Bs1 bibliography

[stumbo]

An initial Hugo script to create a page listing all the bibliographical entries (here expected to be in a folder content/en/publications)

This work is currently live on the staging site at publications.

This is a step back from some of my initial ideas, but is easier to implement and provides a framework discussion and for adding additional functionality.

While the current implementation uses content/en/publications as the location of the bibliography entries my end goal is to replace the existing bibliography page.

I purposely kept this PR separate from Matt's work, #279, to isolate any changes.

Next steps:

• Create the individual page layout. This will be a more extensive presentation of the information within the Zotero bibliographic entry. Links to urls to view the article will be provided
• Creation of css to style the listing and individual pages
• Integration with the work Matt has done to create individual markdown files.
  • I'm currently generating the individual pages with an R script I had laying around. It's ugly put gives me full control over the content. Nice for debugging.
  • Merge into Matt's work and leverage his work on creating individual md files
• Along those lines, handling 404 errors. If a bibliographic entry disappears or is changed, do we want a custom 404 page separate from the Interlisp.org page?
• Longer term, version 2?, we can focus on sorting and filtering of entries, leveraging keywords in the bibliographic entries.

[pamoroso]

The page looks clean. But what about turning the entries from bullet items to just paragraphs with the same formatting (i.e. everything except the bullet)? This would make the entries closer to the left margin. And over 600 bullets may not improve scannability much.

What do others think?

[stumbo]

The page looks clean. But what about turning the entries from bullet items to just paragraphs with the same formatting (i.e. everything except the bullet)? This would make the entries closer to the left margin. And over 600 bullets may not improve scannability much.

What do others think?

I agree - and this is where work on the css will come in. For now, I just went with vanilla html and didn't spend time worrying about the formatting.

[masinter]

404 -- a custom 404 per se isn't necessary and might be difficult. As long as there's a 404?

A date YYYY-01-01 I suspect is how it is printing dates that were just known YYYY. It would be better to not lose the distinction between 'month and day unknown' or 'month known, but day not known' and 'date really is january 01' or 'date really is YYYY-MM'

[stumbo]

I just pushed a commit that makes several major updates. From the commit note:

The bibliography md files are build using the update_bibliography.sh
script. bib-fns.jq provides functions to reformat the Zotero JSON.
bibSplit.pl creates the md file and a data file containing the JSON
for each page. Future work will remove the JSON data files once all the
content is encompassed in the md file.

Added new document type bibliography to generate the list of
bibliography entries. layouts/bibliography/list.html provides this
functionality. It also currently includes some light css to format the
output. The css will eventually be moved into a standalone file within
the Hugo ecosystem.

Removed the original publication type and its associated files in
layouts.

Removed some HugoBlox files. They were left over from an early attempt
to create a bibliography using HugoBlox as a template.

Updated gh-pages.yml, the GHA, to cache the md pages and JSON files.
Rewrote the caching to make it more efficient and easier to follow.

The updates are live on the staging site. Here's a preview of what the interface currently looks like:

Entries are currently sorted alphanumerically by title.

[hjellinek]

This looks marvelous. I see two problems to fix before going live, though:

First, the entries do not link to the relevant entry on the Zotero website. That's what I would assume would happen upon clicking the "Read more" link, but that only takes you to the abstract. How about having two links: one a direct link to the item in Zotero, ("View in bibliography") and another that takes you to the abstract (the current behavior) labeled "View abstract"? "Read more" is too vague.

A big motivation behind this work, as you know, is to get this page, and by extension Interlisp.org, indexed by search engines. So from a SEO perspective, placing the abstract on the page without linking out to it will provide better grist for the indexing mill. (If the abstract is too long, extract the first paragraph or so.)

Second, the plethora of HTML character entities that peek through into the rendered text. (They do not appear in the linked abstract pages.)

For instance, look at the text snippet for A clipping divider:

When compared with a drawing on paper, the pictures presented by today’s computer display

Also in Ambitious evaluation: a new reading of an old issue:

Much has been written about Lazy Evaluation in Lisp—less about the other end of the spectrum—Ambitious Evaluation.

There are many other instances of unrendered character entities in the text. On the staging site, search for "&" and see them.

[MattHeffron]

Looking good, but:

• I do not get vertical scroll bars for either the "main" or "Authors" panes. (Chrome on Windows 11.)
• In list.html the handling of abstract looks like what I had, but for "Special Files on <Test>ARs>" (key: 7KUMIRNC) it is corrupted. It's correct in the title.
  • It appears this may be related to some other condition?? The abstract shown is corrupted as noted above, but it is correct for "SEdit test report", key: RJLLRQDR). However, the <word>, that looks like an HTML tag is removed in both cases on the single page for the items.
• bibSplit.pl isn't handling extended characters in author names, e.g., "Emanuelson, Pär" (in "On compiling embedded languages in LISP" key: ALMXAIQC; and other items). They're correct in the JSON. (Presumably, there are other places where these characters would be an issue.) They're corrupted in the .md files, so it isn't a Hugo problem.

I am testing locally with hugo server -e development and I'm running hugo version 0.152.2.
(The problem that I previously had mentioned about it failing to run as server, was due to a bug in version 0.152.0, which snap had updated to! I'm not using snap for this now.)

[MattHeffron]

• I do not get vertical scroll bars for either the "main" or "Authors" panes. (Chrome on Windows 11.)

Per discussion at meeting 10/29 -- this is a Chrome issue, not restricted to Windows. (I.e., scrollbars display correctly in Firefox.)

[nbriggs]

Not sure if this should be a separate issue -- in the medley timeline view on the staging site, there's oddly inconsistent indentation in the decadal section headers, notably for the 1980s:

[stumbo]

This is ready to merge. What is currently running on the staging site accurately reflects this code, the only material difference is the staging site has a flag setting it to staging. Beyond that there are no differences.

Not sure if this should be a separate issue -- in the medley timeline view on the staging site, there's oddly inconsistent indentation in the decadal section headers, notably for the 1980s:

This is a separate issue and unrelated to the bibliography work. I'll open a separate issue for it.

[hjellinek]

@stumbo, were you planning to replace the "Cite: placeholder for link to retrieve citation from Zotero." text before releasing this?

[hjellinek]

Also, the bug where you click on an author's name (e.g., "Bates, Raymond L."), then click "All Authors" and find yourself on an empty "Authors" page?

[MattHeffron]

I believe that I have a fix for the problem I mentioned:

• bibSplit.pl isn't handling extended characters in author names, e.g., "Emanuelson, Pär" (in "On compiling embedded languages in LISP" key: ALMXAIQC; and other items). They're correct in the JSON. (Presumably, there are other places where these characters would be an issue.) They're corrupted in the .md files, so it isn't a Hugo problem.

Should that be a separate issue after accepting/merging the bs1_bibliography branch, should I make a change here now, or on the bs2_bibliography branch? (bs2_bibliography doesn't have a PR yet)
I noticed that there are other differences between these 2 branches, e.g., hugo version.

[stumbo]

@stumbo, were you planning to replace the "Cite: placeholder for link to retrieve citation from Zotero." text before releasing this?

Good catch, I read right over that multiple times. I'm removing it for now. Something to circle back around and add later. It's been removed.

[stumbo]

Also, the bug where you click on an author's name (e.g., "Bates, Raymond L."), then click "All Authors" and find yourself on an empty "Authors" page?

I was going to fix this as a fast follow. I have some ideas on what's causing the problem. But, it'll take a little time to figure it out.

[stumbo]

I believe that I have a fix for the problem I mentioned:
...
Should that be a separate issue after accepting/merging the bs1_bibliography branch, should I make a change here now, or on the bs2_bibliography branch? (bs2_bibliography doesn't have a PR yet) I noticed that there are other differences between these 2 branches, e.g., hugo version.

Let's fix it as a separate PR. And, no don't use bs2_bibliography that as you rightly noticed isn't what we're releasing. It's a different approach I took in attempting to get by some deployment issues. It's more of a historical artifact right now.