Quantcast

Media and attributes in data model for Gramps 3.0

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Media and attributes in data model for Gramps 3.0

Tim Lyons
Administrator
In Gramps 3.0, can source references contain media, and are  
attributes of references (especially attributes of event refs) output  
in the Descendent report.

(I think I know the answer to the second question, because I have  
downloaded the source, and those attributes are not printed. However  
this is o problem to me because I have modified my own report module  
to do so - once I discovered that Python hates tabs in the source  
files!)

The reason I ask is because I have been trying to do some data  
analysis of how sources should be recorded in Gramps, referring to  
the Silicon Valley PAF Users group "Family History Documentation  
Guidelines", and to a lesser extent "Cite your Sources' by Richard S  
Lackey.

It's obviously possible to use the data model in a variety of ways,  
but I think that the optimum use for census data is:

The event ref should contain information about the person at the  
particular census address, for example the age and occupation as  
reported in the census (hence my desire to print these things out).

The event should be the census at a particular address. Hence address  
and the transcription of the census entry should be here.

The source ref should contain information about the particular  
section of the census document, and hence should contain the page  
reference, and the reference to the scan or photocopy of the census  
page (hence my desire to have the media as a possible property of a  
source ref).

The source should be the census as a whole (e.g. the UK 1901 Census).

I appreciate that an alternative is to have the census page or census  
at the particular address as the concept that is mapped to the  
source. However, the "Family History Documentation Guidelines" in  
particular recommends that the source should be the whole census, and  
the source ref (called a citation) should also be as I suggest.

Regards
Tim Lyons.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Media and attributes in data model for Gramps 3.0

Benny Malengier
We try to avoid duplicate information. That means an error in data is much harder to correct, so is a good thing.

Hence, it is logical a reference only indicates where in a source data can be found, not contain the data itself.

You say correctly:
>  ...hence should contain the page
> reference, and the reference to the scan or photocopy of the census
> page"

In my opinion you are wrong in then extrapolating to:

> hence my desire to have the media as a possible property of a
> source ref

because if you do that, you are adding data to the reference.

I understand it is a difficult balance, and that using shared objects can circumvent the problem (you add everything to source, but share the relavant bits also on the reference, so data is in one place, but you share it in many). Sharing is however something that is also hard to do consistently.

So in all, I'm not convinced the benefits of what you ask, would not be offset by the problems users might encounter like this.
A better solution in my opinion is to 'see' in the reference the exact portions of the source that are reference, by using the reference data (page, section, ...).

As mentioned in another mail, redesigning source and place should probably be a priority after 3.0

Benny

2007/11/21, Tim Lyons <[hidden email]>:
In Gramps 3.0, can source references contain media, and are
attributes of references (especially attributes of event refs) output
in the Descendent report.

(I think I know the answer to the second question, because I have
downloaded the source, and those attributes are not printed. However
this is o problem to me because I have modified my own report module
to do so - once I discovered that Python hates tabs in the source
files!)

The reason I ask is because I have been trying to do some data
analysis of how sources should be recorded in Gramps, referring to
the Silicon Valley PAF Users group "Family History Documentation
Guidelines", and to a lesser extent "Cite your Sources' by Richard S
Lackey.

It's obviously possible to use the data model in a variety of ways,
but I think that the optimum use for census data is:

The event ref should contain information about the person at the
particular census address, for example the age and occupation as
reported in the census (hence my desire to print these things out).

The event should be the census at a particular address. Hence address
and the transcription of the census entry should be here.

The source ref should contain information about the particular
section of the census document, and hence should contain the page
reference, and the reference to the scan or photocopy of the census
page (hence my desire to have the media as a possible property of a
source ref).

The source should be the census as a whole (e.g. the UK 1901 Census).

I appreciate that an alternative is to have the census page or census
at the particular address as the concept that is mapped to the
source. However, the "Family History Documentation Guidelines" in
particular recommends that the source should be the whole census, and
the source ref (called a citation) should also be as I suggest.

Regards
Tim Lyons.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Media and attributes in data model for Gramps 3.0

george legge
Tim and Bennie, I agree with you both.

I am accustomed to basing my citations (= references?) to sources on
the work of Elizabeth Shown Mills, who wrote (1997) "Evidence -
Citation for the Family Historian". That work was based on the earlier
work (1981) "Cite your Sources" by the late Richard Lackey. Both
authors are recognized authorities. Recently, Mills has published a
much larger book, "Evidence Explained", with discussion and examples,
including a greater coverage of web and digital sources.

I believe that a census for an entire country would be a source,
assuming that it represents a recognized body of data available at one
repository, however the collection may have been organized. I would
certainly not wish to treat a single page or a household entry as a
source - such a treatment would leave me with thousands of sources,
rather that the hundreds that I maintain and use.

Yes, the citation (or reference) is specific to an event (in this case
the census record for a household) and so must include the page number
etc where the entry may be found. However, one often has to interpret
what appears in the census; under those circumstances, it may be
desirable to include (in the citation) the exact wording from the
census. It should be possible to place that in a memo belonging to the
citation. There then needs to be a straightforward option to include
or omit such memos when printing a report.

Full citations should be available on all data in any report (text or
graphic), since the data are valueless without citation of their
sources.

Regards, George Legge

[hidden email]

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Media and attributes in data model for Gramps 3.0

Tim Lyons
Administrator
George & Benny,
 
Thank you very much for your thoughtful replies.
 
I take it that, as a matter of fact, source refs can't have media attachments in 3.0 (you don't actually say that).
 
I am not sure why it is logical that a source ref (citation) should only contain a reference to the source, not any data. Surely it depends on what the data is, and whether it belongs to the event, the source or the source ref.
 
My understanding of the data structure is that a source ref is like a structured repeating subrecord within the event record (I know it may not be implemented like that, but my head is in hierarchical/network databases at the moment unfortunately). So I am not sure why this would cause particular problems of shared data. If my understanding is not correct (taking account of the fact that may be done as a relational implementation) then that may well explain why there is a difficulty.
 
At least we all seem agreed that a census for an entire country should be a source rather than treating a single page or a household entry as a source.
 
Now suppose I have an event which is the census at a particular address, and (it being the UK 1881 Census) I have two sources, the transcription, and the census records themselves.
 
As George says, it may be desirable to include the exact wording of the census in a memo belonging to the source ref (citation). In my case, the exact wording for the two sources is different (perhaps there was an error in the transcription). Therefore, it is not appropriate to place the wording in the event itself. Also, it is not reasonable to place the exact wording in the source (record) because this wording only applies to one event (hence there is no need to think about whether the wording is shared or not).
 
Of course everything that I have said about 'exact wording' applies to the media copy of the census page, except that I think you can attach a memo to a source ref, but you can't attach a media.
 
Regards,
Tim Lyons

 

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
Loading...