Poset documentation polishing: New posets from old ones

[f29946bc-ee7b-48cd-9abc-3445948c551d]

Check documentation for product(), dual() and so on.

This continues the serie of #18925, #18941, #18959, #19141 and #19360.

Component: documentation

Author: Jori Mäntysalo

Branch/Commit: e3e3aa9

Reviewer: Kevin Dilks

Issue created by migration from https://trac.sagemath.org/ticket/19435

[f29946bc-ee7b-48cd-9abc-3445948c551d]

Branch: u/jmantysalo/poset_documentation_polishing__new_posets_from_old_ones

[f29946bc-ee7b-48cd-9abc-3445948c551d]

comment:2

This is a long but quite trivial patch. Mostly bikeshedding.

I don't know easy way to make with_bounds() work with non-facade posets. At least now the code gives nicer error message. I also changed completion_by_cuts() to return the empty lattice from the empty poset. It is now consistent with definition "smallest lattice containing..."

---

New commits:

41e3a4b

Slight modifications to documentation.

[f29946bc-ee7b-48cd-9abc-3445948c551d]

Commit: 41e3a4b

[1adc0eef-8957-46d9-975b-2dd71dfbd9ba]

comment:3

• Potential issue with the example in connected_components() giving the covering relations of the first connected component, since cover_relations() returns an ordered list, but strictly speaking is an unordered set. Not really an issue, since I imagine the code will always return it in the order given, but it's something to think about.

• When talking about dual/ordinal sum of a lattice (resp join-/meet- semilattices), the documentation appears to use 'lattice' etc. as adjectives (ie, 'the dual of a lattice is lattice'). All of those should be like 'the dual of a lattice is a lattice'.

• I'm not happy about the definition of Dedekind-Macneille completion being vague about the concept of 'smallest' and not being explicit about the original poset being an induced subposet of the lattice, but I suppose it matches what Wikipedia has, and does link to the full Wikipedia article.

[f29946bc-ee7b-48cd-9abc-3445948c551d]

comment:4

Replying to @kevindilks:

• Potential issue with the example in connected_components() giving the covering relations of the first connected component, since cover_relations() returns an ordered list, but strictly speaking is an unordered set. Not really an issue, since I imagine the code will always return it in the order given, but it's something to think about.

This is true, and I do not know a good way to overcome this. # random order is one possibility. For a user perspective I see no problem when documentation says that a function returns [3, 5], but he/she gots [5, 3]; if the user has any clue at all, he will understand that the list represents a set.

For some posets we could use linear_extension=True, but it only applies to list of elements, and is kind of noise to the user.

[f29946bc-ee7b-48cd-9abc-3445948c551d]

comment:5

Replying to @kevindilks:

• I'm not happy about the definition of Dedekind-Macneille completion being vague about the concept of 'smallest' and not being explicit about the original poset being an induced subposet of the lattice, but I suppose it matches what Wikipedia has, and does link to the full Wikipedia article.

Could we have an option to show this clearly? I.e. somehow show what are "original" and "added" elements on the completion?

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 41e3a4b to 4283464

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Branch pushed to git repo; I updated commit sha1. New commits:

4283464

Some fixes to poset documentation.

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Changed commit from 4283464 to e3e3aa9

[7ed8c4ca-6d56-4ae9-953a-41e42b4ed313]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

f8956b9	Slight modifications to documentation.
f436b4b	Some fixes to poset documentation.
e3e3aa9	Rebase and minor tweaks.

[f29946bc-ee7b-48cd-9abc-3445948c551d]

comment:8

Continuing with this again. Nothing special in this.

[1adc0eef-8957-46d9-975b-2dd71dfbd9ba]

Reviewer: Kevin Dilks

[f29946bc-ee7b-48cd-9abc-3445948c551d]

comment:10

Thanks Kevin. rc0 is out, so I changed milestone.

[vbraun]

Changed branch from u/jmantysalo/poset_documentation_polishing__new_posets_from_old_ones to e3e3aa9