On 05-24-2011 16:26, Tom Wetmore wrote: > That doesn't solve the problem of where to store those links. As Cheryl > points out, if you put a link in a person record, you are making the > explicit statement that the linked-to evidence refers to that person. Always? Many programs support some variation of GEDCOM's TYPE tag. No reason a link couldn't have a TYPE subrecord, or a NOTE or .... Even if GEDCOM doesn't officially support it. Maybe some software out there somewhere has tried that. (Please don't take my comments as an enthusiastic endorsement of GEDCOM) -- Wes Groleau There are two types of people in the world … http://Ideas.Lang-Learn.us/barrett?itemid=1157

On 05-25-2011 11:20, Bob LeChevalier wrote: > Their stinky data is merely the old IGI, PRFs and AFs (along with new > stuff that is being constantly added, and which is generally better). Welll, .... One of the collections is images of marriage records where a column on the left is "date license issued" and a column on the right is "date of marriage." Both columns legibly labeled, yet the index consistently reports the left one as date of marriage, even when the right one legibly says "license returned—marriage did not occur." This makes all of the other collections suspect, since most of them offer the indexes only—no verification possible. Another one estimates the date of birth from the date of death, even though the record _and_ the index has the actual date of birth stated. -- Wes Groleau There are two types of people in the world … http://Ideas.Lang-Learn.us/barrett?itemid=1157

On 05-25-2011 06:59, steve wrote: > Anybody know what kind of approach IBM used in programming Watson for > Jeopardy? http://www.sciencemag.org/content/331/6020/999.short http://www.sciencedirect.com/science/article/pii/S0160289611000262 http://www.medscape.com/viewarticle/740079 http://portal.acm.org/citation.cfm?id=1854275 http://en.wikipedia.org/wiki/Watson_%28computer%29 http://www.pbs.org/wgbh/nova/tech/will-watson-win-jeopardy.html http://www.pbs.org/wgbh/nova/tech/smartest-machine-on-earth.html http://www.google.com/search?q=watson+jeopardy -- Wes Groleau There are two types of people in the world … http://Ideas.Lang-Learn.us/barrett?itemid=1157

On Wednesday, May 25, 2011 12:35:49 PM UTC-4, Wes Groleau wrote: > On 05-25-2011 11:20, Bob LeChevalier wrote: > > Their stinky data is merely the old IGI, PRFs and AFs (along with new > > stuff that is being constantly added, and which is generally better). > > Welll, .... > > One of the collections is images of marriage records where a column on > the left is "date license issued" and a column on the right is "date of > marriage." Both columns legibly labeled, yet the index consistently > reports the left one as date of marriage, even when the right one > legibly says "license returned—marriage did not occur." > > This makes all of the other collections suspect, since most of them > offer the indexes only—no verification possible. > > Another one estimates the date of birth from the date of death, even > though the record _and_ the index has the actual date of birth stated. Wes, I'm sure there are many more horror stories like your example. Personally I think that the LDS should scrap all the data in the nFS tree and then pre-load it with personas created from high-quality data taken from vital records, censuses, and other sources from which it is possible to derive quality, sourced, personas. Then don't allow people to blindly upload GEDCOM files. Force them to enter new personas only by hand, as an attempt to filter out the crappiest of the crap. Personally I think it's a lost cause as soon as you allow any user to add any data they like, which basically means that these family trees of all mankind are inherently flawed. The only solution, in my very humble opinion, is to highly restrict that personas that can be added, highly restrict the ability of users to join personas into persons and to join persons into families (it would not be hard for simple algorithms to check basic sanity of such things). The only hope for these trees is quality input and high quality procedures to create persons and families. Personally I much prefer automatic algorithms to do most of this work, as I have seen these algorithms do wonderful things in analogous contexts (I'll be happy to write about these some time). These are controversial views than many disagree with. Tom

On Wednesday, May 25, 2011 12:40:54 PM UTC-4, Wes Groleau wrote: > On 05-24-2011 16:26, Tom Wetmore wrote: > > That doesn't solve the problem of where to store those links. As Cheryl > > points out, if you put a link in a person record, you are making the > > explicit statement that the linked-to evidence refers to that person. > > Always? Many programs support some variation of GEDCOM's TYPE tag. > > No reason a link couldn't have a TYPE subrecord, or a NOTE or .... > Even if GEDCOM doesn't officially support it. > > Maybe some software out there somewhere has tried that. > > (Please don't take my comments as an enthusiastic endorsement > of GEDCOM) > Wes, I wasn't clear enough. The idea I was getting at is this. You have found an item of evidence that you either copy onto your computer as a file or you have as a URL text string. You are pretty sure this evidence refers to a person you are interested in, but you haven't gotten enough info yet to be sure of this or to know exactly what person it refers to. Cheryl made the point that she would keep a link to that file or URL in a person record in her database. My question was directed to the situation where you don't yet have such a person record to hold the link. My preferred approach is to codify that evidence into new persona records and let them be sit in the database while you collect more data. These persona records are indexed and searchable and manipulable and editable as easily as regular person records. I personally find that this simple mechanism solves all problems I have with designing a single system that can seamlessly handle both record-based and person-based genealogy. I just need the software to give the UI to do this. Tom

On Tue, 24 May 2011 23:25:08 -0400, Wes Groleau <[email protected]> wrote: >On 05-24-2011 10:40, J. Hugh Sullivan wrote: >> Unsourced "facts" are often just conclusions - those are pretty scary. > >Maybe not even that. Could be cousin Jane's memory of what her >93-year-old grandfather told her about events that happened before >his father was born. > >(Been there, read that.) > >-- >Wes Groleau Events do tend to get expanded as the years pass. Hugh

Tom Wetmore <[email protected]> wrote: >I concur with Cheryl. Basically all genealogical applications allow their >records to contain links to external items, either files on the local file >system or to URL on the world wide web. > >That doesn't solve the problem of where to store those links. As Cheryl >points out, if you put a link in a person record, you are making the >explicit statement that the linked-to evidence refers to that person. I'd quibble and say that it MIGHT refer to that person. Genealogical software sometimes does have degree of confidence marking on source data, not that I have found it to be at all useful. >The question I am trying to get to in this thread, is how are you going >to handle that evidence, a link to an external something or other in >this case, when you don't yet know what person should link to it? Not in a genealogical application, because until you can link it to a tree, it is not ***genealogical*** data. The only obvious answer is that you organize the data in whatever way is most useful to you, based on whatever commonalities there are in the data that you think are important. If there are no such obvious commonalities, you might just be reduced to assigning an ID number to the evidence, and keeping a list indexed by ID number. lojbab --- Bob LeChevalier - artificial linguist; genealogist [email protected] Lojban language www.lojban.org

Tom Wetmore <[email protected]> wrote: >Bob, > >I've used nFS and read all the API documentation. They have personas >and persons and the user interface allows you to rearrange >personas within persons. It's a two-tier system. I think it's a great >example of how personas can be made to work. I think their data is >pretty stinky, but that has no bearing on the technique in my opinion. Their stinky data is merely the old IGI, PRFs and AFs (along with new stuff that is being constantly added, and which is generally better). A lot of that old IGI stuff was pretty bad, but they had the option to throw everything away, or keep it and find ways to weed out the bad over time. They seem to be trying the latter, and in most cases it is working. They haven't solved the problem of what to do when someone 20 or 50 years ago conflated two people that weren't related, and recorded that conflation as a "fact". At some point they will possibly decide that certain "personas" aren't salvageable, and remove them. (I have the impression that they can and will do so now, but it is on a record by record basis after being presented with documented evidence.) >>> Eventually every genealogist reaches the point when he or she has delved >>> far enougback in time that the solid, firm trail of records has dried up. > >> I don't ever expect to reach that point. But then my genealogy has >> never been confined to direct ancestors. Try to find all the >> descendants of some ancestor born around the time of the revolution. >> Every single one of those descendants is a cousin, and in many >> families that task alone might take a lifetime. > >What I think you are saying is that you never plan to have to cross the chasm >from person-based genealogy to record-based genealogy. On the contrary, I have had to do some of what you call "record-based genealogy" in the 20th century (especially since many records for "living" persons are locked up), and my French work is back in the 17th century and is entirely record-based. But I never found a chasm to be crossed. Lots of little roadblocks, but they only affect one person or one record out of a zillion lines. When you reach the point where there are no records at all, then there is a chasm. I put no priority on trying to cross them. I have ancestors in pre-Revolution Russia and the Ottoman empire, but no significant likelihood in my lifetime to get access to records that I can use (though they might exist). So I've chosen other pursuits, mostly not for my own family, directly. >The whole genealogical application business, in my opinion, caters to people who >believe that. I think it caters to people who are interested in tree structures rather than source data as an organizing principle. Because that is what sells enough to warrant "catering". >Obviously from all I've written about this, I don't believe >that. At my point in genealogical research I have 1000s of records >that I haven't been able to assign to real people yet. As far as I am concerned, they are all assigned to real people. Whether I can attach those people to my tree is another question. I generally don't enter it into my database until I think such attachment is likely if not definite. The source records I either leave where they lie (because I don't give a damn about building a pile of paper that no one including me will ever look at). Or in some cases, build spreadsheets for my extractions therefrom - I don't need a genealogical application for that. I would not imagine that there is would be any general solution for the problems of organizing records of random types that does not involve linking them together in SOME way, or possibly in MANY ways. >Here is my overall genealogical project. I descend from Loyalist Wetmores >who were exiled to Canada at the end of the American Revolutionary >War. By the 1850s many of the children and grandchildren of those >families were returning to the United States where the economy was >stronger. I descend from one of those returning families. My project >is to understand that return migration by finding all the families who >were involved, what their patterns of migration were, where they >ended up, and where all their descendants are living now. This is a >full research project. It is definitely a record-based project. You can >probably understand my needs for effective ways to record all my >evidence so I can access it in many ways to support the process of >making conclusions. But the ways you want to organize such evidence is not genealogical (i.e. not based on family relationships), because you don't KNOW the family relationships. You want a genealogical application to do a fundamentally non-genealogical task. >You might argue that this project is not a genealogical project, but >rather a historical project, so I have no business expecting a >genealogical application to be able to support me. I don't see it that >way. I believe that genealogy is history, and the farther back in >time we go, the more we have to act like historians to make >progress. I want a genealogical application that can support what >I am doing with this project. While one can consider genealogy a branch of history, once you leave the realm of family relationships, it is no longer genealogy. The market isn't there for such a general application covering historical projects as opposed to genealogical ones, in part because every "historical project" is different both in goals and in methods. If that were not the case, then professional historians would be using such applications. (some of them might be, but you'd have to consult such a professional historian). lojbab --- Bob LeChevalier - artificial linguist; genealogist [email protected] Lojban language www.lojban.org

Ian Goddard wrote: > singhals wrote: >> Yes, PAF too is said to handle audio and video, but since a majority of >> my ancestors died before that technology was invented it's even less >> useful. (g) > > And even if it were invented would you really want a recording of a very > drunk ggfather singing out of tune? > (G) EXcellent point! Cheryl

On Wednesday, May 25, 2011 11:20:59 AM UTC-4, Bob LeChevalier wrote: > Tom Wetmore <[email protected]> wrote: > >Bob, > >What I think you are saying is that you never plan to have to cross the chasm > >from person-based genealogy to record-based genealogy. > > On the contrary, I have had to do some of what you call "record-based > genealogy" in the 20th century (especially since many records for > "living" persons are locked up), and my French work is back in the > 17th century and is entirely record-based. But I never found a chasm > to be crossed. Lots of little roadblocks, but they only affect one > person or one record out of a zillion lines. > When you reach the point where there are no records at all, then there > is a chasm. I put no priority on trying to cross them. I have > ancestors in pre-Revolution Russia and the Ottoman empire, but no > significant likelihood in my lifetime to get access to records that I > can use (though they might exist). So I've chosen other pursuits, > mostly not for my own family, directly. > I disagree. The chasm doesn't occur when you run out of records; it occurs when you cannot automatically assign the records you find to a real person yet. > >The whole genealogical application business, in my opinion, caters to > >people who believe that. > > I think it caters to people who are interested in tree structures > rather than source data as an organizing principle. Because that is > what sells enough to warrant "catering". Agreed. > >At my point in genealogical research I have 1000s of records > >that I haven't been able to assign to real people yet. > > As far as I am concerned, they are all assigned to real people. > Whether I can attach those people to my tree is another question. I > generally don't enter it into my database until I think such > attachment is likely if not definite. This is pretty close to my point. When you refer to real people that you can't attach to your tree yet, I would say you're talking about something close to the persona concept. So to my way of thinking you are storing your evidence that can't go directly into persons in your tree, into "holding personas". This is exactly what I am forced to do with my genealogy program, which I guess I am now ashamed to say, I wrote (LifeLines). It is chockerblock full of person records, some the real tree persons, and some my "on hold" evidence personas. Except for the fact that I have no user interface mechanism that allows me to group the personas into persons when I decide how to do it, the system works middling well. I am still forced to merge, which I now believe to be the wrong thing to do with personas. I am working on a new application now where I use drag and drop to manipulate persona records in person records. I'll be using this to experiment with the mutli-tiered person tree approach. If I can't find a user interface metaphor that makes working with the data easy to comprehend I will likely be forced to conclude that the "person tree" concept, as beautiful as I think it is, just isn't practical for the real world. What do you do when you decide that one of those stand alone persons is the same as one already in your tree, or if you decide that two of those stand alone persons are the same person but you still are not ready to add him to your tree. I assume you have to merge data, that is reduce the number of your person records. I don't like this because I think of the persona records as being a codification of evidence. Of course, you don't have to think of it that way. > The source records I either > leave where they lie (because I don't give a damn about building a > pile of paper that no one including me will ever look at). Or in some > cases, build spreadsheets for my extractions therefrom - I don't need > a genealogical application for that. I agree with the former, but think that genealogical applications can do more for you than spreadsheets can. > I would not imagine that there is would be any general solution for > the problems of organizing records of random types that does not > involve linking them together in SOME way, or possibly in MANY ways. Well, for doing genealogy there aren't that many kinds of record types you have to deal with. I'd add persona, event, place to the list, and maybe have a catch-all object for everything else. > >Here is my overall genealogical project. I descend from Loyalist Wetmores > >who were exiled to Canada at the end of the American Revolutionary > >War. By the 1850s many of the children and grandchildren of those > >families were returning to the United States where the economy was > >stronger. I descend from one of those returning families. My project > >is to understand that return migration by finding all the families who > >were involved, what their patterns of migration were, where they > >ended up, and where all their descendants are living now. This is a > >full research project. It is definitely a record-based project. You can > >probably understand my needs for effective ways to record all my > >evidence so I can access it in many ways to support the process of > >making conclusions. > > But the ways you want to organize such evidence is not genealogical > (i.e. not based on family relationships), because you don't KNOW the > family relationships. You want a genealogical application to do a > fundamentally non-genealogical task. But I'm trying to discover the family relationships. It's not complicated how I want to organize the data -- by name, by names of parents if known, by names of spouses and children when known, by important event dates and places when known. These are exactly the same properties that any genealogists wants any of his records organized by. > >You might argue that this project is not a genealogical project, but > >rather a historical project, so I have no business expecting a > >genealogical application to be able to support me. I don't see it that > >way. I believe that genealogy is history, and the farther back in > >time we go, the more we have to act like historians to make > >progress. I want a genealogical application that can support what > >I am doing with this project. > > While one can consider genealogy a branch of history, once you leave > the realm of family relationships, it is no longer genealogy. Technically, I agree. Practically I don't think it matters much. There is no real difference in my mind between researching persons for the purpose of finding their family tree, versus researching them for the purpose of finding out as much about them and their milieux as possible; one just goes a little further, but the process of finding records and reasoning from them is the same. > The market isn't there for such a general application covering > historical projects as opposed to genealogical ones, in part because > every "historical project" is different both in goals and in methods. > If that were not the case, then professional historians would be using > such applications. (some of them might be, but you'd have to consult > such a professional historian). I agree there isn't a market, but not for the reason you give. Historical projects do have different goals, but their methods are identical and easily supported by software. The methods are search for records, extract evidence from the records, reason about the evidence you have found, make and justify conclusions based on the evidence and your reasoning, and publish your findings. There are certainly lots of record types, and searching for them and extracting evidence from them can be pretty unique, but there is nothing different between projects in the overall process or in the fact that a single software application could support them all. I just think the market is small because it is small! Tom Wetmore

On Tue, 24 May 2011 13:26:52 -0700 (PDT), Tom Wetmore <[email protected]> wrote: >The question I am trying to get to in this thread, is how are you going >to handle that evidence, a link to an external something or other in >this case, when you don't yet know what person should link to it? And that's where I find programs like askSam useful. I put such evidence in there, and when I find a person that I think the evidence might link to, it's easy to find the evidence again to check. One example is gravestones. When I visit a cemetery where families I'm interested are buried, I take photos of gravestones of persons I know are related, and those with the same surname that I don't know are related. I then put them into askSam, and also transcribe the inscription (as much of it as I can read), and also put in fields of name, place, and date of death/burial Then if I find another piece of evidence that links that one to someone in my database, it is quite easy to find, and I have a photo of the original gravestone to refer to. The entry form looks something like this: Surname[ First_Names[ Death_Date[ Burial_Date[ Cemetery[ Place[ Prov/County[ Country[ Date_Entered[ ^D Inscription[ ] Picture[ ] Notes[ ] Research[ ] Other_Sources[ ] The fields with a closing square bracket on the following line are multi-line fields, like memo fields in MS Access. By default askSam will search for any word or phrase in any field, though you can restrict searches by using the usual Boolean arguments. It can produce formatted reports based on any combination of fields, like this: Index of monumental inscriptions Burnard, Magdalena M.C. Died 10 Dec 1906 Stellenbosch, Western Cape Burnard, Sibella Margaretha Died 5 Aug 1949 Stellenbosch, Western Cape Coppin, Rebecca Died 22 Jan 1893 Cardinham, Cornwall Hannan, Thomas Died 4 Oct 1890 Girvan, Ayrshire Hayes, Albert Edward Died 7 Sep 1931 Bristol, Hayes, James Died 16 Aug 1943 Axbridge, Somerset Hayes, John Died 10 Jul 1912 Axbridge, Somerset Raw, Donovan Died 14 Nov 1944 Lidgetton, Natal Raw, Leonard Crick Died 29 Mar 1963 Lidgetton, Natal Raw, Lynda Died 19 Nov 1977 Lidgetton, Natal Raw, Margaret Amy Died 24 Nov 1867 Lidgetton, Natal Raw, William George Died 7 Nov 1956 Lidgetton, Natal Riddle, George Died 29 Nov 1878 Cardinham, Cornwall Riddle, Mary Elizabeth Died 20 Sep 1858 Cardinham, Cornwall Riddle, William Died 23 Apr 1848 Cardinham, Cornwall Riddle, William Died 2 Oct 1854 Cardinham, Cornwall Sandercock, Charlotte Died 11 Feb 1880 Cardinham, Cornwall Sandercock, Henry Died 16 Jan 1887 Cardinham, Cornwall Sandercock, William Died 24 Nov 1786 Cardinham, Cornwall Stooke, Edmund Died Oct 1860 Ashton, Devon Tribelhorn, Johannes Jacobus Ferdinand Died 8 Jan 1968 Stellenbosch, Western Cape ______________ Printed 25 May 2011 (the columns don't match when imported into Usenet). Click on any line and you are taken to the original record to check, edit or annotate. In that case, the Riddle ones in Cardinham, Cornwall, were ones I did not know were related, but if I find other evidence that might link to Riddles, it is easy enough to find these ones. In some cases the date of death on a gravestone differs from the date of death on a death certificate or other record. In a lineage-linked genealogy program you usually have one field for recording your conclusion, which is the correct date. But recording the evidence separately means you can change it if fresh evidence turns up giving you reason to revise your conclusion. If I understood your question correctly, I think this answers it. -- Steve Hayes from Tshwane, South Africa Web: http://hayesfam.bravehost.com/stevesig.htm Blog: http://methodius.blogspot.com E-mail - see web page, or parse: shayes at dunelm full stop org full stop uk

On Tue, 24 May 2011 21:49:41 -0400, Brian <[email protected]> wrote: >On Mon, 23 May 2011 18:23:07 +0200, Steve Hayes ><[email protected]> wrote: > >>On Mon, 23 May 2011 04:52:05 -0700 (PDT), Tom Wetmore <[email protected]> >>wrote: >> >>>This thread is an offshoot from the Linux thread that is going off on a number of tangentsl. >>> >>>How should we store evidence in genealogical databases? >> >>Your message was difficult to read as there is something wrong with your >>wordwrap. > >I was able to read it but it went all the way across the screen on a >15.7" laptop. Seems to be fixed now. -- Steve Hayes from Tshwane, South Africa Web: http://hayesfam.bravehost.com/stevesig.htm Blog: http://methodius.blogspot.com E-mail - see web page, or parse: shayes at dunelm full stop org full stop uk

On Mon, 23 May 2011 09:18:57 -0700 (PDT), steve <[email protected]> wrote: >What a genealogy program can or should do depends on what the user is >doing. A program that is great for presenting a family tree may not >be suitable for doing a one name study. How do you record that the >John SMITH who enlisted in the 1st Regiment of Alabama Infantry just >might be the same John SMITH who married Mary JONES over in the >neighboring county; but you're not sure. In a one-name study, couldn't you use a simpler database format, rather than a lineage-linked one? > >Right now I'm sorta trying to do a locality study. I'm completely >clueless as to how I should organize things. I wind up just making >lots of transcriptions and notes and saving them as text files. >Surely there is a better way. That's where askSam could help you: www.asksam.com it is a better way, in that it is easy to search and produce formatted reports from notes and transcriptions. And no, I don't have shares in askSam, I've just been using it for 20 years to keep track of my genealogical research. -- Steve Hayes from Tshwane, South Africa Web: http://hayesfam.bravehost.com/stevesig.htm Blog: http://methodius.blogspot.com E-mail - see web page, or parse: shayes at dunelm full stop org full stop uk

On May 24, 2:19 am, "Peter J. Seymour" <[email protected]> wrote: > On 2011-05-23 23:08, singhals wrote: > > > Tom Wetmore wrote: > >> Cheryl, > > >> I'm not asking what you should store so you can convince others of the > >> accuracy of your work. > > >> I'm asking what do you want to do with the evidence you have gathered > >> about people you MIGHT be interested in, but BEFORE you have figured > >> out who was who. Since the evidence might apply to persons you are > >> interested in, I assume you wouldn't throw it out. Since you don't > >> know who it refers to yet, you can't add it to any person record > >> already in your database. In what form would you want that evidence, > >> and what would you like to be able to do with it? > > > I want a piece of paper, filed where it seemed to me to be a good idea > > to file it. > > > Paper because I find it easier to shuffle paper than electrons -- for > > one thing, I can spread two dozen pieces of paper on the table-top and > > STILL be able to read them, something I find I cannot do with even 4 > > windows open on the monitor. > > > Cheryl > > Absolutely. For all my interest in computers, my real underlying > interest is in information. Computers are only a tool to help dealing > with this. I keep the evidence (and some printouts) in lever arch files > roughly organised on a surname basis. Yes, you can take several pieces > of paper and spread them out in front of you to study and compare them. > That is something you cannot do on a computer. > > The issue of evidence for people you MIGHT be interested in is a good > point. However, my feeling is that throwing it all straight onto a > computer is likely to be unhelpful even if the software can cope. > > Peter Anybody know what kind of approach IBM used in programming Watson for Jeopardy?

On Tuesday, May 24, 2011 11:29:17 PM UTC-4, Steve Hayes wrote: > On Tue, 24 May 2011 13:26:52 -0700 (PDT), Tom Wetmore <[email protected]> > wrote: > > >The question I am trying to get to in this thread, is how are you going > >to handle that evidence, a link to an external something or other in > >this case, when you don't yet know what person should link to it? > > And that's where I find programs like askSam useful. > > I put such evidence in there, and when I find a person that I think the > evidence might link to, it's easy to find the evidence again to check. > > One example is gravestones. > > When I visit a cemetery where families I'm interested are buried, I take > photos of gravestones of persons I know are related, and those with the same > surname that I don't know are related. I then put them into askSam, and also > transcribe the inscription (as much of it as I can read), and also put in > fields of name, place, and date of death/burial > > Then if I find another piece of evidence that links that one to someone in my > database, it is quite easy to find, and I have a photo of the original > gravestone to refer to. > > The entry form looks something like this: > > Surname[ > First_Names[ > Death_Date[ > Burial_Date[ > Cemetery[ > Place[ > Prov/County[ > Country[ > Date_Entered[ ^D > Inscription[ > ] > Picture[ > ] > Notes[ > ] > Research[ > ] > Other_Sources[ > ] > > The fields with a closing square bracket on the following line are multi-line > fields, like memo fields in MS Access. > > By default askSam will search for any word or phrase in any field, though you > can restrict searches by using the usual Boolean arguments. > > It can produce formatted reports based on any combination of fields, like > this: > > Index of monumental inscriptions > > Burnard, Magdalena M.C. Died 10 Dec 1906 Stellenbosch, Western > Cape > Burnard, Sibella Margaretha Died 5 Aug 1949 Stellenbosch, Western > Cape > Coppin, Rebecca Died 22 Jan 1893 Cardinham, Cornwall > Hannan, Thomas Died 4 Oct 1890 Girvan, Ayrshire > Hayes, Albert Edward Died 7 Sep 1931 Bristol, > Hayes, James Died 16 Aug 1943 Axbridge, Somerset > Hayes, John Died 10 Jul 1912 Axbridge, Somerset > Raw, Donovan Died 14 Nov 1944 Lidgetton, Natal > Raw, Leonard Crick Died 29 Mar 1963 Lidgetton, Natal > Raw, Lynda Died 19 Nov 1977 Lidgetton, Natal > Raw, Margaret Amy Died 24 Nov 1867 Lidgetton, Natal > Raw, William George Died 7 Nov 1956 Lidgetton, Natal > Riddle, George Died 29 Nov 1878 Cardinham, Cornwall > Riddle, Mary Elizabeth Died 20 Sep 1858 Cardinham, Cornwall > Riddle, William Died 23 Apr 1848 Cardinham, Cornwall > Riddle, William Died 2 Oct 1854 Cardinham, Cornwall > Sandercock, Charlotte Died 11 Feb 1880 Cardinham, Cornwall > Sandercock, Henry Died 16 Jan 1887 Cardinham, Cornwall > Sandercock, William Died 24 Nov 1786 Cardinham, Cornwall > Stooke, Edmund Died Oct 1860 Ashton, Devon > Tribelhorn, Johannes Jacobus Ferdinand Died 8 Jan 1968 Stellenbosch, > Western Cape > ______________ > Printed 25 May 2011 > > (the columns don't match when imported into Usenet). Click on any line and you > are taken to the original record to check, edit or annotate. > > In that case, the Riddle ones in Cardinham, Cornwall, were ones I did not know > were related, but if I find other evidence that might link to Riddles, it is > easy enough to find these ones. > > In some cases the date of death on a gravestone differs from the date of death > on a death certificate or other record. In a lineage-linked genealogy program > you usually have one field for recording your conclusion, which is the correct > date. But recording the evidence separately means you can change it if fresh > evidence turns up giving you reason to revise your conclusion. > > If I understood your question correctly, I think this answers it. > Steve Hayes from Tshwane, South Africa Steve, Given that a genealogical application does not support evidence it seems like askSam is almost ideal as the "second program" to handle evidence. I would sure like those askSam capabilities built into my genealogy app so I didn't have to move back and forth between two program. I might be naive, but I don't think it would be difficult to add the capabilities to a genealogy app. Every genealogy record can be viewed as a tree of structured name/value pairs (just think of any GEDCOM, XML, or jSON representation of anything you've seen before). If the askSam capability is that of allowing generalized searches over the values of these pairs, then this is a natural add-on feature for a genealogy app. My inclination is to use fairly strictly defined persona records to hold most codified evidence, but using the askSam approach it seems that much of the strictness could be removed, allowing "evidence records" to be just about anything the user desired, just as long as they were structured in such a way that askSam style searching could handle them. Tom

On Tuesday, May 24, 2011 5:27:03 PM UTC-4, Ian Goddard wrote: > Tom Wetmore wrote: > > I concur with Cheryl. Basically all genealogical applications allow their > > records to contain links to external items, either files on the local file > > system or to URL on the world wide web. > > > > That doesn't solve the problem of where to store those links. As Cheryl > > points out, if you put a link in a person record, you are making the > > explicit statement that the linked-to evidence refers to that person. > > > > The question I am trying to get to in this thread, is how are you going > > to handle that evidence, a link to an external something or other in > > this case, when you don't yet know what person should link to it? > > If I were starting from scratch it would go in a blob. The whole > workflow should be evidence > analysis > reconstruction. You aren't > going to have it linked to your reconstruction until you've got to the > end of that process. > > In essence the database would be arranged as follows: > > Evidence: > Those blobs > Hierarchy of source information providing a provenance and indicating > where the material is currently located > > Analysis: > Events > Personae (preserving names as originally written) > Roles linking Personae to Events > Locations linked to events (also preserving names as originally written) > > Reconstruction: > People (providing a standard spelling plus some de-ambiguation as > needed, e.g. John Smith 3rd) > Links to Personae > Relationships - families, political, military, etc > Places (providing a standard spelling but also put in heirarchies, > civil, ecclesiastical, etc, date-bounded where appropriate) > Links to Locations > > -- > Ian Ian, Bingo. You've just described the ideal model. Tom

On 05-24-2011 10:40, J. Hugh Sullivan wrote: > Unsourced "facts" are often just conclusions - those are pretty scary. Maybe not even that. Could be cousin Jane's memory of what her 93-year-old grandfather told her about events that happened before his father was born. (Been there, read that.) -- Wes Groleau There are two types of people in the world … http://Ideas.Lang-Learn.us/barrett?itemid=1157

Tom Wetmore wrote: > I concur with Cheryl. Basically all genealogical applications allow their > records to contain links to external items, either files on the local file > system or to URL on the world wide web. > > That doesn't solve the problem of where to store those links. As Cheryl > points out, if you put a link in a person record, you are making the > explicit statement that the linked-to evidence refers to that person. > > The question I am trying to get to in this thread, is how are you going > to handle that evidence, a link to an external something or other in > this case, when you don't yet know what person should link to it? If I were starting from scratch it would go in a blob. The whole workflow should be evidence > analysis > reconstruction. You aren't going to have it linked to your reconstruction until you've got to the end of that process. In essence the database would be arranged as follows: Evidence: Those blobs Hierarchy of source information providing a provenance and indicating where the material is currently located Analysis: Events Personae (preserving names as originally written) Roles linking Personae to Events Locations linked to events (also preserving names as originally written) Reconstruction: People (providing a standard spelling plus some de-ambiguation as needed, e.g. John Smith 3rd) Links to Personae Relationships - families, political, military, etc Places (providing a standard spelling but also put in heirarchies, civil, ecclesiastical, etc, date-bounded where appropriate) Links to Locations -- Ian The Hotmail address is my spam-bin. Real mail address is iang at austonley org uk

On Tue, 24 May 2011 19:41:44 +0100, Ian Goddard <[email protected]> wrote: >And even if it were invented would you really want a recording of a very >drunk ggfather singing out of tune? Wouldn't have to be that although those would show up. According to my mother, her father had a wonderful singing voice until a car he was working on fell on him in the 30's. It's too bad there were no recordings.

On Mon, 23 May 2011 10:16:44 -0700 (PDT), Tom Wetmore <[email protected]> wrote: >> >> Your message was difficult to read as there is something wrong with your >> wordwrap. > >If I could fix that I would. I'm using the google interface with safari on a mac. Suggestions? > >Tom I don't know. I use Agent and it has a setting for Word Wrap. Perhaps what you use does also.