the notorious Capt. Wirz, who was held
3 CONT responsible for the infamies of Andersonville prison, and Maj.
3 CONT Knee took command.
3 CONT The regiment was ordered to Mobile in February, was engaged at
3 CONT Spanish Fort in the front line and occupied an exposed
3 CONT position for 13 days and nights. At the conclusion of the
3 CONT siege of Mobile it moved to Montgomery, thence to Selma and
3 CONT remained in guard and garrison duty until the early part of
3 CONT 1866 when it was mustered out. Lieut.- Col. Stibbs received a
3 CONT merited promotion to a colonelcy. The original strength of
3 CONT the regiment was 926; gain by recruits, 55, total, 981.
3 CONT Source: The Union Army, vol. 4
3 CONT <b>Battles Sheldon participated in:
3 CONT </b>Fought on 27 March 1865 at Spanish Fort, AL.
3 CONT Fought on 29 March 1865 at Spanish Fort, AL.
3 CONT Fought on 06 April 1865 at Spanish Fort, AL.
1 EVEN had a photo taken
2 TYPE Misc
2 DATE 1918
2 PLAC Peoria, Peoria County, IL
1 EVEN was listed as a carpenter in the biography listing prepared by Chicago Western Historical Co.
2 TYPE Misc
2 DATE 1878
2 PLAC Belle Plaine, Benton Co., IA
1 FAMS @F33@
1 FAMC @F35@

What hope for standardization?[]

First 15 lines of the above GEDCOM produce an inconsistency:

0 @I130@ INDI
1 NAME Sheldon Riley /Ferree/
2 GIVN Sheldon Riley
2 SURN Ferree
1 _UID 632050E023F3BF4B8FDC306DCE60124C6783
2 DATE 28 Mar 2007
1 SOUR @S58@
2 DATE 13 OCT 1849
2 PLAC Mount Vernon, Knox Co., OH
2 DATE 2 FEB 1918
2 PLAC Peoria, Peoria County, IL

So there's a lot of manual work needed in any really smart conversion. Easy enough, I guess, for you programmers to knock up a few thousand conversion tables for common things, such as "Co., OH == County, Ohio", but individuals' names will be a minefield, as we have discussed in different ways and will do more of. Fred and Rtol could have problems with line 3 if it was a Dutch name, because the "Riley" might not be a "Given". — Robin Patterson (Talk) 06:04, 7 May 2009 (UTC)

These are the realities of what we are up against. It is a tough nut. A lot of these files don't have the highest quality either- some genealogy enthusiasts view this as a game where the one who dies with the longest list of ancestors wins. I am going to be mucking about with some of case studies. If you come across any Gedcom individual data records like this one that are especially rich in volume of information and variety of fields used, that would be great. -~ Phlox 06:20, 7 May 2009 (UTC)
You got my 13,000 people a while ago. Most of them have at least a sort of source (mostly just a reference to my brother's printouts but a few notes about IGI and electoral rolls and probates), and many have Chapman codes (and non-UK equivalents) in their place names. Otherwise probably not as good as this example. I could make you a very small one (from FTM v11) as an extra if it would really help. I trust that you are keeping in touch with User:Rtol on this subject, or at least watching his reports, as he nears breakthrough with his info-oriented improvement to Yewenyi's process. — Robin Patterson (Talk) 13:18, 7 May 2009 (UTC)
Chopping up the location names will probably be a lot lot easier that cracking some of the other fields that are also have important bits that need to be identified. Most critical for serious genealogy is the sources field. Place is easier because there is a strict hierarchy that narrows the possibilities. It makes the guesswork easier. Once you know the place is scotland, you know that a three or four letter abbreviation could be a chapman code. US state abbreviations are fairly uniform, and so once you know Ohio, you can fetch the list of counties (Knox being one of them), and then it knows the county and so on. That sort of thing is indeed table driven, and interestingly, the SMW extensions makes these values directly accessible by tools. Sources is a lot harder, due to the variety of formats, and the tendency of many serious researchers to use their own idiosyncratic shorthand. We can certainly figure out URLs. It's all guesswork though, and these tools will make mistakes. In these cases it is nice to have the original data from the Gedcom right there so a human can check it if necessary. By convention, each GEDCOM INDI record might a /version subpage which would be just like an info page. On the talk page, we could stick the original INDI record in a <pre> block. Or we could stick it in a subpage of the /version page. Makes no difference really. Anyway, then each version has a number and we can collapse the duplicates and merge data into whichever version has the best collection. This then becomes the dominant version that is treated in the manner that we treat info pages now. Hope you are following me. If not, no matter because it may change anyway. But that seems to be the front runner idea at this point in the horse race. -~ Phlox 16:23, 7 May 2009 (UTC)