due to the fact that I'm quite new to TEI but that I am working in the Correspondence SIG "taskforce" to create a ODD-proposal for a correspDesc-element with Peter Stadler and Sabine Seiffert I'm happy that I think I completely understand what you are suggesting (that just to explain why some of my TEI-wording may a bit odd) - I am perfectly happy with the possibilities to organize the semantics of a descriptive bibliography with nested bibleLists and/or p-like structures; the idea with the <div type="entry"> is very interesting, Schematron rules would be necessary - YES.
I think on the other hand - talking about a born-digital bibliography - that the overall question is, if the TEI wants to facilitate the world with - maybe - an even more structured biblStruct so that the semantics of this element would allow a detailed interchange format for bibliographic references.
For me - in the long run - especially in the wide field of bibliography for entities without ISBNs etc. the aim should be to provide a worldwide database (probably the idea of the World Cat though...) - with very detailed information - using - for example - parts of the detailed metadata that comes with modern digitalization-projects like the one of the Michigan university that was circulated some weeks ago (all printed English books before 1800 digitalized and encoded until 2020). If not a data-base so at least a way of using other project's data. Interchange needs flat but well organized structure.
I'm not sure if:
that's the TEI core business
that's a question of creating new sub-elemtens or just a question of creating a best-practice-model using the already existing elements in combination with international standards like FRBR, RNA etc.
the TEI library SIG is already creating something like this
but I'm sure:
that's an interesting field that could provide data for i.e. historical network-research, especially if we find a way to include as much information as possible about the relationship between bibliographic references
Sorry for keeping commenting - before my scholarly-letters-project I would never have thought that I could be so emotional drawn to questions of bibliography. But in our project one could say that the history of the connectedness of bibliographic entities is the history of the whole subject in nuce.
Am 20.02.2014 um 00:07 schrieb Kevin Hawkins <[log in to unmask]>:
> There were enough typos in that message that I'll send this cleaned-up version ...
> The TEI Guidelines ambiguous regarding the FRBR "group 1" entities (work, expression, manifestation, and item). Descriptive bibliography practice varies greatly, and Julian's sample seems to indicate that this particular bibliography will collate editions of a publication (here I intentionally avoid the FRBR terminology in case it's too specific) in an outline form.
> The elements for encoding bibliographic citations -- bibl, biblStruct, and biblFull -- are all inter-level elements that can be embedded inside chunk-level elements such as <p> or as siblings of them. This gives you some flexibility in your encoding.
> When encoding a descriptive bibliography, the first question in my view is whether the annotations for each entry are just phrases or whether there is ever more than one paragraph in an annotation.
> If there are only phrases, you could use <listBibl>, perhaps with <head>s and nested <listBibl>s for various sections. Each entry would be a <bibl> containing a mix of tags for components of bibliographic references (the children of <bilbStruct>) and elements not typically found in bibliographic references (like <emph>) for the description.
> If there is more than one paragraph in the description of each entry, you're going to have to organize this encoded document not using <listBibl> but instead using one <div type="entry"> for each entry. Inside of each entry would be <bibl> elements that, depending on the structure of the source document, either would be siblings of <p>s containing the description or would be children of them.
> Or, as others have suggested, you use the msDesc module, and its typical encoding structure, in order to gain access to specific elements used in descriptive bibliography.
> In any case, as you can see, I recommend using <bibl> instead of <biblStruct> or <biblFull>. When you are attempting to replicate a source document, with its inconsistencies, rather than creating a born-digital bibliography, it's the only workable option. On the other hand, if you are trying to create a new edition of this bibliography, with a regularized structure, you could try to use <biblStruct> or even <biblFull>. Instead, I recommend that you consider sticking with <bibl> but developing a TEI customization that includes Schematron rules to check the structure of each <bibl> to make sure it has all the expected components. In that way it will end up working like <biblStruct> in enforcing some structural requirements.