During the sessions with my French tutor, I share a Google document so that we can mark the words where I need to practice my pronunciation some more or tweak the wording. Using Ctrl+B to make the word as bold is an easy way to make it jump out.

I used to copy these changes into my Org Mode notes manually, but today I thought I'd try automating some of it.

First, I need a script to download the HTML for a specified Google document. This is probably easier to do with the NodeJS library rather than with oauth2.el and url-retrieve-synchronously because of various authentication things.

require('dotenv').config();
const { google } = require('googleapis');

async function download(fileId) {
  const auth = new google.auth.GoogleAuth({
    scopes: ['https://www.googleapis.com/auth/drive.readonly'],
  });
  const drive = google.drive({ version: 'v3', auth });
  const htmlRes = await drive.files.export({
    fileId: fileId,
    mimeType: 'text/html'
  });
  return htmlRes.data;
}

async function main() {
  console.log(await download(process.argv.length > 2 ? process.argv[2] : process.env['DOC_ID']));
}

main();

Then I can wrap a little bit of Emacs Lisp around it.

(defvar my-google-doc-download-command
  (list "nodejs" (expand-file-name "~/bin/download-google-doc-html.cjs")))

(defun my-google-doc-html (doc-id)
  (when (string-match "https://docs\\.google\\.com/document/d/\\(.+?\\)/" doc-id)
    (setq doc-id (match-string 1 doc-id)))
  (with-temp-buffer
    (apply #'call-process (car my-google-doc-download-command)
           nil t nil (append (cdr my-google-doc-download-command) (list doc-id)))
    (buffer-string)))

(defun my-google-doc-clean-html (html)
  "Remove links on spaces, replace Google links."
  (let ((dom (with-temp-buffer
               (insert html)
               (libxml-parse-html-region))))
    (dom-search
     dom
     (lambda (o)
       (when (eq (dom-tag o) 'a)
         (when (and (dom-attr o 'href)
                    (string-match "https://\\(www\\.\\)?google\\.com/url\\?q=" (dom-attr o 'href)))
           (let* ((parsed (url-path-and-query
                           (url-generic-parse-url (dom-attr o 'href))))
                  (params (url-parse-query-string (cdr parsed))))
             (dom-set-attribute o 'href (car (assoc-default "q" params #'string=)))))
         (let ((text (string= (string-trim (dom-text o)) "")))
           (when (string= text "")
             (setf (car o) 'span))))
       (when (and
              (string-match "font-weight:700" (or (dom-attr o 'style) ""))
              (not (string-match "font-style:normal" (or (dom-attr o 'style) ""))))
         (setf (car o) 'strong))
       (when (dom-attr o 'style)
         (dom-remove-attribute o 'style))))
    ;; bold text is actually represented as font-weight:700 instead
    (with-temp-buffer
      (svg-print dom)
      (buffer-string))))

(defun my-google-doc-org (doc-id)
  "Return DOC-ID in Org Mode format."
  (pandoc-convert-stdio (my-google-doc-clean-html (my-google-doc-html doc-id)) "html" "org"))

I have lots of sections in that document, including past journal entries, so I want to get a specific section by name.

(defun my-org-get-subtree-by-name (org-text heading-name)
  "Return ORG-TEXT subtree for HEADING-NAME."
  (with-temp-buffer
    (insert org-text)
    (org-mode)
    (goto-char (point-min))
    (let ((org-trust-scanner-tags t))
      (car (delq nil
                 (org-map-entries
                  (lambda ()
                    (when (string= (org-entry-get (point) "ITEM") heading-name)
                      (buffer-substring (point) (org-end-of-subtree))))))))))

Now I can get the bolded words from a section of my notes, with just a sentence for context. I use pandoc to convert it to Org Mode syntax.

(defvar my-lang-words-for-review-context-function 'sentence-at-point)
(defvar my-lang-tutor-notes-url nil)
(defun my-lang-tutor-notes (section-name)
  (my-org-get-subtree-by-name
   (my-google-doc-org my-lang-tutor-notes-url)
   section-name))

(defun my-lang-words-for-review (section)
  "List the bolded words for review in SECTION."
  (let* ((section (my-lang-tutor-notes section))
         results)
    (with-temp-buffer
      (insert section)
      (org-mode)
      (goto-char (point-min))
      (org-map-entries
       (lambda ()
         (org-end-of-meta-data t)
         (while (re-search-forward "\\*[^* ].*?\\*" nil t)
           (cl-pushnew
            (replace-regexp-in-string
             "[ \n ]+" " "
             (funcall my-lang-words-for-review-context-function))
            results
            :test 'string=)))))
    (nreverse results)))

For example, when I run it on my notes on artificial intelligence, this is the list of bolded words and the sentences that contain them.

(my-lang-words-for-review "Sur l'intelligence artificielle")

• Je l'ai aussi utilisée pour faire des recherches.
• Je peux consacrer une petite partie de mon budget à des essais, mais je ne veux pas travailler davantage pour rentabiliser une dépense plus importante.
• Je n'ai pas le temps de concentration nécessaire pour justifier l'investissement dans mon propre matériel, et sinon, les progrès sont trop rapides pour m'engager dans une configuration spécifique.
• J'ai une conscience aiguë des limites cognitives ou physiques à cause des difficultés de santé de ma mère et de ma sœur, et de mes expériences avec mes limitations à cause du fait que je suis la personne principalement en charge de ma fille.
• Je lis très vite, mais je n'ai pas assez de patience pour les longs contenus vidéo ou audio.
• Je n'aime pas les textes qui contiennent beaucoup de remplissage.
• Beaucoup de gens ont une réaction forte contre l'IA pour plusieurs raisons qui incluent le battage médiatique excessif dont elle fait l'objet, son utilisation à mauvais escient, et l'inondation de banalité qu'elle produit.
• Je réécris souvent la majorité du logiciel à l'exception d'un ou deux morceaux parce que ce code ne me convient pas.
• Je ne veux pas l'utiliser pour les correctifs que je veux soumettre à d'autres projets parce que le code ne me semble pas correct et je ne veux pas gaspiller le temps d'autres bénévoles.
• J'aime pouvoir lui donner trois dépôts git et des instructions pour générer un logiciel à partir d'un dépôt pour un autre via le troisième dépôt.
• Mais je ne veux pas le publier avant de réécrire et tout comprendre.
• Sans l'IA, je pourrais peut-être apprendre plus lentement avec l'aide d'Internet, qui a beaucoup de ressources commehttps://vitrinelinguistique.oqlf.gouv.qc.ca/Vitrine linguistique.
• Je veux profiter davantage, apprendre davantage avec l'aide de vraies personnes, complétée par l'aide de l'IA.
• J'adore les sous-titres simultanés, mais je n'ai pas toujours trouvé une méthode ou un système qui me convienne.

I can then go into the WhisperX transcription JSON file and replay those parts for closer review.

I can also tweak the context function to give me less information. For example, to limit it to the containing phrase, I can do this:

(defun my-split-string-keep-delimiters (string delimiter)
  (when string
    (let (results pos)
      (with-temp-buffer
        (insert string)
        (goto-char (point-min))
        (setq pos (point-min))
        (while (re-search-forward delimiter nil t)
          (push (buffer-substring pos (match-beginning 0)) results)
          (setq pos (match-beginning 0)))
        (push (buffer-substring pos (point-max)) results)
        (nreverse results)))))

(ert-deftest my-split-string-keep-delimiters ()
 (should
  (equal (my-split-string-keep-delimiters
          "Beaucoup de gens ont une réaction forte contre l'IA pour plusieurs raisons qui *incluent* le battage médiatique excessif dont elle fait l'objet, son utilisation à mauvais escient, et *l'inondation de banalité* qu'elle produit."
          ", \\| que \\| qui \\| qu'ils? \\| qu'elles? \\| qu'on "
          )
 )))

(defun my-lang-words-for-review-phrase-context (&optional s)
  (setq s (replace-regexp-in-string " " " " (or s (sentence-at-point))))
  (string-join
   (seq-filter (lambda (s) (string-match "\\*" s))
               (my-split-string-keep-delimiters s ", \\| parce que \\| que \\| qui \\| qu'ils? \\| qu'elles? \\| qu'on \\| pour "))
   " ... "))

(ert-deftest my-lang-words-for-review-phrase-context ()
  (should
   (equal (my-lang-words-for-review-phrase-context
           "Je peux consacrer une petite partie de mon *budget* à des essais, mais je ne veux pas travailler davantage pour rentabiliser une dépense plus importante.")
          "Je peux consacrer une petite partie de mon *budget* à des essais")))

(let ((my-lang-words-for-review-context-function 'my-lang-words-for-review-phrase-context))
  (my-lang-words-for-review "Sur l'intelligence artificielle"))

• pour faire des recherches.
• Je peux consacrer une petite partie de mon budget à des essais
• , et sinon
• J'ai une conscience aiguë des limites cognitives ou physiques à cause des difficultés de santé de ma mère et de ma sœur
• pour les longs contenus vidéo ou audio.
• Je n'aime pas les textes qui contiennent beaucoup de remplissage.
• qui incluent le battage médiatique excessif dont elle fait l'objet … , et l'inondation de banalité
• Je réécris souvent la majorité du logiciel à l'exception d'un ou deux morceaux
• pour les correctifs … parce que le code ne me semble pas correct et je ne veux pas gaspiller le temps d'autres bénévoles.
• pour un autre via le troisième dépôt.
• Mais je ne veux pas le publier avant de réécrire et tout comprendre.
• , je pourrais peut-être apprendre plus lentement avec l'aide d'Internet
• , apprendre davantage avec l'aide de vraies personnes, complétée par l'aide de l'IA.
• qui me convienne.

Now that I have a function for retrieving the HTML or Org Mode for a section, I can use that to wdiff against my current text to more easily spot wording changes.

(defun my-lang-tutor-notes-wdiff-org ()
  (interactive)
  (let ((section (org-entry-get (point) "ITEM")))
    (my-wdiff-strings
     (replace-regexp-in-string
      " " " "
      (my-org-subtree-text-without-blocks))
     (replace-regexp-in-string
      " " " "
      (my-lang-tutor-notes section)))))

Related:

• my-wdiff-strings is in Wdiff
• my-org-subtree-text-without-blocks is in Counting words without blocks

Screenshot:

Nudged by org-daily-reflection (@emacsomancer's toot) and Jeremy Keith's post where he mentions his on this day page, I finally got around to making my own on this day page again. I use the 11ty static site generator, so it's static unless you have Javascript enabled. It might be good for bumping into things. I used to have an "On this day" widget back when I used Wordpress, which was fun to look at occasionally.

The code might be a little adamant about converting all the dates to America/Toronto:

export default class OnThisDay {
  data() {
    return {
      layout: 'layouts/base',
      permalink: '/blog/on-this-day/',
      title: 'On this day'
    };
  }

  async render(data) {
    const today = new Date(new Date().toLocaleString('en-US', { timeZone: 'America/Toronto' }));
    const options = { month: 'long', day: 'numeric' };
    const date = today.toLocaleDateString('en-US', options);
    const currentMonthDay = today.toISOString().substring(5, 10);
    let list = data.collections._posts
        .filter(post => {
          const postDateTime = new Date(post.date).toLocaleString('en-US', { timeZone: 'America/Toronto' });
          const postMonthDay = (new Date(postDateTime)).toISOString().substring(5, 10);
          return postMonthDay === currentMonthDay;
        })
        .sort((a, b) => {
          if (a.date < b.date) return 1;
          if (a.date > b.date) return -1;
          return 0;
        })
        .map(post => {
          const postDateTime = new Date(post.date).toLocaleString('en-US', { timeZone: 'America/Toronto' });
          const postDate = new Date(postDateTime);
          const postYear = postDate.getFullYear();
          return `<li>${postYear}: <a href="${post.url}">${post.data.title}</a></li>`;
        })
        .join('\n');
    list = list.length > 0
      ? `<ul>${list}</ul>`
      : `<p>No posts were written on ${date} in previous years.</p>`;

    return `<section><h2>On this day</h2>
<p>This page lists posts written on this day throughout the years. If you've enabled Javascript, it will show the current day. If you don't, it'll show the posts from the day I last updated this blog. You might also like to explore <a href="/blog/all">all posts</a>, <a href="/topic">a topic-based outline</a> or <a href="/blog/category">categories</a>.</p>
<h3 class="date">${date}</h3>
<div id="posts-container">${list}</div>

<script>
  $(document).ready(function() { onThisDay(); });
</script>
</section>`;
  }
};

function onThisDay() {
  const tz = 'America/Toronto';
  function getEffectiveDate() {
    const urlParams = new URLSearchParams(window.location.search);
    const dateParam = urlParams.get('date');
    if (dateParam && /^\d{2}-\d{2}$/.test(dateParam)) {
      const currentYear = new Date().getFullYear();
      const dateObj = new Date(`${currentYear}-${dateParam}T12:00:00Z`);
      if (dateObj.getTime()) {
        return {
          monthDay: dateParam,
          formatted: dateObj.toLocaleDateString('en-US', { month: 'long', day: 'numeric' })
        };
      }
    }
    const today = new Date(new Date().toLocaleString('en-US', { timeZone: tz }));
    return {
      monthDay: today.toISOString().substring(5, 10), // MM-DD
      formatted: today.toLocaleDateString('en-US', { month: 'long', day: 'numeric' })
    };
  }
  // Fetch and process the posts
  fetch('/blog/all/index.json')
    .then(response => response.json())
    .then(posts => {
      const dateInfo = getEffectiveDate();
      const dateElement = document.querySelector('h3.date');
      if (dateElement) {
        dateElement.textContent = dateInfo.formatted;
      }
      const matchingPosts = posts.filter(post => {
        const postDate = new Date(post.date).toLocaleString('en-US', { timeZone: tz });
        const postMonthDay = (new Date(postDate)).toISOString().substring(5, 10);
        return postMonthDay === dateInfo.monthDay;
      });

      matchingPosts.sort((a, b) => {
        const dateA = new Date(a.date);
        const dateB = new Date(b.date);
        return dateB - dateA;
      });

      const elem = document.getElementById('posts-container');
      if (matchingPosts.length > 0) {
        const postsHTML = matchingPosts.map(post => {
          const postDate = new Date(post.date).toLocaleString('en-US', { timeZone: tz });
          const postYear = new Date(postDate).getFullYear();
          return `<li>${postYear}: <a href="${post.permalink}">${post.title}</a></li>`;
        }).join('\n');
        elem.innerHTML = `<ul>${postsHTML}</ul>`;
      } else {
        elem.innerHTML = `<p>No posts were written on ${dateInfo.formatted}.</p>`;
      }
    })
    .catch(error => {
      console.error('Error fetching posts:', error);
    });
}

I used to include the day's posts as a footer on the individual blog post page. That might be something to consider again.

I write my blog posts in Org Mode and export them to Eleventy with ox-11ty, which is derived from the ox-html backend.

Sometimes I want to link to something in a different blog post. This lets me build on thoughts that are part of a post instead of being a whole post on their own.

If I haven't added an anchor to the blog post yet, I can add one so that I can link to that section. For really old posts where I don't have an Org source file, I can edit the HTML file directly and add an id="some-id" so that I can link to it with /url/to/post#some-id. Most of my new posts have Org source, though. I have a my-blog-edit-org function and a my-blog-edit-html function in my Emacs configuration to make it easier to jump to the Org file or HTML for a blog post.

If the section has a heading, then it's easy to make that linkable with a custom name. I can use org-set-property to set the CUSTOM_ID property to the anchor name. For example, this voice access section has a heading that has CUSTOM_ID, as you can see in the . If I don't mind having long anchor names, I can use the my-assign-custom-ids function from my config to automatically set them based on the outline path.

(defun my-assign-custom-ids ()
  (interactive)
  (let ((custom-ids
         (org-map-entries (lambda () (org-entry-get (point) "CUSTOM_ID")) "CUSTOM_ID={.}")))
    (org-map-entries
     (lambda ()
       (let ((slug
              (replace-regexp-in-string
               "^-\\|-$" ""
               (replace-regexp-in-string "[^A-Za-z0-9]+" "-"
                                         (downcase (string-join (org-get-outline-path t) " "))))))
         (while (member slug custom-ids)
           (setq slug (read-string "Manually set custom ID: ")))
         (org-entry-put (point) "CUSTOM_ID" slug)))
     "-CUSTOM_ID={.}")))

  #+ATTR_HTML: :id interest-development
  That reminds me a little of another reflection
  I've been noodling around on interest development...

/* Add link icons to headings and anchored paragraphs */
function addLinkIcons() {
  document.querySelectorAll('h1[id], h2[id], h3[id], p[id]').forEach((o) => {
    const link = document.createElement('a');
    const article = o.closest('article[data-url]');
    link.href = window.location.origin + (article?.getAttribute('data-url') || window.location.pathname) + '#' + o.getAttribute('id');
    link.innerHTML = '🔗';  // link icon
    link.title = 'anchor';
    link.classList.add('anchor-icon');
    link.addEventListener('click', copyLink);
    o.prepend(link);
  });
}

addLinkIcons();

.entry h2, .entry h3, .entry h4, .entry p, .content h2, .content h3, .content h4, .content p { position: relative }
.content a.anchor-icon:link, .entry a.anchor-icon:link { font-size: 60%; text-decoration: none !important; font-size: small; float: right }
@media only screen and (min-width: 95rem) {
   .content a.anchor-icon:link, .entry a.anchor-icon:link { display: block; position: absolute; left: -2em; top: 0.2em; float: none}
}

Sometimes automatic subtitle timing tools like Aeneas can get confused by silences, extraneous sounds, filler words, mis-starts, and text that I've edited out of the raw captions for easier readability. It's good to quickly check each caption. I used to listen to captions at 1.5x speed, watching carefully as each caption displayed. This took a fair bit of time and focus, so… it usually didn't happen. Sampling the first second of each caption is faster and requires a little less attention.

(defvar my-subed-skim-msecs 1000 "Number of milliseconds to play when skimming.")
(defun my-subed-skim-starts ()
  (interactive)
  (subed-mpv-unpause)
  (subed-disable-loop-over-current-subtitle)
  (catch 'done
    (while (not (eobp))
      (subed-mpv-jump-to-current-subtitle)
      (let ((ch
             (read-char "(q)uit? " nil (/ my-subed-skim-msecs 1000.0))))
        (when ch
          (throw 'done t)))
      (subed-forward-subtitle-time-start)
      (when (and subed-waveform-minor-mode
                 (not subed-waveform-show-all))
        (subed-waveform-refresh))
      (recenter)))
  (subed-mpv-pause))

function getVisibleVideo() {
  const videos = document.querySelectorAll('video');
  for (const video of videos) {
    const rect = video.getBoundingClientRect();
    if (
      rect.top >= 0 &&
      rect.left >= 0 &&
      rect.bottom <= (window.innerHeight || document.documentElement.clientHeight) &&
      rect.right <= (window.innerWidth || document.documentElement.clientWidth)
    ) {
      return video;
    }
  }
  return null;
}

async function skimVideo(video=getVisibleVideo(), msecs=1000) {
  // Get the first text track (assumed to be captions/subtitles)
  const textTrack = video.textTracks[0];
  if (!textTrack) return;
  const remaining = [...textTrack.cues].filter((cue) => cue.endTime >= video.currentTime);
  video.play();
  // Play the first 1 second of each visible subtitle
  for (let i = 0; i < remaining.length && !video.paused; i++) {
    video.currentTime = remaining[i].startTime;
    await new Promise((resolve) => setTimeout(resolve, msecs));
  }
}

function handleSkimButton(event) {
   const vid = event.target.closest('.vid').querySelector('video');
   skimVideo(vid);
 }

document.querySelectorAll('video').forEach((vid) => {
   const div = document.createElement('div');
   const skim = document.createElement('button');
   skim.textContent = 'Skim';
   div.appendChild(skim);
   vid.parentNode.insertBefore(div, vid.nextSibling);
   skim.addEventListener('click', handleSkimButton);
});

(-let* ((files (directory-files "~/proj/emacsconf/2024/cache" t "--main\\.vtt"))
        ((count-subs sum-seconds)
         (-unzip (mapcar
                  (lambda (file)
                    (list
                     (length (subed-parse-file file))
                     (/ (compile-media-get-file-duration-ms
                         (concat (file-name-sans-extension file) ".webm")) 1000.0)))
                  files)))
        (total-seconds (-reduce #'+ sum-seconds))
        (total-subs (-reduce #'+ count-subs)))
  (format "%d files, %.1f hours, %d total captions, speed up of %.1f"
          (length files)
          (/ total-seconds 3600.0)
          total-subs
          (/ total-seconds total-subs)))

I'd like to write about code more often. It's easier for people to try out ideas if they can copy the code without fiddling with selecting the text, especially on mobile browsers, so "Copy code" buttons on source code blocks would be nice. I used this tutorial for adding code buttons as a basis for the following CSS and JS code.

First, let's add the buttons with Javascript. I want the buttons to be visible in the summary line if I'm using the <details /> element. If not, they can go in the div with the org-src-container class.

/* Start of copy code */
// based on https://www.roboleary.net/2022/01/13/copy-code-to-clipboard-blog.html
const copyLabel = 'Copy code';

async function copyCode(block, button) {
  let code = block.querySelector('pre.src');
  let text = code.innerText;
  await navigator.clipboard.writeText(text);
  button.innerText = 'Copied';
  setTimeout(() => {
    button.innerText = copyLabel;
  }, 500);
}

function addCopyCodeButtons() {
  if (!navigator.clipboard) return;
  let blocks = document.querySelectorAll('.org-src-container');
  blocks.forEach((block) => {
    let button = document.createElement('button');
    button.innerText = copyLabel;
    button.classList.add('copy-code');
    let details = block.closest('details');
    let summary = details && details.querySelector('summary');
    if (summary) {
      summary.appendChild(button);
    } else {
      block.appendChild(button);
    }
    button.addEventListener('click', async() => {
      await copyCode(block, button);
    });
    block.setAttribute('tabindex', 0);
  });
}
document.addEventListener("DOMContentLoaded", function(event) { 
  addCopyCodeButtons();
});
/* End of copy code */

Then we style it:

/* Start of copy code */
pre.src { margin: 0 }
.org-src-container {
    position: relative;
    margin: 0 0;
    padding: 1.75rem 0 1.75rem 1rem;
}
summary { position: relative; }
summary .org-src-container { padding: 0 }
summary .org-src-container pre.src { margin: 0 }
.org-src-container button.copy-code, summary button.copy-code {
    position: absolute;
    top: 0px;
    right: 0px;
}
/* End of copy code */

Someday I'll figure out how to make it easier to tangle things to the post's directory and make the file available for download. In the meantime, this might be a good start.

Now that I have a decent workflow for writing, coding would be the next good thing to reintegrate into my life.

I get about an hour or two of discretionary time a day, late at night once A-‘s asleep. It’s not a solid chunk, since A- often wants to nurse, but I can usually get back to what I was doing without losing too much context. Sometimes A- takes a while to settle down, or wakes up midway. Sometimes I’m too sleepy to do much after A- goes to sleep. Still, I usually get a little time to update my journal, do some consulting, or even play video games with my husband.

How does coding fit into the picture? It’s fun. I learn stuff. Sometimes I even build tools that make my life a little easier. It gives me non-baby things to talk about with W- and other people, too.

The time needs to come from somewhere. What are the trade-offs I might make?

• Fewer drawings of non-journal thoughts, balanced by more writing time on phone. Can I figure out a good workflow for drawing on my phone? Not index cards, but maybe I can move my drawing practice to my phone for extra skill-building and mental variety.
• Less consulting, but more personal benefits to code; might also use this to expand my comfortable range for consulting
• Real-life kaizen vs virtual kaizen: shift by doing real-life kaizen while A- is awake
• Other tasks: still do as needed

What could a good setup be like?

• I spend some reading time going through documentation, Q&A, research, etc. This helps me improve my skills and work more efficiently.
• I have a dev environment set up for risk-free experimentation.
• I have a to-do list with prioritized ideas and notes.
• I work on tasks that might be 15-30m in size, ideally with tests.

I think it’ll be worth learning how to properly set things up with Vagrant. Frequent rebuilds will force me to make sure all my dev environment assumptions are documented.

It’ll also be worth cleaning up my technical notes and writing more things down, so I can get back up to speed after months or even years away.

Then I’ll want to sort out my testing environment and get back to writing tests. I wonder if I can set things up so that I can even write tests on my phone. Maybe cucumber cases? It’ll be easier to write behaviour-driven tests than regular tests, since I don’t have to mess with punctuation.

Then I can code, one small chunk at a time. Maybe I can even write pseudo code on my phone.

I’d also like to get back to tweaking my environment and tools, since that pays off multiple ways for me: enjoyment, learning, efficiency, and notes to share.

I can start by sorting out my dev environment and notes. We’ll see how that goes, or if this is something that will be mostly on the back burner until A- grows a little more. =)

While trying out the “after” filter I just added to my blog, I noticed that my paragraph breaks were missing. I hadn't noticed it for a while because I've been building up my weekly and monthly reviews from sketches instead of blog posts. How embarrassing!

(Then A- woke up and it was time for lunch, so I was a bit frazzled. But W- stepped in and took care of her, hooray!)

I saw the paragraph breaks in WordPress' visual editor, but not the exported HTML, which just kept whitespace in between the paragraphs instead of breaking them up with tags. It happened even when I created a new post through the web interface, so it wasn't org2blog's fault.

I checked if the paragraph issue happened on a new install. It didn't.

I checked if the paragraph issue happened with all the plugins deactivated. It didn't. Aha! (Note to self: I really should set up a dev environment again…)

I turned the plugins on one by one, and I narrowed it down to the NextGen Gallery plugin. It worked after I updated that.

Anyway, things should be readable again. Hooray!