[Sources-Citations] Recording your work

Richard Tallent richard at tallent.us
Fri Aug 21 01:16:27 CDT 2015

> On Aug 19, 2015, at 7:33 PM, Rob Hoare <fhiso at robhoare.com> wrote:
> Richard Smith wrote: 
>> 7.  There should be a link from the derived source to the original to 
>> provide a form of layering. ... 
>> Standard derivation types should include:
>>   (a) Media transfers.  ...
>>   (b) Transcription.  ...
>>   (c) Translation.  ...
>>   (d) Indexing.  ...
> As Calum mentioned with the different types of translation, and extracts
> and abstracts, there's perhaps another category ("interpretation"?)
> between b and c.

Probably so. There are four solutions to this I can think of:

1. Come up with a single category that includes everything within the spectrum of B to C, inclusive. Loses too much granularity, IMHO, but it's an option.

2. Come up with one or more categories to squeeze "between" them, but some of the same characteristics you mentioned (corrections based on outside knowledge, adding contextual data such as a missing year, etc.)  could be equally applied while indexing.

3. Allow derivation types to be multi-valued -- more like "flags" or "tags" that describe aspects of the derivation separately from one another.

4. Loosen my rigid, purist definition of “transcription” to allow for certain classes of transcriber edits that are common, such as expanding ditto marks or changing “Wm” to “William”.

> It might sound like these are conclusions, but most transcripts will
> have bits added, filled in or interpreted by the transcriber (or, they
> need to, to be able to be indexed, for example by expanding ditto
> marks). If these are clearly separate from the verbatim transcript
> (another layer) you know who/what to blame when it's been interpreted
> wrong.

I’m not sure many transcribers will feel strongly about creating both an “exact” transcript and a “cleaned up” version as separate layers, but allowing for that possibility is an interesting idea.

>> 9.  There should be an entity or entities to represent publishers and 
>> authors, with attributes for their name and (at least for publishers) 
>> the city they're based in.
> Using OCLC Worldcat (for books) and VIAF.org references (name authority
> for authors and publishers from a consortium of most national libraries)
> would be one way to do some of this. 
> Example author: http://viaf.org/viaf/61911533
> Example book: http://www.worldcat.org/oclc/32130150 (see the data under
> "linked data", includes things like "place of publication", where
> known).

See my note on the last email regarding external source layers, I think this plays very nicely with that concept.

> I like the direction these layers are heading: hopefully the end result
> is a standard that the (online) record providers can provide along with
> each data layer, so that sources and citation become close to automatic.
> So the information needed should be something that the data provider
> already knows or can generate, wherever possible.

I’m definitely on the same page here.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://fhiso.org/pipermail/sources-citations_fhiso.org/attachments/20150821/fce287a7/attachment.html>

More information about the Sources-Citations mailing list