[-empyre-] Engineering the University : Week 03 : Bettivia and Flanders

Julia Flanders j.flanders at neu.edu
Mon Mar 23 03:01:06 AEDT 2015 

Dear Rhiannon,

This is a great question to wrap up with and one that interests me a great deal. I'll start with the familiar tag line, "The wonderful thing about standards is that there are so many to choose from..." This is supposed to be funny because it illustrates the futility of attempting to develop standards, but I think it actually misses part of the point of standards from a digital humanities perspective, and it's a little too cynical about the role that the plurality of standards plays in our data culture.

I'll put in a plug here for a piece I wrote some time ago about collaboration and the role of standards (looking in particular at the TEI):

“Collaboration and Dissent: Challenges of Collaborative Standards for Digital Humanities.” In Collaborative Research in Digital Humanities, ed. Marilyn Deegan and Willard McCarty. Ashgate Publishing, 2012.

Anyone who has ever chaired a meeting at which the goal was to agree on a shared practice of any kind (e.g. what Dublin Core metadata fields will we use in this project? How will my partner and I organize our shared bookshelves? etc.) has participated in a standards-building effort on a very small scale and has experienced its challenges, viz:

--you're trying to agree on a consistent approach that everyone will adopt
--you're trying to reconcile differing motivations and beliefs any one of which (if followed individually) has its own local authenticity and reasonableness, but which if all followed at once would result in inconsistency and hence chaos
--you are all willing to accept the idea that any shared solution is better than inconsistency, but putting this belief into practice means giving up some of the virtues of the individual systems
--the desiderata shaping the local systems don't necessarily translate well across systems

Part of the point here is that standards (in a general sense) quite often operate locally and can do so quite effectively. It is not necessary or desirable for all standards to have universal scope. So in designing a good standard, the first questions should really be "What is the scope of this standard?" "Who are the logical contributors to its development?" "What will be the benefits of its adoption and to whom?" and finally "What will be the costs of its adoption?"

In digital humanities, I think there are three types of standards (I just made up that number):

1. Formal standards that arise from technical necessity and are really non-negotiable (and often invisible): things like IP addresses, file transfer protocols, what order our devices read data packets in, that sort of thing. Although there are people for whom these standards have strategic consequences and a visible politics, those people are really in an entirely different profession. From the DH perspective, we should just be glad that someone else attended those meetings and thought things through for the rest of us, and we hope they did a good job because nothing's going to change them now.

2. Formal standards (or the functional equivalent) that represent widespread, trans-community agreements about things humanists could reasonably care about: things like MARC, ISO codes for representing gender, even XML itself). These standards have perceptible effects on digital humanists' research activities, the shaping of their data, the future of their projects. 

3. "Community standards" (the term I've adopted) that represent more local, voluntary agreements within specific communities concerning things that affect their work: things like discipline-specific controlled vocabularies (e.g. the Getty Art and Architecture Thesaurus), best practices (such as the use of the Data Management Planning Tool) and data representation systems (such as a discipline-specific customization of the TEI, e.g. DALF or EpiDoc). These standards directly and materially affect those who work in the affected domains by creating an expectation (though not a requirement) of conformity.

One role I see standards playing in digital humanities as the field develops is the articulation and formalization of practices, beliefs, and intellectual systems that were formerly either explicit but unformalized, or perhaps even purely implicit. These formalizations help strengthen our ability to use information in systematic ways. 

Another role standards play is that the process of their creation results in greater clarity about where the points of agreement and disagreement are. Within the TEI community, for instance, I think the debates about how that community standard should be used have made it clear that there's a distinction between what I would call "documentary" and "critical" approaches to text markup that are somewhat analogous to (though they don't exactly replicate) the distinction between documentary editing and critical editing. 

Standards can feel coercive, particularly if one wasn't involved in their making, and in this connection I think the TEI is a very interesting example, not only because it does genuinely arise from the community that uses it (with clear and effective mechanisms for involving community members in all aspects of the process), but also because as a standard it includes explicit mechanisms for its own modification by individual users. Users have to decide for themselves how strongly they value the somewhat opposed advantages of conformity and individual expressiveness. A researcher or a  project or a disciplinary group might decide that it's more important to use the TEI in ways that optimize the local research effort, or they might decide that it's more important to optimize for data-sharing with the larger TEI community. The customization mechanism the TEI offers accommodates both approaches. In the paper I reference above, I describe both the specifics of the mechanism and also its significance for digital humanities collaboration, which I think is significant. Essentially this approach to standards-building provides flexibility within the standard so that it can be used at different levels of standardization.

This approach clearly can't work for the first two categories of standards I list above, which require greater levels of universality to really operate at all. But within the third category of "community standards" I think this is a useful approach. It's one that requires direct participation from the community, so I guess the final dimension of my response to Rhiannon's question is to say that the existence of standards in the DH world requires that DH scholars contribute actively, responsibly, and expertly in their creation and usage.

I'd be interested in hearing more from the group about the standards that affect your work--I'm aware that I may be coming at this from a very specific perspective!

Best wishes, Julia

> On Mar 21, 2015, at 7:53 PM, Bettivia, Rhiannon Stephanie <rbettivi at illinois.edu> wrote:
> 
> ----------empyre- soft-skinned space----------------------
> Thank you, Julia, for all the time and consideration you have put into
> responding to all the questions that have arise around this topic.  As we
> get ready to close the week out, I will pose a final question inspired by
> some of the recent discussion and your TEI example.
> 
> You mentioned in your example that you use ISO codes to denote gender in
> TEI.  TEI itself, along with all markup languages, metadata schema, etc
> are standards unto themselves.  Standards come in many varieties, some
> very formal like those approved or commissioned by ISO while others are
> less formal, yet sometimes more pervasive.  I wonder what role standards
> play in the evolving field of DH and, to try to bring this discussion
> round to the general topic of the month, I also wonder what role they play
> in the kind of labor we do in academic spaces.  What kind of work do
> standards do in humanist spaces and what kind of work do we, as scholars
> very generally, do with or around standards in our digital work?
> 
> Thank you again for your participation, Julia.  It has been truly
> wonderful.
> 
> Rhiannon Bettivia
> Doctoral Candidate
> Graduate School of Library and Information Science
> University of Illinois, Urbana-Champaign
> 
> 
> 
> 
> On 3/21/15, 2:49 PM, "Hamilton, Kevin" <kham at illinois.edu> wrote:
> 
>> ----------empyre- soft-skinned space----------------------
>> Fantastic! Thanks, Julia for the generous detail. This is indeed the sort
>> of discussion I was thinking about in my original mention. Fascinating.
>> 
>> Kevin
>> 
>> On 3/21/15 2:43 PM, "Julia Flanders" <j.flanders at neu.edu> wrote:
>> 
>>> ----------empyre- soft-skinned space----------------------
>>> There are several ways gender might register in the markup. I think what
>>> Kevin was probably thinking of is part of a larger mechanism for
>>> representing a variety of personal detail about individuals who are
>>> mentioned in texts. I'm not sure how familiar you are with TEI/XML
>>> markup; if the example below isn't clear, please let me know and I can
>>> elaborate.
>>> 
>>> The transcription of the source material might look like this (a
>>> paragraph of prose with references to individual people):
>>> 
>>> <p>It was on that first day of spring in 2015 that <name
>>> ref="#julia">Julia</name> found herself driving through the back roads of
>>> Smithfield when abruptly she found her way blocked by a white van, buried
>>> up to the axles in a snowbank.</p>
>>> 
>>> Then elsewhere in the TEI file, there's a set of editorial data about the
>>> people named in the text:
>>> 
>>> <listPerson>
>>>  <person xml:id="julia" sex="2">
>>> 	<persName>Julia Flanders</persName>
>>> 	<birth when="1965-02-21">
>>> 		<placeName>New York City</placeName>
>>> 	</birth>
>>>  </person>
>>> </listPerson>
>>> 
>>> So gender in this particular case is represented through a numeric code
>>> (the TEI happens to use the ISO standard codes for representing gender)
>>> but it could just as easily be represented using different terminology
>>> and encoding mechanisms, e.g.:
>>> 
>>> <listPerson>
>>>  <person xml:id="julia" gender="female">
>>> 	<persName>Julia Flanders</persName>
>>> 	<birth when="1965-02-21">
>>> 		<placeName>New York City</placeName>
>>> 	</birth>
>>> 	<gender>Female</gender>
>>>  </person>
>>> </listPerson>
>>> 
>>> If one were interested in representing discussions of gender, rather than
>>> the gender of specific named entities, one could do that using a
>>> mechanism such as this one:
>>> 
>>> <p ana="#gender">[sample paragraph discussing gender; I'm too lazy to
>>> come up with a good example.]</p>
>>> 
>>> <interpGrp>
>>> 	<interp xml:id="gender">Keyword representing discussions of
>>> gender...</interp>
>>> </interpGrp>
>>> 
>>> In the example above, the <interp> element serves to define and anchor a
>>> keyword which can then be applied to the text by using an attribute (@ana
>>> in this case) to reference it.
>>> 
>>> Let me know if you have questions--
>>> 
>>> best, Julia
>>> 
>>>> On Mar 20, 2015, at 7:22 PM, B. Bogart <ben at ekran.org> wrote:
>>>> 
>>>> ----------empyre- soft-skinned space----------------------
>>>> I'd like to here more about this issue of representing gender. We did
>>>> touch on it last week, but it felt like it never really got fleshed
>>>> out.
>>>> How is gender represented in standardized markup language?
>>>> 
>>>> On 15-03-18 07:33 PM, Hamilton, Kevin wrote:
>>>>> Julia wouldn't remember me I'm sure, but I had the pleasure of taking
>>>>> a
>>>>> TEI workshop from her once, in my first (and last) introduction to the
>>>>> complexities of standardized markup language for scholarly texts. I
>>>>> remember hearing her and the other workshop leader talking about the
>>>>> question of gender, for example, and what fields might exist in a
>>>>> standardized markup language for indicating the gender of a character
>>>>> in a
>>>>> narrative, etc. What a rich opening for newbies like me, to see where
>>>>> matters of gender get literally encoded in machine-readable language,
>>>>> but
>>>>> also debated through those looking to set standards, arbitrate them,
>>>>> comment on them, etc. (Sorry for the broad brush Julia, but something
>>>>> stuck in there for me, even if it wasn't what was really going on : )
>>>> _______________________________________________
>>>> empyre forum
>>>> empyre at lists.artdesign.unsw.edu.au
>>>> http://empyre.library.cornell.edu
>>> 
>>> _______________________________________________
>>> empyre forum
>>> empyre at lists.artdesign.unsw.edu.au
>>> http://empyre.library.cornell.edu
>> 
>> _______________________________________________
>> empyre forum
>> empyre at lists.artdesign.unsw.edu.au
>> http://empyre.library.cornell.edu
> 
> _______________________________________________
> empyre forum
> empyre at lists.artdesign.unsw.edu.au
> http://empyre.library.cornell.edu

More information about the empyre mailing list