BRIEF EXPLANATION OF NUMBERING SYSTEM
by Paul Stoneburner

The numbering system used in this genealogy appears at first glance to be cumbersome, confusing, and perhaps even awkward. Let's discuss some numbering systems and clear up what is really going on here.

Unless you have purchased the entire database, covering several disks, what you have here is a rather small subset of collected data. Furthermore, this disk contains family sheet printouts which are intended to facilitate acquiring more data by providing but one family per page, with lots of white space for notes, and not the much more compact database format.

This information is from a genealogical software package called GEAN which was expressedly designed to concurrently process data on multiple families and record all the interrelationships found therein. In semi-landlocked areas such as the Shenandoah Valley, families are much more interrelated that one would at first imagine. If the data were available, GEAN software is capable of recording all the descendents of every pioneer or latecomer to settle in the the valley, showing all the intermarriages and various relationships between any and all descendants! The fruit of my project, when eventually published (in electronic format) will NOT be a genealogy of some family, but a tool for genealogists tracing families which have their roots in the Great Valley. The numbering system used in such a collection should have two objectives: it MUST provide everyone a unique number to keep the computer happy, and, ideally, it should be human useable, i.e. it should "make sense." Therefore,...

First and foremost, this is a developer's numbering scheme and was intended solely for use while collecting information, not necessarily for use in any final publication. Naturally, every person must have a UNIQUE number -- implicit in that statement is that we cannot have unnumbered people running around in the pages of our notebooks. It is simply a fact of life that, even if a person's full name and date of birth is known, that is not sufficient for uniqueness. As proof, consult any of the numerous people who have been arrested because their name and DOB on their driver's license happens to match those of someone else whose name and DOB are on a warrant. And, just as obvious, if the number is not unique, it is worthless.

What about the good old 1, 2, 3..., why not use it? Well, in fact, some people do do that. And, even if we were tracing only a single family, it leads to one of two undesirable consequences while collecting information. Either the numbers jump around all over the place so that, while it is a number for uniqueness, it conveys no other information -- such as what family this person belongs in; or, the numbers are continually being changed so that, for example, if you insert someone new after person # 428, that new person becomes # 429, while the old # 429 becomes # 430, old # 430 becomes # 431, etc. That really happens in some early computer-based genealogy systems, and I find it ugly as sin, particularly because you can never reference anything by that "unique" number since it keeps changing out from under you.

OK then, here's a real simple system which a lot of people have used in the past. It works and anyone can do it. Start with the progenitor of the family you are working on and simply number each of his children. I said EACH of them. No skipping numbers just because they died young or never married. Then, for their children (the progenitor's grandchildren), add one more digit, so that # 32 would mean the second child of the third child. His children, in turn, would be # 321, 322, ....

There is one slight flaw in the above, which is that (particularly in the "old" days) people tended to have more than nine children. Never to be deterred by something as trivial as that, the adherents of this scheme just add a few parentheses as required. So, # 4(12)1 would be the eldest child of the twelfth child of the fourth, whereas # 4121 would be the eldest of the second of the eldest of the fourth.

We can improve on that a little by noting that few people have more than twenty-six children. So by simply substituting letters for numbers we get rid of the parentheses. # 4121 is now DABA and 4(12)1 is DKA. Note a little serendipity here. We have just provided a means of giving the spouse a number as well. Use capital letters for the descendants of the progenitor and lower case for their spouses. We know immediately, for example, that "dka" is married to DKA, and we already know who he is, right?

Furthermore, the system is ever expandable. People can go right on having babies and getting married (not necessarily in that order) and we can fit them in neatly without the need to ever change an existing number.

It only remains to consider a complication or two. GEAN supports concurrently working on multiple lines and we still need uniqueness in the identification numbers. The solution used is to make up some sort of prefix representing the progenitor of the family and put it in front of the alphas. Any old prefix would suffice, but these have two parts, separated by a colon, and the whole thing is then separated from the alphas by a virgule (slash) inserted for readability. Some of these "numbers" use the first three consonants of the last name and then a number which is sequentially assigned as required because those first three consonants are not necessarily unique.

Example of XXX:NN format: WLK:2 happens to be some Wilkins. Then, if that DKA we were using earlier happens to be his descendant, the full number of the descendant is WLK:2/DKA, his spouse, as expected, is WLK:2/dka, second child is WLK:2/DKAB, etc. You get the idea. Historically, because numbers take less space in a computer than letters, we also use numbers of the form number:number, such as 48:27 or 51:3. Same scheme. We can also have a 51:3/DKA who is quite a different person from WLK:2/DKA or 48:27/DKA or 46:143/DKA. These numbers require slightly less space than letters in a computer and when GEAN was originally developed in 1974 that was of paramount importance.

We now have to consider what to do with multiple spouses (I really wish that word were "spice." I sounds classier, or something.) By convention only, the number:number type of identification discussed above will always have a two digit number for the first part of families being traced. Then we can use the single digits to assign numbers to "unrelated" persons such as that extra spouse or the parents of spouses. Nothing can be inferred about a relationship between 3:24 and 3:25, there are just two numbers we happened to need. One need only keep track of the last one used so it isn't reused again on someone else by mistake.

Finally (!), let's make use of that virgule we inserted for readability. We can sneak a little more information in there if we think about it. Adopted children are usually included in family groupings -- if for no other reason than so people won't think you didn't know about them. If we reverse the slash (to make a "\"), we can use that to show that an adoption (or "raising") occurred along the line. Suppose our WLK:2/DKA's first child was adopted, but the second was not. Their numbers would be WLK:2\DKAA and WLK:2/DKAB. Once such an adoption, or break in blood line occurs, it is kept forever in the numbering.

WLK:2Person we are tracing
WLK:2/DHis fourth child
WLK:2/DKTwelfth child of fourth
WLK:2/DKA1 of 12 of 4 of WLK:2
WLK:2\DKAAAdopted child discussed above
WLK:2\DKAABSecond child of adopted
WLK:2\DKAABAGrandchild of adopted
WLK;2/DKAB
Second child (first natural) of DKA
etc.

Notice that when you have everybody sorted in alphabetical order, as in the illustration above, they are in the natural, expected order for making charts, etc.


More information may be obtained by contacting Paul Stoneburner at <pds@wwco.com>. The Stoneburner reunion area is sponsored by Wizard Workshop and Company.     Members of WWCo are given free telnet accounts, a private area on the Website, and virtually unlimited webspace in exchange for testing our multi-user applications.