###Scholarship in Software, Software as Scholarship: From Genesis to Peer Review, Universität Bern, 29-30 January 2015
Live notes, so an incomplete, partial record of what actually happened.
Tags: softwarebern
Programme: https://infoclio.ch/en/node/137184
My asides in []
##Thursday
###Willard McCarty, A Matter of Prepositions: Software in Scholarship and Scholarship in Software?
Herman Goldstine and John van Neumann, Planning and Coding Problems for an Electronic Computing Instrument (1947) - first theory of programming.
Bacon and Swift on speaking with things. Industrial revolution, loom breaking as interaction with automated machinery. Babbage's Difference Engine.
Software like systems already in scholarship - music scores, commentaries on edited texts (instructions on how to deal with scholarship in scholarship).
How do we tell that someone is a scholar? That there work is scholarship? By trusting it, by using it, by testing what we can of it, by building up trust from the evidence we have - but not by knowing every step, we cannot know that.
Obvious point: what is built needs to be used to be made scholarly.
Q: is all software really inevitably a black box?
A: you can publish the code. But how do we know code, we know it in action not in the components in the box.
Q: don't agree that impossible to know computational systems - sets up sublime. And we know these systems have histories. We can assemble and disassemble them
A: dissassembling is a situation in which you can study it but not how it operates in the real world.
[talk very much couched in history of science methodology]
###Panel 1: Assessment and Process
--
Eugene Lyman (Independent Scholar) - Scholarly Software and the Enhancement of Critical Scrutiny
Scholarship connected to expertise, internalised definition. You could look at code for hours and not see scholarship.
Scholarly edition. Testing the quality of the text in an edition the work of a serious reader who wants to reuse the text. Digital scholarly editions enable deeper scrutiny of these texts.
Piers Plowman electronic archive. Problems:
- we have poor visual memories, so comparing transcription with original text can be tricky
- apparatus that is tricky to decipher.
Regular expression search of texts. Hypothesis generator of relationships between texts - pattern building as method of spotting irregular patterns.
--
Aris Xanthos (Université de Lausanne) - By scholars, for scholars: a case study on quality assessment of scientific software
Role of software in research increasing. Good proportion of scientific software authored by scholars. These scholars have little formal training, so we can expect different quality from commercial software. This paper based on experience.
Orange Textable. Extension to data mining software. Flexible way of making data tables on the basis of text. During development, less than a quarter of time devoted to code and another quarter to documentation. But sense that time spent on documentation and testing likely to be cut when time is tight in scholarly work.
Tool open.
Conclusions
- software quality relies on time invested in activities other than coding
- software counts less in academic talks/papers
- yet software increasingly relied upon in academic talks/papers
--
Oleksandr Makarenko (National Technical University of Ukraine) - Mathematical Modeling in Scholarship and their Representation in Software
Complexity of translating scholarly processes/systems into formal descriptions.
Important to stress that software is not a single category: educational, research, commercial, entertainment, gaming
--
Q&A
Q: How does building software turn into scholarly discussion?
A: (Aris) Tools that start out as attempting to give the scholar autonomy have turned out to be a way to facilitate dialogue between scholar and developer. Using the software to get at the models of scholarly practice.
Q: What are the meanings of scholarship and research?
A: (Willard) Increased use of research is a triumph of scientific vocabulary over the academy.
A: (Tara) Wissenschaft not excluding the humanities but inadvertently can do so when translated into Anglophonic discourse.
Q: (Manfred) Models for DH infrastructure: characterised as either providing everything that the dumb humanists needs, or practical small examples that promote reuse. The latter can often benefit the community but not the CV of the researcher.
###Panel 2: Confrontation and Collaboration
--
James Baker (British Library) - Removing Black Boxes: Exposing Scholarship to Researchers
Slides: <iframe src="//www.slideshare.net/slideshow/embed_code/44026365" width="425" height="355" frameborder="0" marginwidth="0" marginheight="0" scrolling="no" style="border:1px solid #CCC; border-width:1px; margin-bottom:5px; max-width: 100%;" allowfullscreen> </iframe>
Notes: https://gist.github.com/drjwbaker/3d54d6712a0a124b0530
--
Pieter Francois (University of Oxford) - Connecting Modes of Scholarship through the Library: The genesis of the Sample Generator for Digitized Texts
Library as a meeting ground between CS, humanities research, collections.
Who digitised what, why, when - biases of digitisation are problematic and there is no safety in numbers
How do 70k books realate to 1.6m?
Sample Generator reveals simple truth: fuzzy curatorial dating, lack of digitisation compared to print.
Shifts bias from digitisation to the paper collections
SG imperfect. Perhaps would not have been accepted in HE, as scholarship.
--
Jonas Schneider (Universität Zürich) - Geovisualizing History
How can we bring together time and space in one medium in order to represent the past in a better way? Unfreeze time in GIS.
<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>Now Jonas Schneider (Historisches Lexikon der Schweiz) showing the work he has done geovisualizing historical events. #SiSBern
— Tara (@tla) January 29, 2015
Functions of tools in the reasoning process.
Putting things on maps is the end of knowledge. Tools that put things on maps are intermediaries of knowledge.
Scholarship in Software: located in code and documentation.
###Panel 3: Creation
--
Manfred Thaller (Universität Köln) - Engineering, Science, Art, Scholarship: On implicit assumptions in the software for semantic image databases
Creating code is comparable to art. Problem is, perhaps, the distinction between humanities and science not so deeply ingrained outside of Anglophonic discourses.
Iconclass - a classification system for art and iconography http://iconclass.org/ -- but what does this have to do with programming.
@j_w_baker Can be used to search Wikidata https://t.co/gGVOnAErIW https://t.co/xgQQDlO8qN (or could, if more of the entries were filled in)
— J. Heald (@heald_j) January 29, 2015@j_w_baker Iconclass was a surface of big debate in BL in 1990s: decided eventually not to use it. Widely used elsewhere.
— Andrew Prescott (@Ajprescott) January 29, 2015<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>Manfred Thaller arguing divide isn’t scientists vs humanists, but between those who have rational view of world & those who don’t. #SiSBern
— Seth Denbo (@seth_denbo) January 29, 2015
Linked open data does not account for weightings of links.
Tricky to fulfil humanities criteria with schemas.
<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>Thaller: need computational way to express that the same thing has two meanings for two different observers! (emphasis his) #sisbern
— Tara (@tla) January 29, 2015
--
Nikolas Churik and Brian Clark (College of the Holy Cross) - Composing living scholarship: applying automated acceptance tests to scholarly writing
Unifying prose and code in scholarship. Concordion. Follow the logic of the argument by embedding URNs to texts. [seems very orientated towards work on critical editions]
--
Joris van Zundert (Huygens ING) and Gregor Middell (Independent Scholar) - Code and Authorship in the Humanities
In the DH bubble it may seem unreasonable that software is not treated as scholarship. But from software development for humanities settings it is clear that most humanists don't see rejecting the scholarly merit of programming a problem.
Authorship is a social construct. Appropriation of authorship the foundation of research in the humanities.
So we end up with questions such as where is the theory in your database. Well the answer, is the database is the theory.
@j_w_baker @brandaen I think even in Computer Science, merely developing software isn’t considered scholarship. +
— Amanda French (@amandafrench) January 29, 2015@j_w_baker @brandaen Has to be a new problem addressed and analyzed, which may result in code.
— Amanda French (@amandafrench) January 29, 2015@j_w_baker @brandaen Which is just to say that I don’t see rejecting scholarly merit of programming as a problem either.
— Amanda French (@amandafrench) January 29, 2015Theoretically (Foucault et al) code creation should be recast as authorship. We should read, interpret, critique, study, analyse as we do text.
Authoring is a human activity that occurs through the technologies we have available to us: pen, paper, computer, whatever.
iPython Notebook a nice example of code authorship.
--
QA
Q (Willard) Avoiding binary confrontation -- implicit assignment of argument as finished product -- designation as author from an authors point of view is valuable: what form of multi-author designation works for all authors?
Q Who do we credit and at what level?
A (Joris) We don't credit the creator of the printing press... Software would need the credit roll of a movie.
General... why take authorship and apply it to software. Why not change the authorship model - all depends on what you economy of credit is - old professors will die but then there are young fogies too (and pressures that create young fogies) - and even if the volume of young and old fogies make it hard (futile?) to introduce change, then agency needed for revolution, to shout loudly that things are wrong - journals are proliferating in the humanities that accept these new modes of authorship, but are they preaching to the choir - differences in publishing cultures between hums and science greater than their methodologies -
###David Berry, Softwarization, Archives, and the Digital Humanities
Software is what offers DH an interesting perspective on the digital transformations in the academe.
Whether or not we believe that to softwarize is to industrialise the academe we must ask the question.
Bernard Stiegler (2015) - thought has failed to think techniks - at what point does computation become poison for the humanities? (for some, of course, any...)
Knowledge in the humanities conceived traditionally not as explanation, rules, patterns, but as understanding particularity in context. Explanation could (and does) turn people into objects, robs them of their humanity.
What does it mean for a society to be disenfranchised from its knowledge? (i.e., the code it can't read) And what does it require to overcome this?
Turmoil of traditional scholarly publishing and new forms of knowledge is an important signal for DH.
DH needs to engage more with materiality. 'Exploring how technologies are organised into complex assemblages which create the conditions under which new knowledge and practices are created'
Bringing back the material to the softwarization of the archive.
Softwarization ~ software is eating the world, clearer than ever in our post-Snowden world. But that does not mean it is deterministic, inevitable.
Historicising a system, an assemblage, critical.
Prototype that comes out of a DH project often not what you (the humanist) had in mind: important to be critically reflective of signifiers that are maybe not meaningful outside your (humanities) frame of reference.
Confident that boundaries between academic and tool based work will break down via what is happening to the archive.
The nexus at which DH is located is both interesting and problematic.
Post-digital is a useful position for us to think behind the virtual and the real.
Archive interesting place to think about these issues. It represents materiality. And how it is transformed it brings to the fore the political economy of archiving, publishing, organisation, and institutions.
The memory institution capture and give stability. Curate, care for, make choices, give judgements. When you softwarize an archive you bring with you a lot of baggage: changes things for good and/or for bad.
Logic for storage delegated to algorithms. Amazon - for example - radically archives based on how the algorithm thinks archiving can be optimised for the benefit of its purpose. Illogical and almost non-human. The warehouse has changed in relation to the code rather than the code changed in relation to the warehouse. And of course then people around the edges of the process become objects...
@MechCurBot as softwarized curation.
##Friday
###Roundtable on Peer Review of Digital Scholarly Work
--
Philip Steinkruger, (KU Leuven and Institute for Documentology and Digital Editing (IDE)) Editor of RIDE (Review Journal for digital editions and resources) – Toward a catalogue of criteria for the review of digital editions
Digital editions. Challenge finding reviewers who know the content side and the technical side. To get what RIDE wants reviewers need to fill out a questionnaire - gets data they can work with. Don't include digital editions that are closely yoked to print
--
Nicholas Thely (Université de Rennes) – Toward an evaluation grid for Digital Humanities projects
DH community building in Brittany. Questionnaire to help the speakers understand the expectation of the seminar. Evaluation around the problems of HSS and CS
--
Seth Denbo (AHA) - Making digital history count
AHA's ad-hoc committee on professional evaluation of digital scholarship by historians.
Caleb McDaniel's work on @every3minutes - should that be considered scholarship?
It isn't a very useful question to ask is a tweet scholarship or is a blog scholarship - as this puts the mediums before the content.
The AHA recommend... Argue that work should be valued for its contribution to the conversation not the choice of medium. Nonetheless, it must be evaluated in its native medium. Must be understood as always unfinished. Suggest that scholars document their digital projects well to better enable evaluation. Departments need to be careful thinking about what they expect from a junior digital scholar. Nothing here radical. Not setting aside digital scholarship.
We need a more capacious conception of publication and forms that can be evaluated.
--
Ingrid Kissling (Head of the Humanities and Social Sciences division, Swiss National Science Foundation) – Peer review under revision – The digital challenge for funding agencies
Focus on infrastructure and edition projects. Dissemination and visibility of humanities work.
Peer review from funding agency perspective: bias towards mainstream/established fields/researchers; reviews tend to be generous; no interaction between those proposing and those reviewing.
--
Eugene Lyman (Boston) - Publishing digital project reviews: practical suggestions
<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>#sisbern Gene Lyman: Deans and provosts need to be talking about it if you want software connected to scholarship.
— Joris van Zundert (@brandaen) January 30, 2015
<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>#sisbern Lyman: reviews of digital editions LLC: 120 (>1990), DHQ 13 (>2007), DM-L: 15 (>2006). Let's celebrate and expand that.
— Joris van Zundert (@brandaen) January 30, 2015
This lack of reviews is a problem if we consider the volume of reviews mainstream journals publish...
--
Claire Clivaz (Laboratoire des cultures et humanités digitales, Université de Lausanne) - Reshaping the peer-review process: heretic remarks in a digital time
Peer review at the heart of our identity. Can prove our work. So highly dependant on subjective peer review of work.
Heretic remarks:
- we need peer review work to be valued and excellent peer-review skills to be rewarded
- peer review will become a step post publication rather than pre-publication. Are we ready to imagine new peer-review processes disconnected from publishers?
- we are not ready to peer-review humanities work produced in plural modes
- we need to move from blind peer review to collection peer-review (not least because the blind peer review is no longer blind). Perhaps even on open process.
--
Sacha Zala (Director of the Swiss Historical Association and Director of the Diplomatic Documents of Switzerland) - Some dogmatic postulates for the digital historical sciences
<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>#SiSBern @SachaZala talking about Diplomatic Documents of Switzerland http://t.co/x3OFHjhz8t. Been around since 1995 & online since ’98!
— Seth Denbo (@seth_denbo) January 30, 2015
Postulates:
- open access - more of it
- open source - peak under the hood
- open data - on which are own tools can be applied
--
QA
Funders are 'ahead' of scholarly communities with regards to reviewing scholarly work.
How do we peer review - what do we want to matter?
Editorial role - reviving it (digital scholarly work tends to go ahead without being peer reviewed, collecting together later can add value)
What are quality criteria and research - even within humanities these differ between disciplines.
[Discussion sliding between peer review in terms of funding, pre-publication, and post-publication reviews]
How is software known? By what it does not by how it does what it does.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Exceptions: embeds to and from external sources, and direct quotations from speakers