TEIC / TEI

The Text Encoding Initiative Guidelines
https://www.tei-c.org
Other
276 stars 88 forks source link

seriesStmt needs better examples, use of biblScope #1870

Open ebeshero opened 5 years ago

ebeshero commented 5 years ago

See TEI listserv conversation, issue raised by Daniel Schwarz: https://listserv.brown.edu/cgi-bin/wa?A2=ind1904&L=TEI-L&P=18938. Here are the issues: 1) The seriesStmt element spec provides an unclear (and likely fictitious) example of the use of biblScope using a decimal notation (1.2) for a volume. Can we provide a less ambiguous example? 2) The prose explanation of seriesStmt doesn't list biblScope in the list of elements it contains. 3) A question: Do we want to do more with biblScope in seriesStmt, than we typically do in a bibl or biblStruct? Since seriesStmt is an (optional) part of the teiHeader that "groups information about the series, if any, to which a publication belongs," if we use it, don't we want to indicate that what we're presenting is part of a larger whole? So, supposing that our example in seriesStmt is indicating Vol. 1, No. 2 of a larger series, don't we also want a means of indicating how many volumes in total there are in the series, and wouldn't that information belong in seriesStmt?

dlschwartz commented 5 years ago

May I also add here that biblScope needs better documentation in general. The guidelines list over 150 elements that may appear inside of biblScope but not one example has anything other than a text node. I'm contemplating the use of title inside of biblScope but don't understand what the TEI would intend that to mean. I can define it in a way that meets a certain need I have, but that wouldn't necessarily mean that I'm using it properly.

ebeshero commented 5 years ago

@dlschwartz Well, those elements are available to <biblScope>'s content model, but we don't have clear precedents for their use. My sense is that<biblScope> is flexible and not limited to things like page and volume numbering--designed to hold standardized information in its attributes about unit and range of any kind to do with documents. What's your idea for using <title> inside it? Let's take a look.

dlschwartz commented 5 years ago

@ebeshero Thanks for the reply. And sorry that this is likely more than you wanted by way of a response. The suggestion I make appears later in the email exchange you reference above. @kshawkin does not think it works. I'm open to that position and continue to think through the issues but I think the whole situation needs to be a lot clearer in the guidelines. The files I am dealing with are a part of multiple series. I need to represent this in order to give proper credit to the work people have done on these series, but also to allow for visualization, search, and browse of groups of files according to the multiple series.

The example in the email that I reiterate below is from our place data but the issues are most easily shown in our person data: http://syriaca.org/persons/index.html. [Please note that the questions I am asking come from recent efforts to standardize things across all of our data, produce the needed schema customizations, and document the various aspects of our project. The headers of the TEI records displayed at that link are part of this process.]

We have a series (title level="s") that is all of our person data and users may want to engage with all of that data. Some records deal with persons who were authors. They constitute a series (title level="m"). Some records deal with persons who were saints. They constitute a series (title level="m"). Some records are for persons who were both authors and saints. They are a part of both title level="m" series. Some records are for people who were neither saints nor authors, they are only a part of the series with title level="s". The seriesStmt in the header is used to facilitate these different approaches to displaying our data.

See this diagram for a rather complicated example. Here we have one TEI document containing data about Ephrem that is actually a part of four different series, two with title level="s" and two with title level="m". Each title level="m" is also a part of a title level="s". I would like to be able to express this in the TEI header for the document. The current practice in the record mentioned doesn't allow me to express editorial credit as broadly as I need to be able to do.

This is what I proposed in the email discussion. It's (mercifully) a simpler example than the diagram above but it gets at the same issues:

The Syriac Gazetteer ... http://syriaca.org/geo Beth Qaṭraye Gazetteer ... http://syriaca.org/bethqatraye The Syriac Gazetteer http://syriaca.org/geo

Multiple seriesStmt elements would indicate that the particular document is a part of multiple series. I also want to indicate in the xml that the Beth Qatraye Gazetteer is a volume of The Syriac Gazetteer. The best option I can find to express that relationship is biblScope. I understand the biblScope in this example to indicate something about the series in the seriesStmt element containing it, namely that it is volume number 1 of something else. My problem is how to be clear in the xml about the thing that this series is a volume of. Most examples of biblScope in the guidelines seem to assume the connection they indicate. We're just supposed to know that titles of level="a" get contained by titles of level="j" and that the biblScope provides some information about that relationship. I'm considering using the title and idno elements inside of biblScope in the above example to indicate the thing of which the second series is a part.

Perhaps that is significantly out of keeping with how the TEI intends title inside of biblScope to function. However, the guidelines aren't clear about how elements inside of biblScope should be used. The original question I posed to the list was about the use of biblScope inside of seriesStmt. I clearly also have questions about the use of any element inside of biblScope.

dlschwartz commented 5 years ago

@ebeshero You asked for what I was trying to accomplish and I'm afraid I may have gone a bit overboard in explaining it. There is a simple question in all of this, and one that gets at my original comment about documentation of biblScope as well as seriesStmt. The TEI allows the following:

  <fileDesc>
     <titleStmt>
        <title>X</title>
     </titleStmt>
     ...
     <seriesStmt>
        <title>Y</title>
        <biblScope unit="volume" n="1">
           <title>?</title>
        </biblScope>
     </seriesStmt>
  </fileDesc>

Title "X" is the title for the current document. Title "Y" is the title of the series to which title "X" belongs. How then would the TEI expect someone use the title inside the biblScope? The question also applies to the other elements allowed inside biblScope.

I propose to use it to assert something about title "Y", namely to provide the title for the entity of which it is volume number 1. As far as I am aware, current documentation does not suggest some different use for the content of a biblScope inside seriesStmt, nor does it give examples of elements inside biblScope in other contexts to use as an analogy for this context. I hope this helps simplify the issues related to my initial comment.

sydb commented 4 years ago

Council VF2F subgroup thinks (1) and (2) are both quite reasonable and should be GO for @ebeshero to go ahead with.

(3), however, is a bit complex. First off, we want to ensure that OP @dlschwartz understands that there is no need to put any elements of any sort inside <biblScope>. Second, we’re a bit worried that at some points (both in the discussion and probably in examples in the Guidelines) folks have used <biblScope> when <citedRange> is the more correct element. Last, we think some of the confusion arises from the fact that <biblScope> (like many other TEI elements) serves two purposes: a. For when you are generating metadata, e.g. writing your TEI header. b. For when you are encoding running prose that includes a bibliographic citation that includes some scoping. It may well make sense to put a <title> inside a <biblScope> for (b), but it probably does not for (a).

The thought of multiple <seriesStmt>s (which does not seem to be part of original ticket) is something we think Council should invite OP @dlschwartz to discuss on a conference call.

lb42 commented 4 years ago

It think the use of <title> inside <biblScope> is unlikely but not implausible. It would suggest that (referring to OP's second example above) vol 1 of series Y has some particular title (other than X) which you wish to record. It might just be "volume one", for example. The more general point to note is that <biblScope> is one of many TEI elements permitting a whole shedload of elements, or none, for simplicity's sake. You may think this shows culpable laziness on the part of the TEI designers, or their willingness to recognize the limits of their abilities. In either case, you can of course constrain your usage of <biblScope>to contain only a subset of what's available.

dlschwartz commented 4 years ago

I am open to the idea that I have peculiar needs for my born digital project and that I might just need to customize and document that customization to meet those needs.

Regarding the use of multiple <seriesStmt>'s, however, I believe there is applicability outside of my born digital project. Here's an example from my field. The Corpus Scriptorum Christianorum Orientalium (CSCO) is a series. It is made up of multiple sub-series: Scriptores Aethiopici, Scriptores Arabici, Scriptores Armeniaci, Scriptores Coptici, Scriptores Iberici, Scriptores Syri. Each volume published is a part of the overall CSCO and receives a CSCO volume number. It is also part of one of the sub-series and receives a separate volume number in that sub-series. This example shows a book that is volume 261 of CSCO and volume 51 of Scriptores Aethiopici.

What I have proposed in the first code block I posted above is, I believe, a reasonable way to model that kind of series and sub-series situation in the TEI. I have a slight concern about how I've used <biblScope> and <title> due to the limited documentation in the Guidelines. That was the impetus for my original question. I would be happy to entertain other ideas for how to model in the header that a file is part of two series, one of which is a sub-series of the other series. I would prefer any customization I make to be as minimal as possible. I am available to discuss. Thank you.

ebeshero commented 4 years ago

Thanks @dlschwartz ! Council discussed this again today in light of your examples; these are super helpful. I'm going to work on an example for the Guidelines uses more than the standard limited things we've been using in examples of biblScope. Also, we're all pretty clear now that we need the capacity for multiple <seriesStmt>s in the TEI header.