Prez pour la présentation d'un papier au LIP - retour accueil
git clone git://bebou.netlib.re/prez-lip
Log | Files | Refs |
prez.slides (9041B)
1
2
3
4
5
6
7 Presentation of
8 "A Case for Feminism in Programming Language Design"
9 Felienne Hermans and Ari Schlesinger
10
11
12
13
14
15
16 › Self presentation
17
18 Arthur Pons
19 Computer engineer by trade
20 Worked as a research engineer at the Université de Strasbourg
21 Interested in the relationship between IT and environmental issues
22
23 PhD thesis with Commown, LIP and CITI
24 "Eco-conception de services d'infogérance low-tech pour la sobriété et la résilience"
25 "Eco-design of low-tech IT services for sufficiency and resilience"
26
27
28 › Why read a paper about feminism in a computer science context
29
30 1. Gain awareness about our biases, mainly in what type of knowledge we
31 produce and how we produce it
32 2. Use feminist theory, notably the concept of care, to help foster a
33 resilient, sustainable and empowering IT infrastructure for everyone
34
35 This paper focuses on point number 1
36
37 › Disclaimer
38
39 I don't do research in programming language design
40 I don't do research in feminism
41
42 › The authors
43
44 Felienne Hermans is a scientist at Vrije Universiteit Amsterdam
45 PhD in software engineering
46 Her interests are :
47 programming education
48 spreadsheets
49
50 Ari Schelsinger is a scientist at Georgia University
51 PhD in Human-Centered Computing
52 Her interests are :
53 understanding how social issues are encoded into technical objects and
54 infrastructure
55
56
57 › The main questions
58
59 Directly for the abstract
60
61 1) What does it mean to design a programming language?
62 2) Why does minimal demographic diversity persist in the programming
63 language community?
64
65 › Feminism as a critical thinking tool
66
67 "I understand thinking of feminism is confusing for PL people, trust me, it did
68 not come naturally to me either! I thought feminism was just about gender. How
69 can gender, of all things, play a role in programming language design?"
70 › Feminism as a critical thinking tool
71
72 "I understand thinking of feminism is confusing for PL people, trust me, it did
73 not come naturally to me either! I thought feminism was just about gender. How
74 can gender, of all things, play a role in programming language design?"
75
76 The authors rely on modern feminist epistemology, more specifically on a
77 framework built for a glaciology paper :
78
79 Carey, Mark, M. Jackson, Alessandro Antonello, and Jaclyn Rushing.
80 “Glaciers, Gender, and Science: A Feminist Glaciology Framework for Global
81 Environmental Change Research.” Progress in Human Geography 40, no. 6
82 (December 1, 2016): 770–93. https://doi.org/10.1177/0309132515623368.
83
84 › Feminism as a critical thinking tool
85
86 "Feminism Can Help Question Values and Priorities of Programming Languages"
87
88 "Feminism Can To Better Us Understand Science"
89
90 "A central principle in feminism [...] is that knowledge is shaped by the
91 context in which it is made"
92
93 "Feminism Is About More Than Gender" (inclusivity)
94
95 "Feminism takes into account how discrimination across many different
96 social and cultural identities is interconnected"
97
98 "Feminism Is For Everyone"
99
100
101 › The paper's thesis
102
103 "Diversity in both the design of PL and the demographics of the community
104 are limited because of the dominant culture that prioritizes theory and
105 formalism over people and social impact."
106
107
108 › The paper's thesis
109
110 "Diversity in both the design of PL and the demographics of the community
111 are limited because of the dominant culture that prioritizes theory and
112 formalism over people and social impact."
113
114 The goal here is not to argue the dominant culture is intrinsically bad or that
115 it produces bad research.
116
117 "This essay will not answer, but raise questions about how we can make the
118 field inclusive of more types of research, and more types of people,
119 including all genders."
120
121 › Structure
122
123 1. CS and PL domain are not very diverse
124 2. Gendered Science and Knowledge
125 3. Systems of Scientific Domination
126 4. Knowledge Production
127 5. Alternative Representations
128
129 › Structure
130
131 1. CS and PL domain are not very diverse
132 2. Gendered Science and Knowledge
133 3. Systems of Scientific Domination
134 4. Knowledge Production
135 5. Alternative Representations
136
137 The authors provide statistics and anecdotes backed up by feminist literature
138 to argue for each of those points. I feel a strong will to convince their
139 peers this is a worthwhile endeavour.
140
141 › CS domain is not very diverse
142
143 "Only 10% of CS papers are authored by women"
144
145 Girls are less likely to enter out-of-school programming clubs
146
147 "Girls feel that they cannot enter the world of computers without
148 endangering their sens of femininity" (which the rest of society pushes on
149 them)
150
151 In maths "research has found teachers view girls as successful [...] thanks to
152 their hard work, while they believe boy's success comes from their talent."
153
154 › It has concrete consequences on the PL domain
155
156 PL domain historically was, and still is, largely western
157 => by default only 0..9 numerals
158 ١+١ won't work
159
160 Most people are not visually impaired
161 => "system.cout>> reads "system <pause> c out greater than greater than" in
162 screen readers"
163
164 › Gendered Science and Knowledge
165
166 PL was made simultaneously mathematical and masculine
167
168 * At first an unusually feminine line of work
169 * With more complex architectures and automation came formalisation and
170 prestige*
171 * Now considered a domain that " requires vast mental powers, a kind of genius
172 with formalism akin to that of the mathematician"
173
174 *Very well documented via the "Humble Programmer" speech of Dijkstra
175 › Gendered Science and Knowledge
176
177 Traditionally valued :
178
179 * Complex maths*
180 * Research on language features
181 * Building tools
182 * Quantitative methods
183 * Research that is "difficult"
184
185 Less valued :
186
187 * Use of said tools
188 * Qualitative methods
189 * A will to broaden research topics
190 * Research that looks or feel "easy"
191
192 * "a senior academic jokingly say that a paper was rejected because ‘it did
193 not have enough Greek letters for POPL’"
194
195 › Gendered Science and Knowledge
196
197 "We are robbing ourselves of a place for conversations on the different
198 perspectives on the ways people use with programming languages."
199 › Systems of Scientific Domination
200
201 Structural domination through norms
202 * Guidelines for conferences usually don't explain
203 what constitutes good qualitative work
204 * Without the right tools reviewers are incentivize to place these kind of
205 work out of scope
206 * Change in norms have to come through the community but people who are
207 influential enough to do it are usually there because they upheld those
208 norms themselves
209
210 Interpersonal domination
211 People with less mainstream research topics, methodologies or a non
212 majority identity (e.g being a black women) spend (conference) time :
213
214 * answering entry level questions
215 * doing "diversity work"
216
217 instead of seizing opportunities to learn and build connections about their
218 scientific interests
219
220 › Knowledge production
221
222 History of PL mainly retains men (although less in recent history)
223 builders
224
225 This lens omits the perspectives of maintainers, teachers, advocates or critics.
226
227 › Knowledge production
228
229 Our biases go as deep as informing what even counts as a PL
230
231 "A recent paper that investigated how to best teach programming to people
232 in the prison system describes a struggle to install the right software to
233 do so. When at their conference presentation they were asked why they did
234 not consider Excel (which in addition to formulas includes a VBA
235 interpreter), which was available, the audience burst out in laughter."
236
237 › Alternative representations
238
239 Alternative methodologies :
240
241 * SocioPLT - What factors lead to programming language being adopted
242 * Controlled Experiments - What impact error messages have on the use of PL
243 * User-centered Design - Involving users in the design process of PL en citer
244
245 Alternative PL :
246
247 firefox https://github.com/wordplaydev/wordplay/blob/main/LANGUAGE.md
248 yt c https://www.youtube.com/watch?v=77KAHPZUR8g
249 firefox https://wy-lang.org/
250 firefox https://www.hedy.org
251
252 feh --fullscren -. crochet.jpg
253
254
255 › Further reading and watching
256
257 Ensmenger's paper on the genesis of masculine archetypes in IT (the hacker, the
258 corporate guy à la Bill Gates, the scientist etc), how they are related and how
259 they differ :
260
261 Ensmenger, Nathan. “‘Beards, Sandals, and Other Signs of Rugged Individualism’:
262 Masculine Culture within the Computing Professions.” Osiris 30 (January 1,
263 2015): 38–65. https://doi.org/10.1086/682955.
264
265 Isabelle Collet presentation "Femmes et numérique : pratiques égalitaires,
266 dispositifs inclusifs", in french and in a "Science communication" tone :
267
268 https://jres.ubicast.tv/videos/femmes-et-numerique-pratiques-egalitaires-dispositifs-inclusifs_n26a9jqx6e/
269
270 Presentation sources :
271
272 git : http://git.bebou.netlib.re/prez-lip/log.html
273 ssh : ssh guest@bebou.netlib.re -p 1459 prez-lip
274
275
276
277
278
279 › Les slides
280
281 git : http://git.bebou.netlib.re/prez-lip/log.html
282 ssh : ssh -t guest@bebou.netlib.re -p 1459 prez-lip