Got a proof-of-concept working of a pingback script that accepts
URLs (backlinks) of reply posts and checks the remote content for
a => link to one's capsule before appending the URL to the list of
received replies.

Maybe others want to expand on that for their own pleasure.
Cheers

On 2020-11-13, in the mailing list:
> If clients would send referers, servers could collect (cache) and
> present those to clients asking for links _to_ the current page
> ("backlinks" in wiki parlance), thus creating an interconnected gemini
> sphere.
>
> Although not simple, I would enjoy bi-directional linking in Gemini.
> --
> https://lists.sr.ht/~adnano/gemini/
%3C157076343ad69ee6833f16684b5a0c2f%40sdfeu.org%3E

--
```bash
#!/bin/bash
# URL of own capsule (grep string for mention)
capsule="gemini://gemini.capsule/~username/"
# Local file to append received reply URLs (chmod?)
replies="/home/username/public_gemini/replies.gmi"

function urldecode() { : "${*//+/ }"; echo -e "${_//%/\\x}"; }

if [[ -n $QUERY_STRING ]]; then
printf "20 text/gemini\r\n"
echo "# Thank you"

urlraw="$QUERY_STRING"
url=$(urldecode "$urlraw")
echo "Received the following reply URL:"
echo '```'
echo "$url"
echo '```'

domainpart=$(echo "$url"|awk -F/ '{print $3}')
echo "Trying to retrieve your reply content to check for the mention ..."
echo "Links back to my capsule found mentionend in your reply content:"
echo "$url" | openssl s_client -crlf -quiet -ign_eof -connect
"$domainpart:1965" \
| grep "^=> $capsule"

if [[ $? -eq 0 ]]; then
echo
echo "Adding your URL to my list of received replies, thanks!"
echo "=> $url $(date "+%F %T")" >> "$replies"
else
echo "Could not find a link back to my capsule (or had other issues)."
fi

else
printf "10 Please Enter Gemini URL of Your Reply\r\n"
fi

Plain Text wrote:

> Got a proof-of-concept working of a pingback script

I know very little about gemini, but that sounds like an early sign of building
into gemini the sort of crap that weighs down http/html and I though gemini was
meant to act as an escape from?

Plain Text <text@sdfeu.org> writes:

> Got a proof-of-concept working of a pingback script that accepts
> URLs (backlinks) of reply posts and checks the remote content for
> a => link to one's capsule before appending the URL to the list of
> received replies.
>
> Maybe others want to expand on that for their own pleasure.
> Cheers
>
> On 2020-11-13, in the mailing list:
>> If clients would send referers, servers could collect (cache) and
>> present those to clients asking for links _to_ the current page
>> ("backlinks" in wiki parlance), thus creating an interconnected gemini
>> sphere.
>>
>> Although not simple, I would enjoy bi-directional linking in Gemini.
>> --
>> https://lists.sr.ht/~adnano/gemini/
> %3C157076343ad69ee6833f16684b5a0c2f%40sdfeu.org%3E

Cool idea. Though I think something like this would be better served as
a separate protocol. I think it's cool to have a server that you run on
your machine which checks links in your pages and announces it P2P-style
and anyone else with the same link that also runs that same server can
ping back to you and that would give you a rough backlinking feature.

Of course, I'm not software engineer nor a computer programmer in that
regard. But I thought that's a neat idea after I've read your post.

--
|----Give them an inch and they will take a mile.----|
|----------------------------------------------------|
|---------{gopher,gemini}://kalayaan.xyz-------------|
|-C4AE 5D53 46A0 01DF 6E92 CB46 92D7 9FBB AB9F 3E37-|

Andy Burns <usenet@andyburns.uk> wrote:
> I know very little about gemini, but that sounds like an early sign of building
> into gemini the sort of crap that weighs down http/html and I though gemini was
> meant to act as an escape from?

Well, it is an optional feature that I suppose that you do not have to use if
you do not want to do. It also isn't a capability of the protocol or of the file
format; it is a separate program using them, as much as you can do like any other
CGI program can do.

(It is done without needing a Referer header. I do not want to add things like
the HTTP Referer header and that stuff, either, but the program that was written
does not use it; it uses only the existing format.)

However, I think that it is not needed, anyways. Back links isn't really
something that I need, and if you want to write articles with comments, then
the NNTP could be used. You could even cross-post, even if with Unusenet to
multiple servers with different newsgroups, and then set follow up to and can
follow up on your own server, and can still find the links.

(You could also make same messages exposed with multiple protocols, and can
also use "Content-type: text/gemini" in NNTP and still readable even in a
program that only understands plain text, too. This way, it is better than HTML.)

--
Don't laugh at the moon when it is day time in France.

On Sun, 27 Feb 2022 11:12:38 +0000, Andy Burns wrote:

> Plain Text wrote:
>
>> Got a proof-of-concept working of a pingback script
>
> I know very little about gemini, but that sounds like an early sign of
> building into gemini the sort of crap that weighs down http/html and I
> though gemini was meant to act as an escape from?

In this case, this is not something that clients will automatically do. If
the user has written a response to a Gemini page, and they notice the page
mentions a link to submit your response, they can go in and fill that
form. The script checks if the link actually exists in order to prevent
spam to some degree.

This is similar to a manual webmention [1][2].

[1]: https://indieweb.org/Webmention
[2]: https://www.w3.org/TR/webmention/

--
Leo

Hi,
Just seeing this thread now.

> On Sun, 27 Feb 2022 11:12:38 +0000, Andy Burns wrote:
> > Plain Text wrote:
> >> Got a proof-of-concept working of a pingback script
> > I know very little about gemini, but that sounds like an early sign of
> > building into gemini the sort of crap that weighs down http/html and I
> > though gemini was meant to act as an escape from?
>
> In this case, this is not something that clients will automatically do. If
> the user has written a response to a Gemini page, and they notice the page
> mentions a link to submit your response, they can go in and fill that
> form. The script checks if the link actually exists in order to prevent
> spam to some degree.
>
> This is similar to a manual webmention [1][2].
>
> [1]: https://indieweb.org/Webmention
> [2]: https://www.w3.org/TR/webmention/

For what is worth, I've worked and implemented a similar idea that I
think still stay aligned with Gemini idea:
gemini://gmi.bacardi55.io/gemlog/2022/02/27/my-take-on-gemlog-replies/

It isn't popular at all and most responses I got where negative but
the concept works (and is in place on my capsule) 🤷.

Cheers,
--
bacardi55

On Mon, 4 Apr 2022 10:32:42 +0200, bacardi55 wrote:
> It isn't popular at all and most responses I got where negative but
> the concept works (and is in place on my capsule) 🤷.

I remember seeing this a while back, probably via CAPCOM or Antenna.
One of the sentiments that struck me, which seems to be a common theme
these days is this bit from your post:

> A "simple" solution would be going to the capsule, find the contact
> of the author and send an email… But not every capsule has a contact
> page and the process is very manual, requiring to use another tool and
> protocol to communicate.

It seems to me that the whole point of application protocols like e-mail,
IRC, netnews, and Gemini itself is to accomplish a specific task. Yes, you
can extend most of them to do what the other does (I used to run an IRC to
e-mail bridge for reasons best left to history) but why? Anything powerful
enough to run Gemini can certainly run e-mail or netnews.

I'd love to know the source of the sentiment and if it happens to be an
aversion to 'switching applications', I'm sure a member of the church of
emacs will step up with alacrity.

--Matt

--
"The avalanche has started, it is too late for the pebbles to vote."
--Kosh

On 2022-04-06, Matthew Ernisse <matt@going-flying.com> wrote:
> It seems to me that the whole point of application protocols like e-mail,
> IRC, netnews, and Gemini itself is to accomplish a specific task. Yes, you
> can extend most of them to do what the other does (I used to run an IRC to
> e-mail bridge for reasons best left to history) but why? Anything powerful
> enough to run Gemini can certainly run e-mail or netnews.
>
> I'd love to know the source of the sentiment and if it happens to be an
> aversion to 'switching applications', I'm sure a member of the church of
> emacs will step up with alacrity.

Interacting with disparate protocols doesn't necessarily mean having a
dissonant experience. Your browser or OS can open your mail client
when encountering a "mailto" link. Pan and Lynx open your browser for
web links.

Hi Matt,

> > A "simple" solution would be going to the capsule, find the contact
> > of the author and send an email… But not every capsule has a contact
> > page and the process is very manual, requiring to use another tool and
> > protocol to communicate.
>
> It seems to me that the whole point of application protocols like e-mail,
> IRC, netnews, and Gemini itself is to accomplish a specific task. Yes, you
> can extend most of them to do what the other does (I used to run an IRC to
> e-mail bridge for reasons best left to history) but why? Anything powerful
> enough to run Gemini can certainly run e-mail or netnews.
>
> I'd love to know the source of the sentiment and if it happens to be an
> aversion to 'switching applications', I'm sure a member of the church of
> emacs will step up with alacrity.

I don't have an "aversion to switching applications" :). I'm fine
opening neomutt to send an email or weechat to have a direct convo :)
(And no, I don't need to switch to emacsOS :) </troll>)

The main reason for me doing this was:
Not everyone wants to put their email address publicly available.
I've been "blocked" multiple time trying to find a contact info in a
capsule without finding one to reach out to the author.
This could be true for the author of the response too (not wanting
to share email address).
This made me think about a solution that doesn't require sharing any
more information that you want/need on your capsule.

Then this idea morphed into finding a "standard" type of solution for
other interested too.

To be honest, I disagree with the fact that this is against gemini,
and seems that even solderpunk had some thoughts about this way back:
=> gemini://gemini.circumlunar.space/~solderpunk/gemlog/replies-in-geminispace.gmi

At the end of the day, it was for me a fun exercise to think, write
and code around this solution, even if it's not used :] 🤷

Cheers,
--
bacardi55 - https://bacardi55.io/ - gemini://gmi.bacardi55.io

On Thu, 7 Apr 2022 12:13:22 +0200, bacardi55 wrote:
> At the end of the day, it was for me a fun exercise to think, write
> and code around this solution, even if it's not used :] 🤷

Ah! I can totally understand wandering down the thought experiment
path. I hope it was rewarding.

--Matt

--
"The avalanche has started, it is too late for the pebbles to vote."
--Kosh

In the past I proposed my own solution to this problem and like
mentioned already, there have been many such proposals which ultimately
haven't been widely accepted.

Almost a year later my opinion is different. I now strongly believe that
gemini shouldn't care about this problem. Instead, maybe we should see
it as a read only medium and let other protocol handle followup
comments.

What if the authors that are interested in replies to their posts would
cross-post their content, or maybe just a link to their gemini content,
here on Usenet?

I know it's only a tiny niche of the geminauts that still use NNTP, but
what else could we agree upon as a common means to communicate in a
relaxed way? Besides, the ones who are already on this newsgroup could
perhaps persuade others to get back or try Usenet for the first time.
This would bring some influx of like-minded tech savvy people back here,
which in my book would be a nice side effect.

--
Vasco Costa

AKA gluon. Enthusiastic about computers, motorsports, science,
technology, travelling and TV series. Yes I'm a bit of a geek.

Gemini: gemini://gluonspace.com/

Vasco Costa <vasco.costa@invalid.invalid> writes:

> What if the authors that are interested in replies to their posts would
> cross-post their content, or maybe just a link to their gemini content,
> here on Usenet?

Usenet is clearly the technically correct choice for public
discussions. But on the other hand, I also feel like it puts an undue
burden on people wanting to reply or comment.

What I kind of like is the idea of there being an expectation that
authors put a contact email on the front page of their site (whether
that's at the top level of the domain, or under a ~user directory), and
ideally also in a file under /.well-known. It should also be generally
accepted that this address could be a secondary, anonymous, or throwaway
address.

--
Jason McBrayer | “Strange is the night where black stars rise,
jmcbray@carcosa.net | and strange moons circle through the skies,
| but stranger still is lost Carcosa.”
| ― Robert W. Chambers,The King in Yellow

On 11/04/2022, Jason McBrayer <jmcbray@carcosa.net> wrote:
> What I kind of like is the idea of there being an expectation that
> authors put a contact email on the front page of their site (whether
> that's at the top level of the domain, or under a ~user directory),
> and ideally also in a file under /.well-known. It should also be
> generally accepted that this address could be a secondary, anonymous,
> or throwaway address.

I like this idea very much too. I often get frustrated by not finding
any contact information. Has there been any attempt at choosing a
standard location/name for that file?

--
Vasco Costa

AKA gluon. Enthusiastic about computers, motorsports, science,
technology, travelling and TV series. Yes I'm a bit of a geek.

Gemini: gemini://gluonspace.com/

Vasco Costa <vasco.costa@invalid.invalid> writes:

> I like this idea very much too. I often get frustrated by not finding
> any contact information. Has there been any attempt at choosing a
> standard location/name for that file?

I don't know of any discussions in this direction.

Earlier this afternoon, I was thinking of a file at
/.well-known/contacts that would consist of lines like:

<PATH><WHITESPACE><EMAIL>

Where PATH is a path segment, below which the contact in this line
applies, unless there is a more specific path match. So:

/ admin@my.site
/~bob "Bob" <bob@my.site>
/~alice alice@my.site
/~alice/cat "Alice's Cat" <alices_cat@my.site>

This would be easily machine parsable, and you could add a "contact"
button in the client that opens an email compose program for any given
page.

But I also remembered "humans.txt", which is more freeform and perhaps
more humane.

=> http://humanstxt.org/ Humans.txt

--
Jason McBrayer | “Strange is the night where black stars rise,
jmcbray@carcosa.net | and strange moons circle through the skies,
| but stranger still is lost Carcosa.”
| ― Robert W. Chambers,The King in Yellow

This all seems to me like trying to add the "finger" protocol to gemini,
is not a bad idea, but in those files will be nice to add a bit more
information like gpg public keys and such.

On 2022-04-12, Jason McBrayer <jmcbray@carcosa.net> wrote:
> Vasco Costa <vasco.costa@invalid.invalid> writes:
>
>> I like this idea very much too. I often get frustrated by not finding
>> any contact information. Has there been any attempt at choosing a
>> standard location/name for that file?
>
> I don't know of any discussions in this direction.
>
> Earlier this afternoon, I was thinking of a file at
> /.well-known/contacts that would consist of lines like:
>
><PATH><WHITESPACE><EMAIL>
>
> Where PATH is a path segment, below which the contact in this line
> applies, unless there is a more specific path match. So:
>
> / admin@my.site
> /~bob "Bob" <bob@my.site>
> /~alice alice@my.site
> /~alice/cat "Alice's Cat" <alices_cat@my.site>
>
> This would be easily machine parsable, and you could add a "contact"
> button in the client that opens an email compose program for any given
> page.
>
> But I also remembered "humans.txt", which is more freeform and perhaps
> more humane.
>
>=> http://humanstxt.org/ Humans.txt
>

--
gemini://hispagatos.org
gemini://rek2.hispagatos.org
https://hispagatos.org
https://hispagatos.space/@rek2

rek2 hispagatos <rek2@hispagatos.org.invalid> writes:

> This all seems to me like trying to add the "finger" protocol to
> gemini, is not a bad idea, but in those files will be nice to add a
> bit more information like gpg public keys and such.

You're not wrong about that, and it would almost make sense to just use
finger, but finger is generally keyed off of username and host, which is
typically what you're trying to find out here.

humans.txt is closer to finger than my suggestion, of course.

--
Jason McBrayer | “Strange is the night where black stars rise,
jmcbray@carcosa.net | and strange moons circle through the skies,
| but stranger still is lost Carcosa.”
| ― Robert W. Chambers,The King in Yellow

On Mon, 11 Apr 2022 10:09:31 -0400, Jason McBrayer wrote:
> What I kind of like is the idea of there being an expectation that
> authors put a contact email on the front page of their site (whether
> that's at the top level of the domain, or under a ~user directory), and
> ideally also in a file under /.well-known. It should also be generally
> accepted that this address could be a secondary, anonymous, or throwaway
> address.

I think this is reasonable, and if one wants to make it easy mailto: is a
valid URI scheme and is widely understood so one could even provide a
link eg:

=> mailto:matt@going-flying.com Contact Me

I don't think one needs to get any more complex than that.

--
"The avalanche has started, it is too late for the pebbles to vote."
--Kosh