#social 2018-03-25

2018-03-25 UTC
cdchapman, xmpp-social, fr33domlover, rigelk, jankusanagi_ and eprodrom joined the channel
#
eprodrom
strugee, I'm here too
#
eprodrom
All the chats
#
xmpp-social
[ajordan] Same!
#
eprodrom
Dangit
JanKusanagi and fr33domlover joined the channel
#
@HironobuSUZUKI
Authors: Christopher Lemmer Webber Jessica Tallon Erin Shepherd Amy Guy Evan Prodromou https://www.w3.org/TR/activitypub/
(twitter.com/_/status/977922940166705152)
#
fr33domlover
Hi everyone
#
fr33domlover
aaronpk, cool!
#
fr33domlover
If I have an actor that is managed by software, not a person,
#
fr33domlover
Should it have a "following" property that resolves to 404, or empty collection, or not have the property at all?
#
fr33domlover
(It doesn't follow anything, it just has an inbox you can POST and outbox you can GET, and a list of followers)
#
csarven
fr33domlover: no property > empty collection > 404
#
fr33domlover
csarven, so no property is the best?
#
csarven
I think so. To publish as well as for the consumer.
#
csarven
Most importantly you are not announcing some URI.. and the consumer doesn't have to follow-up on anything
#
csarven
empty collection is costly. 404 is just bleh..
#
csarven
open-world assumption - incomplete information is not wrong
#
fr33domlover
csarven, great points!! thank you so much for this input
#
fr33domlover
i have so many questions :p
#
csarven
and "pay as you go".. if/when "following" is useful, just add
#
csarven
You can extend the Agent/Application's description by saying that it has a "foaf:maker" or something.. or some concept along the lines of controller, maintainer if you want to connect hte dots
#
csarven
See also Description of a Project (DOAP) http://usefulinc.com/ns/doap if you want to describe the software further.
#
csarven
as to its abilities, documentation, repo etc..
#
fr33domlover
Thanks csarven
#
fr33domlover
I'm still at the very beginning of implementing AS2 and AP though ^_^
#
csarven
I use DOAP eg at https://dokie.li/ with info like which specs it implements.
#
csarven
http://csarven.ca/linked-specifications-reports describes how the LDN spec, its implementation reports, and software are all human and machine-interpretable. Each doing its own thing on the Web.
travismiller and jankusanagi_ joined the channel
#
fr33domlover
csarven, interesting! And dokieli looks really cool
#
fr33domlover
It's nice to see cool stuff done with these technologies
JanKusanagi joined the channel
#
fr33domlover
Has decentralized search been discussed?
#
fr33domlover
For example the option of maintaining a DHT for certain types of search
#
fr33domlover
Idk if it's good for plain text search, but maybe it's still good for something
#
saranix
for public data, standard web search engines are about as decentralized as it gets (anyone can run one), and for private data, usually you are searching meta which has federated, from within your own pod
#
fr33domlover
saranix, yeah I mean public data
#
fr33domlover
saranix, I mean something like this: Suppose some person is using a centralized sevice like faceb00k or twitter or linkedin or github or whatever, they want to switch to the fediverse but worried they won't be found - sort of like tags on mastodon are not federated so if you're on a small instance your stuff won't be noticed as much - what do we do and what do we tell this person
#
fr33domlover
To allow them to be really be found, their profile and their projects etc. whatever they publish
#
saranix
err uh, like I said, they are indexed by the search engines.
#
saranix
You know what isn't very well indexed? the silos.
#
saranix
when other pages mention your blog/whatever, it becomes more and more relevant
#
saranix
so the more decentralized sites where people are commenting on your stuff and sharing, the more linkages you have
#
saranix
this ranking does not occur with silos, because whenever anyone mentions a facebook page, what they are mentioning is all a single site called facebook.com. Individual pages don't get ranked as heavily as entire sites.
#
csarven
"You know what isn't very well indexed? the silos." -- can you expand on this, saranix?
#
saranix
you can't even view instagram through tor. How can it be indexed?
#
saranix
js is necessary to render the page
#
csarven
Is there a correlation?
#
csarven
re: silos and JS-requirement
#
saranix
@someone tweeting about @someoneelse, is 2 urls pointing inside twitter.com -- https://one.example/someone mentioning https://two.example/someoneelse is a bigger relevance boost
#
csarven
or do you just mean that URLs with JS-required content is not well indexed by some services?
#
saranix
I mean urls with js-requied content is not well indexed, that's one way that silos are at a disadvantage. The internal vs external linkages is the bigger factor.
#
saranix
You're much more likely to get ranked decentralized. On Facebook you are just one 2billion + pages on that site.
#
saranix
missing a couple of prepositions obviously. but you get it.
#
csarven
I'm not sure what you are getting at with ranking either, but the major search engines seem to do an okay job showing a relevant/accurate serp
#
csarven
Depends on the name that's searched of course.
#
saranix
Let's say you have a POSSE post on yourname.example. 10 twittizens share it (and let's assume the links are not hidden behind t.co). elsewhere on the web popularblog.example and news-site.example also mention your article, linking to yourname.example. ...
#
saranix
... a search engine might show one of the twitter links, and each of the other sites
#
saranix
with t.co, twitter might not even show up at all on some search engines
#
saranix
and that's if you are specifically searching for a page you know exists
#
saranix
if your search terms are more general, and it is the article which is relevant, the search engine will try it's best to find the canonical article. The one on your site.
#
saranix
thus, you will always be *better* found on decentralized than on silo
#
saranix
this is actually a very important point that I convey to businesses
#
saranix
and if you are a hub of one - i.e. yourname.example --- yourname.example is more relavent for 'yourname' search term, than somesite.example/u/yourname
#
csarven
Makes sense. Depends on the document/what's being searched. I think you were getting at any concept in a document. I initially thought about profile pages.
#
saranix
profile pages too
#
saranix
also remember that profile pages on hubzilla show up on hundreds of directory servers
#
saranix
other software will have of course different SEO
#
saranix
but yeah, if your name is actually *in the hostname* - that trumps all
#
saranix
almost always ranked #1 result
#
csarven
I'm not sure re profiles pages on "silos" not being indexed well on services with zillion users. I tend to see profile pages ranking well-enough. If you have stats on the contrary, I'll take your word for it.
#
saranix
profiles should be able to be found just about anywhere -- the whole point of a search engine is to find relevant results. If you are searching for a name, what's relavant is profiles.
#
csarven
That's what I'm trying to get to the bottom of when you say they (all?) are not indexed well. There is a bunch of stuff all compacted.. .is it the JS issue you want to raise? the backlinks to an URL under a "silo" service... having a personal website in comparison?..
#
saranix
I don't understand why you are obsessing so much with *just profiles* instead of the whole picture.
#
saranix
a lot of people don't even have profiles BTW.
#
csarven
I'm not.. just taking an category. First I don't quite understand what you consider as silo, then whether the issue is specific to JS-requirement and/or being part of a service with many other URLs.. and whether it was specific to a particular category of pages. It seemed like you are pulling all of that under a single explanation.
#
saranix
There's of course a placeholder profile page, but it's so irrelevant to the whole web
#
saranix
"It seemed like you are pulling all of that under a single explanation" - yes, I was giving a list of ways, not just one
#
saranix
and just like not all decentralized software is the same, neither is all silos. They of course exist on a scale
#
saranix
there is much more opportunity to affect your SEO if you are the site admin than if Zuckerberg is though. Can't change that fact.
#
csarven
I personally don't care about SEO or where I am in the SERPS, but I understand why others do or need to.
#
csarven
I used to care long ago.
#
csarven
For personal websites, I don't think there is much more to it then just following best/basic publishing practices.
#
csarven
The energy that could go into trying to figure out that tiny bit of extra optimisation isn't worth it I think.
#
saranix
For personal websites, so long as you are found at all, people will probably explore your whole site. That too is less likely to happen on a silo though, when it just feels like scrolling through one feed in time, and there are so many others calling for your attention.
#
saranix
really should put all this into a slidedeck. I keep meaning to but it's so much easier to just talk about it when it comes up.
#
csarven
The end destination varies. People jump around to find what they are looking for.
#
saranix
Right, and jumping *away* from your stuff and onto someone elses on the same site is kind of how they deliberately design their interactions to work on silos.