9
votes
Can Bearblog links show authors?
Both Medium and Substack include the author name in the site title (next to the favicon). Bearblog hosted blogs seem to be frequently posted, especially in ~tech, and only showing that it's from bearblog doesn't give insight into authorship. (Authors are being added to tags, but this is not immediately visible on mobile clients until clicking into the comments.)
Is this something that can be updated? If it's a technical limitation of some sort, I respect that. It would just be a "nice to have" feature. :)
I believe this is an issue with Bearblog, and not with Tildes.
Substack adds the meta tag for authors like so:
Bearblog does not set that tag. I think if they did, it would likely get posted? I will test that in a bit.
Edit: Well, it's not the entirety of the issue, but it is the reason it's a non starter.
I checked it out with a couple of links, and Tildes only gets author information from that author tag. If you have the author tag set up properly, then it will get pulled in to meta information about that post. I did confirm that Bearblog doesn't set this - and you can see that it doesn't scrape the meta information on this bearblog topic. But meta information is gathered if it is presented; for example this test post picks up the meta information that I give.
However, that test post doesn't add the author to the topic URL area. To do that, I think that all that is required is a) that Bearblog make that adjustment and b) making an adjustment to
lib/site_info.py(see here) so that bearblog.dev is one of the sites that is configured to show the author.But as I said, it's kind of a non-starter because bearblog doesn't add this meta tag. To that end, I have not added it as an issue to the tracker, since it's not really something that can work (sorry about that).
Thinking about this further, what about the name of the blog? Discord is able to extract that in embeds, so I assume it's in the metadata somewhere.
At first I thought this might not be useful because a blog can be named anything, and the author could have a very different name.
However, this would be comparable to listing the channel name next to YouTube on those posts.
Bearblog does include the site_name and title OpenGraph meta tags so it's probably possible to scrape them, but I'm not sure if Tildes stores those items specifically. I would have expected it to grab the title and the word count for the meta information, but it's possible that there's something else preventing that from happening.
Tildes uses Embedly's Extract API behind the scenes to grab that information, you can actually try this out yourself here and see the results:
In the omeru.bearblog.dev example it doesn't grab any authors (presumably because it lacks any kind of author indication in the meta tags) but with your blog it does grab "Andrew Phoenix".
You can also see the other data from the
og:...and other meta tags Embedly can grab.I see what it can grab, but I was really wondering more about how it displays the information that it scrapes. For example, in the test I submitted it scrapes and displays Title, Author, Word Count. For bearblog sites, it seems to grab publish date and word count, but not title?
Also, maybe I'm crazy, but this afternoon I was pretty sure it didn't show either publish date or word count when I looked at this link but it's clearly there now, so maybe I'm just crazy.
Ahh right, so the actual API call to Embedly can be found here (this is also how it gets stored in the database) and then how that data actually gets used can be found here and in the HTML here.
Also this might be why it doesn't show the title for some cases.
I'm confused on how tildes grabs word count... unless I'm missing something, it's not present in embedly, and not referenced in that python file. (It's also not specified anywhere in the embedly API docs.)
I don't mind putting in a suggestion with bear, but if we're trying to grab the word count too, I'd like to understand how that is supposed to work.
I think it's just counting the words in the content, and it appears to do that just fine with bearblog, though I'm not sure when it scrapes the info.
"provider_name" does contain the blog name, though!
I tried this with another one too: https://embed.ly/docs/explore/extract?url=https%3A%2F%2Frango.bearblog.dev%2Fadults-ruin-everything-a-joke-mostly%2F
Send an email to Herman. If this is just a matter of adding a tag, he might do it.
He always answers me in a day or two, and I am kind of an annoying non-technical user.
EDIT: there is also a suggestion board and a GitHub.
As adding the new meta tag is a feature request, it should be posted on the suggestion board.
Source:
https://github.com/HermanMartinus/bearblog/?tab=readme-ov-file#no-javascript-no-stylesheets-no-trackers-just-your-words