#dev 2023-04-06
2023-04-06 UTC
geoffo joined the channel
geoffo, superkuh and jamietanna joined the channel
# 
jamietanna gRegor I was literally just coming here to report the same - https://telegraph.p3k.io/webmention/17TgVhAJG3i1wtZa45/details is what I get back from BF when Webmentioning it
# jamietanna Also `Enqueued task projects/bridgy-federated/locations/us-central1/queues/webmention/tasks/7736403492543298562 to deliver to followers...`
# jamietanna But I'm seeing no notes of mine being published for a few days
# jamietanna But _am_ seeing listens from a couple of days ago
# jamietanna Not sure what's changed on the BF side?
jonnybarnes, gRegor and [jamietanna] joined the channel
# [jamietanna] And GWG yep will have a look this weekend
IWSlackGateway, gRegor and [snarfed] joined the channel
[KevinMarks] joined the channel
# 
[KevinMarks] A scraper answer to account mapping https://whosum.com/
[tantek], geoffo and gRegor joined the channel
[jacky] and [jamietanna] joined the channel
# [jamietanna] Thanks snarfed! Tbh I'll probably just leave the posts as-is and wait for new posts to get syndicated out
[schmarty] joined the channel
# 
gRegor https://fed.brid.gy/user/gregorlove.com shows "gRegor Morrill None" instead of "gRegor Morrill posted", but the log appears ok otherwise
# 
[snarfed] gRegor I've cleared https://gregorlove.com/2023/04/wawg-distribution/, feel free to retry
geoffo, [campegg], [chrisbergr] and [capjamesg] joined the channel
# 
[capjamesg] Iād love to hear more [tantek].
# 
Loqi It looks like we don't have a page for "ARIA" yet. Would you like to create it? (Or just say "ARIA is ____", a sentence describing the term)
# bkil [chrisbergr]: Yeah, I think SEO doesn't really understand footnote links that well. At least it would cause me quite some headache if I were to implement a scraper & crawler. I don't see how descriptive footnote links could help here and if you copy & pasted it between the two sites it would probably hurt your rating.
# bkil capjamesg: It's not about _following_ the link per se. It's attributing links to (1) contextual content within its occurrence (2) attributing a title to it via the link description.
# bkil Algorithms should be specifically fine tuned if they wanted to inherit this context transitively.
# 
[chrisbergr] @bkil The engines need context for the hyperlinks. This is the reason why links with just the url or "here" as link text are irrelevant. So if you write "Jame's post about his implementation of footnots" in your footnote link, this will help james. If you don't use nofollow.
# bkil By the way, another implementation specific detail about footnotes. I see many are lacking just 1 line of CSS to highlight the footnote that I have just clicked on and takes me quite some time to find it on the page. This is especially problematic if you have a huge screen and the footnote is close to the edge, as browsers will only scroll the top of the screen to the anchor.
# bkil See here for some example code: https://github.com/indieweb/chat.indieweb.org/issues/58 https://developer.mozilla.org/en-US/docs/Web/CSS/:target
# bkil [chrisbergr]: Yes, that's half of what I mean. The other half you will miss is _where_ the given phrase (and link) occurs within the paragraph. Better search engines have been considering that for decades. Some even support NEAR-style queries with such knowledge.
# [chrisbergr] @bkil then maybe I don't get your point. Why exactly would the rating got hurt by using footnotes with descriptive linktexts?
# bkil [chrisbergr]: Could you perhaps share a URL for me to see? I think we are confusing terms here.
# bkil Oh, and my previous message about CSS opens another can of worms. Some web developers fall into the trap of being satisfied with putting an ID on "something" as long as clicking on it scrolls to that part of the page. However, little do they think about whether what really should be holding the ID, the heading, the body, a given sentence or `<section>` perhaps?
# [chrisbergr] <a name="something"/> right before the element I'd like to be seen
# [chrisbergr] I don't have an explicit url for you, I don't use footnotes š But let's say wikipedia for example. Their footnotes are pretty good https://en.wikipedia.org/wiki/Psychopathy#References
# bkil They follow a good example according to my CSS woes: they place the ID on the `<li>` that contains the footnote and even change the color of its background when it is targeted in the URL!
# bkil But now that you mention it, I added some userCSS for it: li:target
{ border-left: 1px solid; padding-left: 0.5em; }# bkil Some may find this too much: .references li:target
{ border-left: 1px solid; padding-left: 0.5em; } .reference:target a { font-weight: bolder; background-color: black; color: white; padding: 2px; }# bkil So about SEO. Basically if a given sentence or coupled short paragraph links to a certain external article, that also tells the search engine something about the referenced article. We usually only make a few words or a short phrase clickable for the URL itself, hence why the engine must evaluate its context as well.
# bkil How should such context be attributed if the link target within the source code is an internal link? This can be especially messy if in case of Wikipedia, the footnote itself contains valuable novel information about the linked article, such as full citation details about the publication as per the scientific method. Actually, both the citation and the context where the footnote is referenced can tell something about the linked article and its connection with
# bkil the given webpage, so neither should be discarded.
# [chrisbergr] Ok, got it. But it won't downgrade your rating, it just won't raise it
# bkil I don't see straightforward algorithms through which you could detect and connect external links to be attributed the same edge weight and metadata context between indirect footnote-linking and inline linking. Unless you (1) use an AI (2) write a bunch of heuristics and parsers for most known CMS and libraries and only handle their output.
# bkil The outward pointing links from your website do not contribute (noticeably) for most search engines. It is the inbound links and their context that count more. One who is experimenting with link indirection are doing the network as a whole a disservice. Mind you this is not a big disservice, only a nuance, but I hold that every little counts.
# bkil I.e., many new, small independent search engines just process documents and their links (or link texts) in total isolation and as a single big unit as they don't have advanced contextual cues implemented. Go figure why their results are... "different"