Konrad Hinsen - Revision history

Jonny at 00:47, 12 November 2022

2022-11-12T00:47:22Z

← Older revision		Revision as of 00:47, 12 November 2022
Line 9:		Line 9:
	\|Fediverse=https://scholar.social/@khinsen		\|Fediverse=https://scholar.social/@khinsen
	\|Github=khinsen		\|Github=khinsen
	}}		}}
	{{Workshop Submission		{{Workshop Submission
	\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.		\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.

Jonny at 00:47, 12 November 2022

2022-11-12T00:47:11Z

← Older revision		Revision as of 00:47, 12 November 2022
Line 9:		Line 9:
	\|Fediverse=https://scholar.social/@khinsen		\|Fediverse=https://scholar.social/@khinsen
	\|Github=khinsen		\|Github=khinsen
	}}		}}
	{{Workshop Submission		{{Workshop Submission
	\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.		\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.

Jonny at 00:46, 12 November 2022

2022-11-12T00:46:17Z

← Older revision		Revision as of 00:46, 12 November 2022
Line 9:		Line 9:
	\|Fediverse=https://scholar.social/@khinsen		\|Fediverse=https://scholar.social/@khinsen
	\|Github=khinsen		\|Github=khinsen
	}}		}}
	{{Workshop Submission		{{Workshop Submission
	\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.		\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.

Khinsen at 09:44, 5 November 2022

2022-11-05T09:44:32Z

← Older revision		Revision as of 09:44, 5 November 2022
Line 35:		Line 35:

	And, of course, the code: https://github.com/khinsen/leibniz-pharo		And, of course, the code: https://github.com/khinsen/leibniz-pharo
	\|Organizer Topics=Research Data, Reproducibility, Federation, Gradual Enrichment		\|Organizer Topics=Research Data, Reproducibility, Federation, Gradual Enrichment, Documents, Wikis, Knowledge Graphs, Transdisciplinarity, Scientific Publishing
	}}		}}

Khinsen at 09:40, 5 November 2022

2022-11-05T09:40:02Z

← Older revision		Revision as of 09:40, 5 November 2022
Line 35:		Line 35:

	And, of course, the code: https://github.com/khinsen/leibniz-pharo		And, of course, the code: https://github.com/khinsen/leibniz-pharo
	\|Organizer Topics=Research Data, Reproducibility, Federation		\|Organizer Topics=Research Data, Reproducibility, Federation, Gradual Enrichment
	}}		}}

Khinsen at 09:59, 2 November 2022

2022-11-02T09:59:33Z

← Older revision		Revision as of 09:59, 2 November 2022
Line 1:		Line 1:
	{{Participant		{{Participant
	\|Timezone=CET (GMT+01:00/GMT+02:00)		\|Timezone=CET (GMT+01:00/GMT+02:00)
			\|Homepage=http://khinsen.net/, https://science-in-the-digital-era.khinsen.net/
	\|Affiliation=CNRS		\|Affiliation=CNRS
	\|Projects=Leibniz		\|Projects=Leibniz
Line 17:		Line 18:

	Although my current prototype is a single-user environment, the long-term objective is a collaborative computational medium for scientific models. My main inspiration for this aspect is the consensus-enabling structure of Federated Wiki, but I really hope to learn more about this problem space from other participants of the workshop.		Although my current prototype is a single-user environment, the long-term objective is a collaborative computational medium for scientific models. My main inspiration for this aspect is the consensus-enabling structure of Federated Wiki, but I really hope to learn more about this problem space from other participants of the workshop.
			\|Frame=Practitioner, Tool-builder
	\|Materials="Computational science: shifting the focus from tools to models" https://f1000research.com/articles/3-101/v2		\|Materials="Computational science: shifting the focus from tools to models" https://f1000research.com/articles/3-101/v2
	An in-detail explanation of the problem, with an outline of a solution, written in 2014 before I actually started working on my prototypes.		An in-detail explanation of the problem, with an outline of a solution, written in 2014 before I actually started working on my prototypes.

Khinsen: Add ORCID and social network handles

2022-11-02T07:53:28Z

Add ORCID and social network handles

← Older revision		Revision as of 07:53, 2 November 2022
Line 4:		Line 4:
	\|Projects=Leibniz		\|Projects=Leibniz
	\|Table Assignment=Table 1		\|Table Assignment=Table 1
	}}		\|ORCID=0000-0003-0330-9428
			\|Twitter=khinsen
			\|Fediverse=https://scholar.social/@khinsen
			\|Github=khinsen
			}}
	{{Workshop Submission		{{Workshop Submission
	\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.		\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.

Jonny at 07:05, 26 October 2022

2022-10-26T07:05:21Z

← Older revision		Revision as of 07:05, 26 October 2022
Line 4:		Line 4:
	\|Projects=Leibniz		\|Projects=Leibniz
	\|Table Assignment=Table 1		\|Table Assignment=Table 1
	}}		}}
	{{Workshop Submission		{{Workshop Submission
	\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.		\|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.

Jonny at 06:57, 26 October 2022

2022-10-26T06:57:21Z

← Older revision		Revision as of 06:57, 26 October 2022
Line 3:		Line 3:
	\|Affiliation=CNRS		\|Affiliation=CNRS
	\|Projects=Leibniz		\|Projects=Leibniz
			\|Table Assignment=Table 1
	}}		}}
	{{Workshop Submission		{{Workshop Submission
Line 12:		Line 13:

	Although my current prototype is a single-user environment, the long-term objective is a collaborative computational medium for scientific models. My main inspiration for this aspect is the consensus-enabling structure of Federated Wiki, but I really hope to learn more about this problem space from other participants of the workshop.		Although my current prototype is a single-user environment, the long-term objective is a collaborative computational medium for scientific models. My main inspiration for this aspect is the consensus-enabling structure of Federated Wiki, but I really hope to learn more about this problem space from other participants of the workshop.
	~~\|Frame=Tool-builder~~
	\|Materials="Computational science: shifting the focus from tools to models" https://f1000research.com/articles/3-101/v2		\|Materials="Computational science: shifting the focus from tools to models" https://f1000research.com/articles/3-101/v2
	An in-detail explanation of the problem, with an outline of a solution, written in 2014 before I actually started working on my prototypes.		An in-detail explanation of the problem, with an outline of a solution, written in 2014 before I actually started working on my prototypes.

Jonny: Created page with "{{Participant |Timezone=CET (GMT+01:00/GMT+02:00) |Affiliation=CNRS |Projects=Leibniz }} {{Workshop Submission |Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulati..."

2022-10-26T06:54:49Z

Created page with "{{Participant |Timezone=CET (GMT+01:00/GMT+02:00) |Affiliation=CNRS |Projects=Leibniz }} {{Workshop Submission |Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulati..."

New page

{{Participant
|Timezone=CET (GMT+01:00/GMT+02:00)
|Affiliation=CNRS
|Projects=Leibniz
}}
{{Workshop Submission
|Interest=Ever since I started doing computational biophysics research 25 years ago, I have been unhappy with the state of the art in dealing with complex computational models. Then and still now, they are represented in two unconnected forms: a narrative providing the scientific background and a summary, and an optimized (and hence unreadable) piece of simulation software the embeds the model. The narrative is usually subject to peer review, but review has to remain superficial in the absence of the complete model. The software source code usually undergoes no independent evaluation because that's far too much effort to ask from a reviewer. For the potential user of the model, there is a version they can run but not understand, and a summary that they can understand but not compare to what they are running. Given the increasing importance of computational models in many fields of science, that's a problem. Scientific models ought to take center stage in scientific debate, but this is currently not possible.

My first attempt to improve the situation was the move to high-level programming languages. In 1995, I co-founded the "Numerical Python" project, which became the nucleus of the Scientific Python ecosystem. While Python code was indeed more readable and better modifiable than the previously common Fortran or C codes, this didn't change the fundamental situation of two unconnected representations. Also, Python code is still too alien for human readers to make independent evaluation of scientific software a reality.

Another obstacle to doing computational science is that computational models are, in general, specifications rather than algorithms. A runnable implementation is thus needlessly specific, making it difficult to use computational tools (formal methods, code generators, future AI-based tools) to reason about the model. My current second attempt at improving the situation is therefore (1) a specification language and (2) embedded into a human-readable narrative, much like traditional mathematical notation. The second major iteration of this project, which I named "Leibniz" in honor of Gottfried Wilhelm Leibniz, just reached the point where I could make a public demo.

Although my current prototype is a single-user environment, the long-term objective is a collaborative computational medium for scientific models. My main inspiration for this aspect is the consensus-enabling structure of Federated Wiki, but I really hope to learn more about this problem space from other participants of the workshop.
|Frame=Tool-builder
|Materials="Computational science: shifting the focus from tools to models" https://f1000research.com/articles/3-101/v2
An in-detail explanation of the problem, with an outline of a solution, written in 2014 before I actually started working on my prototypes.

"The structure and interpretation of scientific models" http://blog.khinsen.net/posts/2020/12/10/the-structure-and-interpretation-of-scientific-models/
A blog post explaining why computational models are, in general, specifications rather than algorithms.

"Verifiability in computer-aided research: the role of digital scientific notations at the human-computer interface" https://peerj.com/articles/cs-158/
A report on the first iteration of Leibniz, written in 2018.

"Liberating computational science from software complexity" https://www.youtube.com/watch?v=YbznItQpALo&t=2104s
A recorded talk at RacketCon 2020, with a motivating introduction and a demo of the first iteration.

"Using Glamorous Toolkit for Scientific Notation" https://www.youtube.com/watch?v=f10NpsMmbis
A first demo of the second iteration, in discussion with Tudor Girba, the chief architect of the platform I chose to build on.

And, of course, the code: https://github.com/khinsen/leibniz-pharo
|Organizer Topics=Research Data, Reproducibility, Federation
}}