Pubmed RSS feed the summary only displays the conclusion (end of the summary)

hello

a few weeks ago i created a new feed. The summary display only contains the end (the conclusion, when the summary is structured).
My other feeds from pubmed (created for several years) are OK, the summary display is good
I tested the display of my new feed on thundebird: the summary is very well displayed
I created another feed from another base (medrxiv.org) the display of the summaries is good
I deduce that the writing of the feed by pubmed has changed and that zotero does not correctly translate the display of summaries

Eric
I am dyslexic and I do not master English writing well enough, I must trust the translators
  • You would have to share the feed URL in question.
  • for all the feeds I create on pubmed and whatever the theme, zotero only displays the end of the summary
    I'm under ubuntu
    I also tested several streams on another computer under ubuntu, the problem is also present
    Currently, I cannot do a test on a Windows computer

  • I specify the symptoms
    if the summary is structured then zotero displays the conclusion
    if the summary is not structured then zotero displays the start of the summary and ends with ",..."
  • That's not what's happening.

    RSS feeds can have a "description" element and a "content" element. Zotero displays the "description" in the Abstract field in the feed view.

    PubMed seems to sometimes include the "description" text within the "content" element as well, but that's irrelevant ā€” that's not what Zotero is using.

    (The correct usage of "description" vs. "content" is complicated, and depends somewhat on whether the contents are plain text or HTML. See here for an in-depth discussion of it. But regardless, "discussion" is what Zotero shows in Abstract, which is a plain-text field.)
  • Thank you for your answer

    Unfortunately I just found out that the problem is more serious
    Before, Pubmed did not use the "content" element, the abstract was in the "description" element.

    before :
    https://eutils.ncbi.nlm.nih.gov/entrez/eutils/erss.cgi?rss_guid=1LgINZmdEb7dTigG4QJac2AOT0ivyLK27StNI0TS1Okojj5cyO

    In recent weeks, Pubmed has introduced the element "content" in these feeds. the abstract is put in "content" and they have to choose to put the conclusion (when they can identify it) in "description" otherwise it puts the beginning of the abstract.

    Now :
    https://pubmed.ncbi.nlm.nih.gov/rss/search/1toH0ZWNzwdwzQ0_FHIiRPTFsFZUeT0UYUXJCWqtL1hhfXBstP/?limit=15&utm_campaign=pubmed-2&fc=20201228001145


    In theory, I find this approach more consistent with the functions of the "description" and "content" elements.

    The fact that zotero displays "description" will severely handicap pubmed's flow management via zotero. This will affect many people who will have to find another solution to manage pubmed flows. It is essential to have the summary to sort it in the health RSS feeds, because of the very (far too) high number of publications and the very low percentage retained

    I suggest adding an option to the zotero feed parameters: display "content" yes / no
  • That frankly sounds like inappropriate use of the description field by PubMed. Iā€™d recommend that you contact them and encourage them to revert that change. This would negatively impact all RSS readers, not only Zotero.
  • edited January 10, 2021
    Please do contact them. You can use the "Comment" button at the bottom of a PubMed search results page. I already sent my request about this and received a reply that this was done intentionally "for consistency" across "other feeds". I haven't yet replied to their reply but I'll both ask for more details about what they mean and point out that their new practice isn't consistent with RSS feeds from most journal publishers.

    I didn't mention Zotero in my comment and I urge you not to mention Zotero. Including Zotero will be a distraction from this issue being their responsibility.

    The PubMed tech support _are_ quite responsive to questions and requests. They _will_ make changes if they like the idea. For example, they brought back the "order results by journal" option after I requested it. They asked me for my use-case and I explained my need. Within a few days they sent me a message that the journal sort had been restored. A month later I received a follow-up message informing me of the proportion of searches displayed in journal-order and thanking me for my request. Clearly, the tech folks at NLM like their jobs and are willing, even eager, to improve the site's UI, etc.
  • 'description' is poorly specified ā€” it can be "the full content, a summary or some other form of excerpt at the publisher's discretion". But 'content:encoded' is meant to include the full-text HTML content, so putting the abstract there, even as HTML, does seem wrong. And if you put the full text into 'content:encoded', it seems much more appropriate to put the abstract, not some extracted conclusion sentence, into 'description'.

    So yes, it would be good if they would revert this. This isn't something we can fix without a clumsy feed-specific preference, since our goal really is to display the abstract, not the full content.

    Reference: https://www.rssboard.org/rss-profile#namespace-elements-content-encoded
  • @dstillman Yes. Please do not take the time to make changes to Zotero to accommodate this. I feel sure that PubMed will revert if they receive enough comments. You might consider putting what you wrote above into a comment for the NLM. Somebody gave them the idea that the new way to organize their feeds was a good idea. I don't think that it will take much to change their minds. A decade ago I was on first name friendship with several of the tech people there. They were wonderful people! Not any more of that closeness but it still feels like the team wants to do what is best.
Sign In or Register to comment.