Research:Newsletter/Production process

These notes document the tasks involved in the production and distribution of each new issue of the research newsletter every month. Names in brackets indicate members (or former members) of the editorial team who have experience with carrying out the corresponding step. Help is welcome; leave a comment at Research_talk:Newsletter or contact the editorial team directly if interested.

Newsletter content preparation[edit]

During the month[edit]

Find new publications to cover, via Google Scholar alerts, Tweetdeck (for @wikiresearch), or tips received via Twitter DMs etc. (Miriam, Tilman, formerly: Dario)
Tweet/retweet from @wikiresearch as appropriate (Miriam, Tilman, formerly: Dario). NB: The texts of all tweets authored by Tilman or Miriam (excluding retweets) are licensed under CC BY-SA 3.0.
Import tweeted publications into Zotero and the Etherpad todo list, via this spreadsheet that automatically updates with new tweets (Dnshitobu, Masssly, Tilman, formerly: Dario). There is also a screencast walking through most of these steps:
- Go to a spreadsheet entry (tweet) that is not yet marked as processed in the "Zotero" and "Etherpad" columns
- Decide whether the tweet's link(s) should be imported, i.e. whether they concern a research publication that's in scope for the research newsletter. (Generally, everything that has a DOI or is peer-reviewed should be imported. But many tweets - e.g. those announcing a new edition of the newsletter itself - should be left out. It's also good to check for duplicates, i.e. whether the same paper has already been tweeted/imported before. See the "notes" tab in the spreadsheet for further remarks.)
- If it should not be imported: Mark the tweet's entry accordingly, using the dropdown selectors in the "Zotero" and "Etherpad" columns
- If it should be imported:
  - Import the publication linked in the tweet into Zotero, e.g. using the Zotero Connector browser extension (documentation) or via https://www.zotero.org/save . More specifically, into the Zotero web library (not, or not only, into the desktop client; see also explainer).
  - Check the imported item in the Zotero web library; it can be found by sorting entries by "date added" and looking at the most recent entry. (If there is no "date added" column, click the column selector icon on top right to include it. Sometimes if the import via Zotero Connector fails, or results in an item that is lacking certain fields like the authors' names, it can be fixed by retrying.)
  - If the item looks OK, mark the tweet's entry in the spreadsheet as "Done" in the "Zotero" column, using the dropdown selector. If the automated Zotero import fails, consider importing the paper manually or leaving a "Zotero fail" comment in the "Notes" column.
  - Export the paper's resulting entry from Zotero into the Etherpad todo list for the current month, using the "Wikipedia Citation Templates" export format. In the Zotero web library, this is done by clicking the list entry (so that it is highlighted), and then the export icon above the list. This downloads a small text file whose content should be copypasted into the Etherpad. New entries go on top.
  - For paywalled papers (or other publications like printed book chapter that are not freely available online), add {{closed access}} right after the citation on the Etherpad. In case there is an alternative freely available version (e.g. on arXiv as a preprint, on researchgate.net or on an author's home page), try to list it too on the Etherpad
  - In case the tweet contains supplementary links, add them to the Etherpad too (no need to use Wikipedia citation templates). These can include replication code, datasets, slide decks, blog posts or press releases accompanying the paper.
  - Optional: Add an excerpt of the publication's abstract to the entry on the Etherpad, focusing on the parts that are the most relevant and interesting for our audience. (E.g. leave out generalities about "Wikipedia, a volunteer-written encyclopedia" or such.) This helps potential reviewers who read through the Etherpad decide which publications they may be interested in covering. Also, we use these excerpts as a fallback later in case a paper won't get a full writeup and instead goes into the "Other recent publications" section (see below).
  - Optional: Check whether the publication is released under an open access license that allows its images to be uploaded to Wikimedia Commons (normally a CC BY or CC BY-SA license). If yes, check whether the publication contains images that might be suitable illustrations for the newsletter. In that case, add a note to the Etherpad like "CC BY, has illustrations".
  - Mark the tweet's entry in the spreadsheet as "Done" in the "Etherpad" column, using the dropdown selector.

About one week before the planned publication date[edit]

(Publication usually occurs on the last weekend of the month, tied to the Signpost's publication schedule. See the Signpost newsroom page for the most up-to-date information.)

Send invitations to potential reviewers (Masssly, Tilman, formerly: Dario)
Add note about the upcoming issue in the Signpost newsroom page (example) (Tilman)
Set up draft page for the new issue, using the standard layout (example), at w:Wikipedia:Wikipedia Signpost/Next issue/Recent research (Tilman and other Signpost contributors)

Up to the Signpost's publication time:

Curate incoming reviews, including copyedits and formatting: ...
Fill the "other recent publications" section (example) with items from the Etherpad that look unlikely to be read and reviewed soon, optionally quoting from their abstract (Tilman)
Add per-review bylines, in addition to the general bylines on top (Tilman, Signpost editors, Masssly, formerly: Hanteng, Heather)
Optional: Add illustrations, in particular by extracting images and figures from those covered papers that are under an open access license which is compatible with Wikimedia Commons (normally CC BY or CC BY-SA).

After publication of the Signpost version[edit]

Prepare the version published here on Meta (Sir Amugi, Masssly, Tilman, formerly: Dario). The necessary steps can be gleaned from these edits):

Import the page from enwiki via Special:Import (needs user rights). Check "Copy all history revisions for this page", but do not check "Include all templates". As comment, one can use something like "for syndication as Wikimedia Research Newsletter.
Adjust wikilinks, using regex search and replace (e.g. [[Wikipedia:Verifiability|verifiability]] --> [[:w:Wikipedia:Verifiability|verifiability]])
Fix images if necessary (the Signpost sometimes uses a special layout template for them)
Fix citations if necessary: All used citation templates should be present on Meta too. Some individual template parameters may not be supported for the time being and can be removed (see e.g. Template talk:Cite_arXiv). In case a template doesn't work at all after the import, consider replacing it with the output obtained by pasting the original wikitext into en:Special:ExpandTemplates.
Add standard WRN header and footer (these can copypasted from previous issues, but don't forget to update template parameters)
On the talk page, add a link to the discussion page of the Signpost version (example)
Update Template:WRN current and Research:Newsletter/Archives

Newsletter publication[edit]

Publish and announce in other channels:

Tweet the Meta-wiki version from @wikiresearch, pinning the tweet (Tilman, formerly: Dario).
Consider posting it manually from the WikiResearch Facebook page (delete the post mirrored from Twitter in that case), pinning the post
Send announcement to the Wikiresearch-l mailing list example (Masssly, Tilman, formerly: Dario)
Send out HTML email version to subscribers (Tilman):
- Start a draft email in Gmail
- Add subject line (e.g. "Wikimedia Research Newsletter, March 2024")
- At the bottom, add the following boilerplate text (centered and in small font, replacing your email signature if you use one):
  "Thanks for subscribing to the HTML email distribution of the Wikimedia Research Newsletter! Click "Email" above to manage your subscription. To be able to view the illustrations, remote loading of images must be enabled in your email client. Images are only loaded from Wikimedia Foundation servers and are not used for tracking purposes. Feedback about this service is welcome at wikiresearch-contactgooglegroups.com"
- Open the mobile Meta-wiki version (example)
- Click on the "Contents" box to expand TOC
- Select an area around the content (i.e. try avoiding the MediaWiki header and footer)
- Copypaste into the Gmail draft (above the footer)
- In the Gmail editor, manually remove content above the WRN header and below the WRN footer, leaving divs intact
- Check for formatting issues (e.g. <math>...</math> code might not have survived the copypaste)
- Send the result to research-newsletterlists.wikimedia.org
- Log into [1] and approve your mail

Post-publication[edit]

Notify authors of reviewed papers (Musahfm, Masssly, Tilman): ....
Update stats... (old spreadsheet / PAWS notebook)
Set up new Etherpad to coordinate reviewers' work on the upcoming issue (it is linked from Research:Newsletter#How_to_contribute, the link there should update automatically). (Tilman)
- Copy over the previous month's Etherpad, updating dates (Tilman)
- Delete all entries that were covered in the just published issue (Tilman)

(Optional, not yet done for more recent years:)

Prepare Zotero dump of all papers covered in that year's volume, and a PDF book with all 12 issues (Dario, Tilman): ...