From: <bh...@me...> - 2002-06-22 10:34:43
|
Hi! I posed a question about the following on the `open discussion' forum on sourceforge, and was encouraged to present them here. 1) Does or will gramps support an `evidence/conclusion'-model for genealogical research? By an evidence/conclusion-model I mean the ability to clearly distinguish between the _facts_ (from sources) I enter (i.e., the evidence) and the _conclusion_ I draw from the facts. For instance, in one church book record it says that a person _Anders Foer_, son of Hans Andersen and Malena Christensdatter, was christened November 22 1778. In another record, it says that a man _Anders Hansen Nerfore_ married a woman Martha Hansdatter on September 25 1804. These are pieces of _evidence_. Are `Anders Foer' and `Anders Hansen Nerfore' the same person? Based on other information in these (and other) sources, I believe they are. This is my _conclusion_. After having drawn the conclusion that two or more persons referenced in sources, are the same, I want to treat them as one person for `all practical purposes' when using the data base. But if I simply store the informations as events (christening and marriage) on a person (with alternative names), I lose the information that this is based on a conclusion. If I later on find contradicting evidence, I have no way to backtrace my reasoning and correct the error. Similarly, it makes it very hard for other people to trust my data, because they can't see what is evidence and what is my conlusions. 2) Does or will gramps support a `source oriented' way of entering data? This can be somewhat related to the above, but it doesn't have to. By source oriented, I simply mean that when I enter the information from a source excerpt, I don't want to have to manually spread that information to a lot of persons. For instance, in the christening record above, it also says that Mons Jelstad, Bi=F8rn Walnum Wigdell, Hans Christophersen F=F8lvigen and Sara Iversdatter Foer were godfathers and godmothers for Anders. So the record contains information about the name and residence of seven people. It's tiresome, at best, to have to manuall update (and often create) all those person records. Even further, I would like to be able to extract or collect the information from a source excerpt later on. (This becomes very useful and important with an evidence/conclusion-model.) I am searching for a genealogy program that have these features (or at least plans to have them), and I would be willing to put some work in it. I realise that it might not be easy to implement these features, but I think they would bring the development of genealogical software a great step forward. I'm very interested in finding and developing new ways that a computer can _help_ me in my research. Today, most genealogical software is not much more than a specialised data base application. --=20 Regards, Bj=F8rn-Helge Mevik |
From: Mark L. <ma...@fo...> - 2002-06-22 12:25:53
|
I've run into some strange behavior with the latest CVS version of Gramps that looks to me like a bug in the Python 1.5.2 interpreter that ships with RH 7.2. Has anybody else seen this? The following snippet from AutoComp.py lines 60-62 causes an ugly stack-trace and corresponding bad behavior whenever I try to open the EditPerson dialog: # This doesn't work. cnv = string.lower self.nlist = [] self.nlist = map((lambda n: (cnv(n),n)),plist) The above code throws a name error on cnv within the lambda. But if I replace the cnv(n) call with string.lower(n) inside the lambda, everything works as expected: # This works cnv = string.lower self.nlist = [] self.nlist = map((lambda n: (string.lower(n),n)),plist) Am I missing something here?? I'd hate to have to leave the string.lower in there because it's slower. -- Mark Lewis |
From: Don A. <dal...@us...> - 2002-06-22 13:02:11
|
Mark, I encountered the same thing last night. I've been looking at it but have not come up with a solution other than the one you came up with. Don On Sat, 2002-06-22 at 06:29, Mark Lewis wrote: > I've run into some strange behavior with the latest CVS version of > Gramps that looks to me like a bug in the Python 1.5.2 interpreter that > ships with RH 7.2. Has anybody else seen this? > > The following snippet from AutoComp.py lines 60-62 causes an ugly > stack-trace and corresponding bad behavior whenever I try to open the > EditPerson dialog: > > # This doesn't work. > cnv = string.lower > self.nlist = [] > self.nlist = map((lambda n: (cnv(n),n)),plist) > > The above code throws a name error on cnv within the lambda. But if I > replace the cnv(n) call with string.lower(n) inside the lambda, > everything works as expected: > > # This works > cnv = string.lower > self.nlist = [] > self.nlist = map((lambda n: (string.lower(n),n)),plist) > > Am I missing something here?? I'd hate to have to leave the > string.lower in there because it's slower. > > -- Mark Lewis > > > > ------------------------------------------------------- > Sponsored by: > ThinkGeek at http://www.ThinkGeek.com/ > _______________________________________________ > Gramps-devel mailing list > Gra...@li... > https://lists.sourceforge.net/lists/listinfo/gramps-devel -- Don Allingham dal...@us... http://gramps.sourceforge.net |
From: Don A. <dal...@us...> - 2002-06-22 20:14:52
|
I like both of these concepts. The first one, in particular, maps to the original concept of GRAMPS being an analysis system, not just a database. I like the idea of being able to link two or more people together as potentially the same person, and treat them as a single person without doing a full merge. I also like the idea of being able to break the person back into the original parts if you discover that you made a mistake. I would need to think on how to make this fit nicely into the current database. Right now we have a Person object and a Family object. Somehow a Person object would need to be able to consist of multiple Person objects. I'm not clear on the second one. Are we really talking about "Events" and not "Sources"? A christening is an event, and I can see multiple people linked to the same event. Is this what is meant? Don On Sat, 2002-06-22 at 04:47, Bj=F8rn-Helge Mevik wrote: > Hi! >=20 > I posed a question about the following on the `open discussion' forum > on sourceforge, and was encouraged to present them here. >=20 > 1) Does or will gramps support an `evidence/conclusion'-model for > genealogical research? By an evidence/conclusion-model I mean the > ability to clearly distinguish between the _facts_ (from sources) I > enter (i.e., the evidence) and the _conclusion_ I draw from the > facts. >=20 > For instance, in one church book record it says that a person > _Anders Foer_, son of Hans Andersen and Malena Christensdatter, was > christened November 22 1778. In another record, it says that a man > _Anders Hansen Nerfore_ married a woman Martha Hansdatter on > September 25 1804. These are pieces of _evidence_. Are `Anders > Foer' and `Anders Hansen Nerfore' the same person? Based on other > information in these (and other) sources, I believe they are. This > is my _conclusion_. >=20 > After having drawn the conclusion that two or more persons > referenced in sources, are the same, I want to treat them as one > person for `all practical purposes' when using the data base. >=20 > But if I simply store the informations as events (christening and > marriage) on a person (with alternative names), I lose the > information that this is based on a conclusion. If I later on find > contradicting evidence, I have no way to backtrace my reasoning and > correct the error. Similarly, it makes it very hard for other > people to trust my data, because they can't see what is evidence > and what is my conlusions. >=20 > 2) Does or will gramps support a `source oriented' way of entering > data? This can be somewhat related to the above, but it doesn't > have to. By source oriented, I simply mean that when I enter the > information from a source excerpt, I don't want to have to manually > spread that information to a lot of persons. >=20 > For instance, in the christening record above, it also says that > Mons Jelstad, Bi=F8rn Walnum Wigdell, Hans Christophersen F=F8lvigen > and Sara Iversdatter Foer were godfathers and godmothers for > Anders. So the record contains information about the name and > residence of seven people. It's tiresome, at best, to have to > manuall update (and often create) all those person records. >=20 > Even further, I would like to be able to extract or collect the > information from a source excerpt later on. (This becomes very > useful and important with an evidence/conclusion-model.) >=20 > I am searching for a genealogy program that have these features (or at > least plans to have them), and I would be willing to put some work in > it. I realise that it might not be easy to implement these features, > but I think they would bring the development of genealogical software a > great step forward. I'm very interested in finding and developing new > ways that a computer can _help_ me in my research. Today, most > genealogical software is not much more than a specialised data base > application. >=20 --=20 Don Allingham dal...@us... http://gramps.sourceforge.net |
From: <bh...@me...> - 2002-06-22 22:02:54
|
Don Allingham <dal...@us...> writes: > I would need to think on how to make this fit nicely into the current > database. Right now we have a Person object and a Family object. Somehow > a Person object would need to be able to consist of multiple Person > objects. I like to think of it in terms of `atomic' and `composite' person objects. One thing that has to be decided is which `compositions' should be allowed. Can a composite person consist of only atomic persons or of both atomic and composite persons? Should it only be allowed to combine two persons (atomic or composite) at a time (so we get a `tree' of compositions)? (This would be the most interesting if one wanted to assign probabilities to the compositions; it makes it much clearer what the probabilities mean.) Should a person be allowed to be part of more than one composite person? > I'm not clear on the second one. Are we really talking about "Events" > and not "Sources"? A christening is an event, and I can see multiple > people linked to the same event. Is this what is meant? I think there is a difference. An event can be described by several sources, with varying information. For instance, in Norway, there are usually two sets of church books, one written by the vicar and one by the parish clerk. They often contain slightly different information. Also, a source record sometimes contain information about more than one event; for instance some marriage records in Norwegian church books give the dates for birth, christening and confirmation. Finally, sources not only describe events but attributes, such as names of people or places. I'm sorry if I'm getting too philosophical, but: In my opinion the source excerpt (is there a better word?) is perhaps the most important entity in a genealogical research process. You start with the source excerpts. It is the source excerpts you can assign a confidence to. It is from the information in the sources that you can draw conclusions about which persons (or events, or places) are the same. Therefore I'd like to keep the source excerpts as `intact' as possible, and at least be able to enter it into the data base as one. As I see it, the genealogical research process consists of three more or less separate `parts': - evidence, i.e., sources - conclusions - presentation It is customary to separate the last from the two first in genealogy software (i.e., different reports from the same data base), and I believe it would be a Good Thing to separate the two first as well. --=20 Bj=F8rn-Helge Mevik |