forked from UKRSE/UKRSE.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
talk.html
280 lines (279 loc) · 10.1 KB
/
talk.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
---
title: The Craftsperson and the Scholar
layout: reveal
short: Talk
---
<section><section id="the-state-of-research-software" class="titleslide slide level1"><h1>The State of Research Software</h1></section><section id="the-siro-problem" class="slide level2">
<h1>The SIRO Problem</h1>
<p><del>Garbage in Garbage Out</del></p>
<p>Sensible In, Reasonable Out.</p>
</section><section id="phdware" class="slide level2">
<h1>PhDWare</h1>
<ul>
<li>Don't look if anyone's done it before</li>
<li>Code till it works</li>
<li>Generate a figure</li>
<li>Throw it away</li>
</ul>
</section><section id="labware" class="slide level2">
<h1>Labware</h1>
<ul>
<li>Understood by one genius</li>
<li>Implements great science, now</li>
<li>FORTRAN in any language</li>
<li>Code not engineered for readability</li>
<li>Can't add new science</li>
</ul>
</section><section id="hpcware" class="slide level2">
<h1>HPCWare</h1>
<ul>
<li>Get a 5% improvement in performance</li>
<li>On a particular architecture</li>
<li>Publish a scaling graph</li>
<li>Selection against:
<ul>
<li>Readability</li>
<li>Maintainability</li>
<li>Adaptability</li>
</ul></li>
</ul>
</section><section id="consultantware" class="slide level2">
<h1>ConsultantWare</h1>
<ul>
<li>Little understanding of the science</li>
<li>Overengineered</li>
<li>Unmaintainable by the research group</li>
</ul>
</section></section>
<section><section id="research-software-developers" class="titleslide slide level1"><h1>The Craftsperson And The Scholar</h1></section><section id="the-craftsperson-and-the-scholar" class="slide level2">
<h1>The Craftsperson And The Scholar</h1>
<figure>
<img src="http://www.software.ac.uk/sites/default/files/images/content/ScholarAndCraftsman.jpg" />
</figure>
</section><section id="research-software-developers-1" class="slide level2">
<h1>Research Software Engineers</h1>
<ul>
<li>Facilitative, supportive, and collaborative
<ul>
<li>Deep engagement with research</li>
<li>Sometimes central support</li>
<li>Sometimes part of research group</li>
<li>Can read and understand the papers</li>
<li>Evaluated based on software</li>
</ul></li>
<li>Sustainable and long term
<ul>
<li>Institutional memory</li>
<li>Continuity, stability, maintainance</li>
</ul></li>
<li>Training and teaching
<ul>
<li>Share software best practice in research</li>
</ul></li>
</ul>
</section><section id="the-best-job" class="slide level2">
<h1>The Best Job</h1>
<ul>
<li>Read papers</li>
<li>Write code</li>
</ul>
</section></section>
<section><section id="the-ucl-story" class="titleslide slide level1"><h1>The UCL Story</h1></section><section id="activities" class="slide level2">
<h1>Activities</h1>
<ul>
<li>Projects</li>
<li>Consulting</li>
<li>Training</li>
<li>Infrastructure</li>
</ul>
</section><section id="domain" class="slide level2">
<h1>Domain</h1>
<ul>
<li>From HPC codes to simple scripts</li>
<li>From theoretical physics to digital humanities</li>
</ul>
</section><section id="ucl-staffing" class="slide level2">
<h1>UCL Staffing</h1>
<ul>
<li>3 Permanent Roles</li>
<li>3 research-grant funded roles</li>
<li>Two so far</li>
</ul>
</section><section id="free-projects" class="slide level2">
<h1>Free projects</h1>
<ul>
<li>Judged by executive</li>
<li>Every quarter</li>
<li>Half FTE for term</li>
<li>Worth £8k</li>
</ul>
</section><section id="paid-projects" class="slide level2">
<h1>Paid Projects</h1>
<ul>
<li>Via research grants</li>
<li>Or existing college money</li>
<li>Displaces free calls
<ul>
<li>Until recruit</li>
</ul></li>
<li>Secured ~£0.25M</li>
</ul>
</section></section>
<section><section id="some-sample-projects" class="titleslide slide level1"><h1>Some sample projects</h1></section></section>
<section><section id="low-template-dna" class="titleslide slide level1"><h1>Low-Template DNA</h1></section><section id="whodunnit-code-low-template-dna" class="slide level2">
<h1>Whodunnit code: low-template DNA</h1>
<figure>
<img src="images/whodunnitdna.jpg" />
</figure>
</section><section id="whodunnit-code-low-template-dna-1" class="slide level2">
<h1>Whodunnit code: low-template DNA</h1>
<ul>
<li>Low-template DNA</li>
<li>Relative likelihood of defence and prosecution hypotheses</li>
<li>Optimise over relative degree of contribution from various individuals</li>
</ul>
</section><section id="structural-work-on-likeltd" class="slide level2">
<h1>Structural work on likeLTD</h1>
<ul>
<li>Broken down into functions</li>
<li>Separate objective function from simulated annealing optimiser</li>
<li>Look at using different optimiser libraries</li>
<li>Documentation added</li>
<li>Ready for publication to CRAN</li>
</ul>
</section><section id="performance-improvements-in-likeltd" class="slide level2">
<h1>Performance Improvements in likeLTD</h1>
<ul>
<li>4 times from moving to C</li>
<li>8 times from parallelisation</li>
<li>10 times from change of optimisation algorithm</li>
<li>300 times total</li>
</ul>
</section><section id="performance-improvements-in-likeltd-1" class="slide level2">
<h1>Performance Improvements in likeLTD</h1>
<figure>
<img src="images/ltdFigure.jpg" />
</figure>
</section></section>
<section><section id="dcprogs" class="titleslide slide level1"><h1>DCProgs</h1></section><section id="dcprogs-1" class="slide level2">
<h1>DCProgs</h1>
<figure>
<img src="images/colquhoun.jpg" />
</figure>
</section><section id="old-state-of-the-code" class="slide level2">
<h1>Old State of the Code</h1>
<ul>
<li>Hasn't compiled since 2006</li>
<li>Underpins Nature-published research</li>
</ul>
</section><section id="new-state" class="slide level2">
<h1>New State</h1>
<ul>
<li>Reimplemented in Python</li>
<li>Use standard linear algebra and root finding libraries</li>
<li>Just as performant as the old code</li>
</ul>
</section></section>
<section><section id="hemelb-setup-tool" class="titleslide slide level1"><h1>HemeLB Setup Tool</h1></section><section id="hemelb-setup-tool-1" class="slide level2">
<h1>HemeLB Setup Tool</h1>
<figure>
<img src="images/tubes.png" />
</figure>
</section><section id="voxelisation" class="slide level2">
<h1>Voxelisation</h1>
<figure>
<img src="images/blockgrid.jpg" />
</figure>
</section><section id="robustness" class="slide level2">
<h1>Robustness</h1>
<figure>
<img src="images/cgal.png" />
</figure>
</section><section id="robustness-1" class="slide level2">
<h1>Robustness</h1>
<ul>
<li>~95% to 100%.</li>
<li>Voxelisation in the loop</li>
</ul>
</section></section>
<section><section id="priorities" class="titleslide slide level1"><h1>Priorities</h1></section><section id="priorities-1" class="slide level2">
<h1>Priorities</h1>
<ul>
<li>Readability</li>
<li>Reliability</li>
<li>Performance</li>
</ul>
<p><strong>In that Order</strong></p>
</section><section id="priorities-2" class="slide level2">
<h1>Priorities?</h1>
<ul>
<li>Readability</li>
<li>Reliability</li>
<li>Deployability</li>
<li>Sustainability</li>
<li>Replicability</li>
<li>Usability</li>
<li>Performance</li>
</ul>
<p><strong>In that Order</strong></p>
</section><section id="acknowledgements" class="slide level2">
<h1>Acknowledgements</h1>
<ul>
<li>Clare Gryce, Gavin McLachlan, Mike Cope, Rex Knight</li>
<li>Anthony Finkelstein, Richard Catlow, David Price, John Shawe-Taylor, Simon Arridge, Peter Coveney</li>
<li>Timo Betcke, Michail Stamatakis, David Balding, Miguel Bernabeu, Remis Lape, David Colquhoun</li>
<li>Mayeul D'Avezac, Jens Nielsen, Bruno Silva, Owain Kenway, Ian Kirker, Brian Alston</li>
<li>Neil Chue Hong, Simon Hettrick, Simon Baxter, Kenji Takeda, David de Roure</li>
<li>Andrew Smith, Ben Waugh, Matt Clarkson</li>
<li>Greg Wilson, Fernando Perez</li>
</ul>
</section><section id="contact" class="slide level2">
<h1>Contact</h1>
<p><img src="http://www.justinfoley.ca/transparent-twitter-logo-icon.png" /> <span class="citation" data-cites="uclrcsoftdev">@uclrcsoftdev</span> <span class="citation" data-cites="jamespjh">@jamespjh</span></p>
<p><img src="images/rss.png" /> blogs.ucl.ac.uk/research-software-development</p>
<p><img src="images/email.png" /> [email protected]</p>
</section></section>
<section><section id="the-campaign" class="titleslide slide level1"><h1>The Campaign</h1></section><section id="our-campaign" class="slide level2">
<h1>Our campaign</h1>
<ul>
<li>People <em>want</em> to be RSEs</li>
<li>We need a name, and some status</li>
<li>The research system needs to
<ul>
<li>Promote and preserve its soft benefits better</li>
<li>Measure and incentivise correctly</li>
<li>Integrate RSEs within HEIs</li>
<li>Support institutional innovation</li>
<li>Teach good research software engineering</li>
</ul></li>
</ul>
</section><section id="the-committee-20132014" class="slide level2">
<h1>The Committee 2013/2014</h1>
<ul>
<li><a href="http://www.all-day-breakfast.com/cannam/">Chris Cannam</a></li>
<li><a href="http://dirkgorissen.com/">Dirk Gorissen</a></li>
<li><a href="https://iris.ucl.ac.uk/iris/browse/profile?upi=JHETH53">James Hetherington</a></li>
<li><a href="http://www.software.ac.uk/about/people/simon-hettrick">Simon Hettrick</a></li>
<li><a href="http://cassjohnston.wordpress.com/">Caroline Johnston</a></li>
<li><a href="http://www3.imperial.ac.uk/bioinfsupport/people/mark_woodbridge">Mark Woodbridge</a></li>
</ul>
</section><section id="the-campaign-so-far" class="slide level2">
<h1>The Campaign So Far</h1>
<ul>
<li><a href="http://www.software.ac.uk/blog/2012-11-09-craftsperson-and-scholar?mpw">SSI Collaborations Workshop 2012</a></li>
<li>oERC <a href="http://digital-research-2012.oerc.ox.ac.uk/papers/the-research-software-engineer/view">Digital Research 2012</a></li>
<li><a href="http://www.software.ac.uk/workshop-research-software-engineers">Research Software Engineers meeting 2013</a></li>
<li>Interim Committee</li>
<li><a href="http://www.software.ac.uk/blog/2013-08-23-ten-reasons-be-research-software-engineer">Ten reasons to be an RSE</a></li>
<li>This Website</li>
<li><a href="http://www.timeshighereducation.co.uk/news/save-your-work-give-software-engineers-a-career-track/2006431.article">Press Coverage</a></li>
<li><a href="http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6886129">Lobbying</a></li>
</ul>
</section><section id="what-next" class="slide level2">
<h1>What next?</h1>
<ul>
<li>New Committee</li>
<li>Institution Building</li>
<li>Sharing ideas</li>
</ul>
</section></section>