Stéphane BortzmeyerGood morning, Belval! First day of the Undone conference in computer science. Researches that should have been done but were not.
Let's start #UndoneCS with "Video streaming: how do the socio-economical models shape our research questions?" by Natacha Lapeyroux https://project.inria.fr/videoimpact/
65 % of total Internet traffic!
Is data compression political? Does PeerTube compress differently than YouTube? Why do we compress since in the end it will just lead to more videos sent?
"ICT environmental impact evaluation at sixes and sevens: Leveraging social studies for better use of quantification" by Clément Morand
Is quantitative evaluation the only way to study environmental impact?
"Beyond benchmarks: The undone science of model validation" by Alex Kindel
(Remote presentation, which is an opportunity to watch in real time that research in CS should work more on screen sharing.)
"Building Inclusive Datasets: From Values to Design" by Payal Arora ("professor of inclusive AI" https://payalarora.com/ )
Author of https://mitpress.mit.edu/9780262049306/from-pessimism-to-promise/
She insists that techno-pessimism (and critiques of the tech world) is mostly a thing of the Global North. People in the Global South are much more optimists/enthusiasts about new tech, for instance AI.
(Most optimism about AI in India, least in Australia - for once, the French people are not the most negative 🙂 )
She criticizes the "binary" approach of AI which is trendy in the Global North.
Arora: We need better datasets to train AI, more inclusive. The model of crawling existing data reproduces bias because the AI are trained on the data of the people who publish the most.
After testing chinese payment apps at #IETF125, I discover in Arora's talk at #UndoneCS that there is an indian payment app, Aadhaar Pay https://play.google.com/store/apps/details?id=com.tcs.merchant.cags.boi&hl=ln
From the room: "why has techno-pessimism become so dominant in the western academic world?"
"Mostly a matter of age"
"The Computer Science Undone: How The Social Construction of Disciplinary Boundaries and Disciplinary Hierarchies Shape a Field" by Felienne Hermans https://www.felienne.nl/
A very lively talk, with lot of body language, perfect to avoid going to sleep after lunch.
"Computer science is a science without any interest for research methods. Or for people."
"A bias against the present: Recurring sociotechnical oversights in HCI’s cyclical visions of the future" by Jingyi Li
"Why can't we work on boring UX problems like how to display slides on the big screen [reference to this morning problem in the room] ?"
Panel "The Politics of Computer Science: margins and centres" https://www.undonecs.org/2026/programme.html#panel
Zoe Kahn about her field work in Togo: "people are not worried about access to their data by governments or corporations, they are worried about access by their neighbors" (it echoes the remark of Payal Arora that an homosexual in Pakistan preferred to trust ChatGPT to talk about homosexuality than trusting his family/neighbours).
Just noticed that it is a one-gender panel. (Like a DNS or cryptography meeting at IETF.)
{ Pengouin(BSD|Pdt) }Purée, ils ont une vision à très long terme !
> In June 20243…
Du futur passé ;)
Ou tout simplement une erreur typo, et une absence de relecture correcte ?!
France Natouille 🍷 🥃 🍾@bortzmeyer moien Esch-sur-Alzette 😁