Quoting "Myers, John F." <[log in to unmask]>:


> Now that the preliminaries are out of the way, I would like to see  
> discussion regarding a communications medium that allows us to "have  
> our cake and eat it too" - one in which the transcription of data  
> that fosters matching the resource in hand with the record  
> describing it and that simultaneously offers the degree of  
> structured data for coherent organization and effective selection of  
> resources.

John, we should be able to discuss this without reference to any  
particular record formats, because my gut feeling is that any modern  
data format can handle the problem.

 From a technical viewpoint I think that we *should* have both  
transcribed strings and "data" in our descriptions, since they have  
different functions. In fact, I think we should assume that as a goal.  
The only caveat being that we don't want to increase the amount of  
work that the inputter has to do -- we want to reduce the redundant  
keying as much as we can so as to reduce the work load.

I could imagine a system that helps the cataloger by inferring some  
suggestions from the transcribed place, much like when you type  
"Paris" into Google maps. Most of the time you find your place in the  
list; occasionally you need to supply it. As long as we design a data  
element for this, it can be stored like your XML examples, or in some  
other format that achieves the same thing. And it shouldn't matter to  
the cataloger if it is stored as an identifier from the GeoName  
database, a set of coordinates, or some other controlled list. Note  
that it would also have to be possible to leave it blank... for  
fictitious places, lost continents on Venus, or whatever.

To me this would be even simpler than the 008 place of publication,  
and could be made easier to include in our data.

And, BTW, there are other areas in our data where analogous pairings  
could be made, such as your name example. So the question to me is not  
"how can we do this in XML?" but "Why should we do this? What  
functionality does this approach provide?"

kc

>
> In XML, this seems possible via the use of attributes within the  
> tags.  A tag pair could frame a transcribed piece of data, while an  
> associated attribute could encode the corresponding controlled form.  
>  An alternative would be the opposite arrangement (controlled form  
> between tags with transcribed form in the attribute) but I think  
> that would work less well (experts feel free to chime in).
>
> For example (and without engaging the debate over language-neutral  
> MARC equivalent tagging and "natural language" tagging):
>
> We might have various forms for New York City on different  
> resources: New-York; New York; New York, N.Y.; New York, NY; N.Y.,  
> N.Y.; etc.
>
> Using XML tags and attributes we could offer:
> <PlaceOfPublication place="New York, N.Y.">New-York</PlaceOfPublication>
> <PlaceOfPublication place="New York, N.Y.">New York</PlaceOfPublication>
> <PlaceOfPublication place="New York, N.Y.">New York,  
> N.Y.</PlaceOfPublication>
> <PlaceOfPublication place="New York, N.Y.">New York, NY</PlaceOfPublication>
> <PlaceOfPublication place="New York, N.Y.">N.Y., N.Y.</PlaceOfPublication>
>
> The display could drive off the transcribed data within the tag  
> pairs, while the indexing could drive off the data recorded as the  
> attribute value.  If desired, the data could be further transformed  
> in a display to offer the controlled form via a mouse hover.
>
> There are other areas where this treatment might be desirable, owing  
> to the same competing tasks - the Publisher element comes readily to  
> mind.  I am mindful also of elements like Series statement and  
> series access points or Statement of responsibility and  
> corresponding access points.
>
> As to the latter, what if we employed this kind of structure:
> <StatementOfResponsibility>
> by
> <Name controlledname="Public, John Q. (John Quartermain),  
> 1967-">John Q. Public</Name>
> and
> <Name nameid="n987654321">Jane Doe</Name>
> </StatementOfResponsibility>
>
> where the attributes controlledname and nameid offer two options for  
> articulating the controlled form of the name, via a character string  
> or via a control number.  A further attribute addressing whether the  
> name is to be a main or additional access point could also be  
> included.  There are some issues here and I've played fast and  
> loose, since the nameid attribute probably warrants a proper URI  
> structure and the controlledname attribute disregards current  
> conventions for breaking the string into specific subelements, but  
> hopefully the picture is adequate for conversation and elaboration.
>
> Some questions that I lack the expertise to answer:
> Does the use of XML tag attributes seem to be a useful solution?
> Is development of XML tag attributes and associated  
> control/validation structures feasible?
> Are there solutions outside of XML that similarly balance  
> description and access/identification and selection?
>
>
> John F. Myers, Catalog Librarian
> Schaffer Library, Union College
> 807 Union St.
> Schenectady NY 12308
>
> 518-388-6623
> [log in to unmask]
>
>



-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet