ALIS-structure

[petermr]

Overall structure of manuscript.

This will determine the sections/ subsections, etc.

Each section will have a different Discussion

[petermr]

Introduction

Ranganathan's legacy

climate - immediacy, complexity

semantic publishing and need for renaissance

Open (access, source, notebooks)

semantic publishing model

structure

nanostructure

semantics

role

ontology - lightweight

multimedia

tooling

conversion

identifiers

versions

knowledge graph

Examples

• static extensible corpora

• homogeneous vs heterogeneous

• UNFCCC

• statis

• PDF, tractable

• templatable

• common ontology

• IPCC Glossary

• scattered over central and reports

• central

• PDF parsing

• headless HTML browsing

• reports

• PDF

• abbreviations

• IPCC Reports

• PDF double-column 

• or headless "Explore" - can be scraped

• scholarly literature

• pygetpapers

• EPMC

• OpenAlex

• preprints

• docanalysis

• classification

• search

• amilib

• annotation of corpus

Toolkit Extensible to:

• city climate
• theses
• legal documents
• NGO/gov reports

Strategy of Semantification

• 2x1 hour Zoom

• video, transcripts

• Slack

• Github

• OpenNotebook

• Issues

• per-corpus

• per-software

• git (technical)

• Python

• PyPI

• git clone

• task-driven

• learn-by-doing

• individual

• chapter

• glossary letter

• knowledge graph

• wikimedia

• wikipedia page

• wikidata item

Software:

• non-specialist , task-driven, legacy conversion

• dynamic

• Python

• libraries

• Colab

• cmdline

• gui?

• packages

• pygetpapers

• docanalysis

• amilib

• amiclimate

Trajectory

• universal climate dictionary
• IPCC knowledge graph
• intelligent scholarly software reader

Discussion on AI/ML(Transcript only)

• how to search for complex concepts (@Anmol Negi)

• search 70 chapters of IPCC for term

• build problem-specific dictionary (e.g. crop pathogens)

[petermr]

list of transcripts

2024-06-03

Components of a dictionary
links to Wikidata, Wikipedia
How ami_dict (just being integrated now) can do much of this automatically

2024-06-04

ALIS article:
I am pleased with the presentation (SESSION 2) https://semanticclimate.slack.com/archives/C011C642XKK/p1717498347585329
that I gave this morning (72 min) . It covers much of what we want to cover in the article and if you play it at 1.5X (so 48 min) I think it's quite listenable (a pseudo podcast!).
It will make a useful artefact for readers of our article to
There are various things that I missed out. They include:
why we are doing semanticClimate
mitigation (if we are lucky)
adaptation (certainly)
understanding of wht is happening and what is likely to happen
Open NotebookScience and our use of repositories, Zoom, Slack, etc.
discovery (pygetpapers)
tools for semantifying IPCC and UNFCCC:
PDF conversion
headless download
stripping publisher cruft
macro-components (documents and immediate subdocuments)
micro-components (annotated sentences, with IDs)
identification of terms (docanalysis)
creation of dictionaries
markup of terms and link to dictionaries
knowledge graphs and visualisation
Tomorrow I plan to have a "1 hour session" where I cover these topics
PLEASE LET ME KNOW IF THERE ARE OTHER THINGS THAT HAVE TO GO IN THE PAPER.
We have to reduce all this to 10 PDF pages. This will be very high-level like the synopsis of a book in Wikipedia or the plot of a movie. It will have multiple links into our resources. For example we can link into the different parts of my presentation , the structures and contents of our major github repositories, etc.
I have scraped the transcript of the presentation (I don't think there is an easy download) and will write software to tidy it.
This is a major step forward in semantic publishing

2024-06-05

This was a 72-min monologue by
@Peter Murray-Rust
which represents about half the content of what I think we want to say in the paper. It's been recorded above and is best played at 1.5X as I take time to think. It sounds a lot better! Makes me 27 years younger. If you have time please listen - if you don't have time, please listen to a bit to get a feel (I want to link to it from the article).
It was transcribed (Slack uses an Amazon service). It's good except Indian names 😞 . It's not easy to read because of additional commas, spaces, lines, etc. I have automatically tidied it with a short regex program in amilib .
Current word count ca 7500.
I intend to add two more pieces:
motivation for semanticClimate in the current apocalypse. Very much welcome your comments. ca. 1200 UK == 1630 IST
technical aspects of converting IPCC text to semantic form including dictionaries
The text will then probably be about 1500 words - verbose and slightly repetitive. We can then run it through an Open Source summariser (I am sure HuggingFace will have one). And I don't care if it's "self-plagiarism" or "cheating" - the purpose of the manuscript is to alert the world to the climate crisis, not worry about academic glory.

2024-06-06

This should conclude what we want to put in the article.
SESSION 2
first chunk Climate and Ranganathan, are the key notes of the article. How semantic can help climate science
second chunk , our technology - what we have to process and the tech we have developed,
OpenNotebookScience
software:
pygetpapers
docanalysis
amilib
amiclimate
dictionaries

2024-06-11

DRAFT of ARTICLE . PMR monologue drafting the toplevel narrative.

These two sessions run together - the break is to drink tea
THIS IS THE COMPLETE NARRATIVE FOR THE ALIS ARTICLE.
It can be listened as audio only - the screen doesn't change . Sessions of 56 and 65 minutes.
For those with fluent English you may find x1.5 speed saves time and boredom.
We will work on the transcript and create a structured HTML narrative with hyperlinks and diagrams
SESSION 1:
Ranganathan's legacy
climate - immediacy, complexity
semantic publishing and need for renaissance
Open (access, source, notebooks)
semantic publishing model
structure
nanostructure
semantics
role
ontology - lightweight
multimedia
tooling
conversion
identifiers
versions
knowledge graph

SESSION 2
Examples
static extensible corpora
homogeneous vs heterogeneous
UNFCCC
statis
PDF, tractable
templatable
common ontology
IPCC Glossary
scattered over central and reports
central
PDF parsing
headless HTML browsing
reports
PDF
abbreviations
IPCC Reports
PDF double-column 😞
or headless "Explore" - can be scraped
scholarly literature
pygetpapers
EPMC
OpenAlex
preprints
docanalysis
classification
search
amilib
annotation of corpus
Toolkit Extensible to:
city climate
theses
legal documents
NGO/gov reports
Strategy of Semantification
2x1 hour Zoom
video, transcripts
Slack
Github
OpenNotebook
Issues
per-corpus
per-software
git (technical)
Python
PyPI
git clone
task-driven
learn-by-doing
individual
chapter
glossary letter
knowledge graph
wikimedia
wikipedia page
wikidata item
Software:
non-specialist , task-driven, legacy conversion
dynamic
Python
libraries
Colab
cmdline
gui?
packages
pygetpapers
docanalysis
amilib
amiclimate
Trajectory
universal climate dictionary
IPCC knowledge graph
intelligent scholarly software reader
Discussion on AI/ML
(Transcript only)
how to search for complex concepts (
@Anmol Negi
)
search 70 chapters of IPCC for term
build problem-specific dictionary (e.g. crop pathogens)

2024-06-12

1 SESSION

catchup and introductions
ALIS manuscript production

2024-06-13

SESSION 1 Thursday
"Ranganathan2024"
SRR would have been actively involved with F/OSS , Wikimedia and Mozilla
comparison of Open Access with F/OSS/Wiki

SESSION 2

• banned from Github
• How we create our article

2024-06-14

• @harshul explains transformers
• collection of all transcripts