Skip to content

Latest commit

 

History

History
64 lines (47 loc) · 2.8 KB

WorkNotes.md

File metadata and controls

64 lines (47 loc) · 2.8 KB

Notes on work-related things

Links

Datasets

List of target datasets

  • Abstract polytopes and maniplexes (Gabe, Katja)
  • Symmetric graphs (Katja, Gabe)
  • Lattices (Jukka)
  • Small Groups Library export (Michael T., Katja)

Things to do with each

  1. Prioritize (make list in order of importance) properties of datasets for the other dataset tasks.
  2. Write schema theories
  3. Formalize relevant math
  4. Import into
  5. Provenance

Initial assignments:

  • 1: Gabe, Jukka, Katja
  • 2 and 3: Gabe, Jukka, Katja, Michael T. with Dennis
  • 4: Gabe, Jukka, Katja, Michael T. with Tom

Other

  • Connect the small groups with Group Names and the Group Explorer.
  • How to deal with the above in general way, (so that the below can also be handled) so that we can add such things without too much work - Florian?
  • How to deal with p-adic
  • Come up with a good name for the fourth kind of mathematical data (previously known as concrete/relational/...)
  • Think about recording provenance
  • Think about a journal of math datasets

Datatypes and Codecs

Our MDDL (math data description language) defines two kinds of objects:

  • datatypes specify types using their mathematical properties, such as ** basic types: integers, rationals, booleans, etc. ** collection types: vectors, matrices, finite sets, multisets, etc. ** higher mathematical types: polynomials, rings, etc.
  • codecs specify translations (and back) from datatypes to primitive types supported by databases, such as ** integers as strings, integers as lists of base 2^64 digits, etc. ** sparse vector codec (i.e., vectors as sparse lists), taking a codec for the entries as an argument

The datatypes are formally specified and the codecs are named and documented in the file https://gl.mathhub.info/ODK/discretezoo/blob/master/source/MDDL.mmt

This file serves as the reference for our implementations. It will probably be extended substantially during the course of the workshop.

Interface

  • Choosing collections (defaults?), choosing columns displayed by default
  • Support the filter box with autocomplete-type search (in case there are many)
  • Usability: intro tour of the interface (see intro.js), pop-up definitions
  • Out there: new graphic image and/or colour scheme to integrate with MathHub?

Computer algebra system integration

  • How to do it? - MitM
  • Sage explorer connection?