💾 Archived View for gmi.bacardi55.io › gemlog › 2023 › 01 › 03 › gemini-mention-coincidence captured on 2023-09-28 at 15:57:09. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-01-29)
-=-=-=-=-=-=-
Posted on 2023-01-03
Nota: I started this entry as a short fun fact about a coincidence, and ended writing a long post responding to Martin and Sean, I'm sorry, but it seems I can't be concise…
On Sunday, while publishing my review for 2022 from my new laptop, I had to reinstall a few things for the deployment of my capsule and blog to work. I was using an old version on kiln[1] (static capsule generator) on my previous laptop, so after installing the latest one, I had some issues to fix. I took the "opportunity" to clean a bit my capsule templates.
I decided to remove the "gemini mention" text displayed at the bottom of each pages. For those unfamiliar with the gemini mention concept, the basic idea is about a minimalist version of webmention for gemini to alert fellow geminauts that you wrote a response about an page they wrote themselves. For more information, read my gemlog entry about it[2] or the RFC itself[3].
Anyway, I remove this on Sunday and thought I would forget about it. I was surprised and amused that literally the next day, Martin[4] wrote a post about it where he found a solution to avoid the CGI script (but still requires a script to analyze the logs instead). I was even more surprised to read another post about gemini mention today on Sean's capsule[5].
Random coincidences like this often makes me smile :). So I decided to re-enable gemini mention on my capsule again (link at the end of each post)
Sean wrote:
As far as the actual proposal itself [8], I don't have much to comment about it, except to say I don't like the mandated text it uses in the link. I think just finding the link should be good enough.
Indeed, the proposal indicates that the text link must start with `RE:'. The complete proposal for the link is:
=>[<whitespace>]<URL><whitespace>RE:<whitespace><USER-FRIENDLY LINK NAME>
The main goal of the `RE:' part is to simplify parsing. A gemtext page may contain many links, and the `RE:' starting text is simplify a way of simplifying parsing if you want to make sure the URI provided indeed reference a specific page. If this is too restrictive, I'm happy to remove that limitation. The RFC I wrote is, as the acronym RFC implies, open for comments and discussions! So let me know your thoughts and I'm fine making it easier for other to embrace the idea.
Sean also wrote:
Even better in my mind would be two links, much like webmention uses, but that doesn't seem to be a popular opinion.
If I understand Sean's proposal (read his page[6]), it means I need first to define a full url like this:
gemini://gemini.conman.org/gemmention?source=gemini://gmi.bacardi55.io/gemlog/2023/01/03/gemini-mention-coincidence/&target=gemini://gemini.conman.org/boston/2023/01/02.2
And then copy it to my browser (or send it via a cli tool like gemget).
Which, I think, is a little bit more complex than having a mandatory "RE:" at the start of the link description and then just posting the link of the response itself without the need of 2 arguments in the url. But that's just my take on it.
But as said above, I'd rather have a "community agreed" proposal than one that 100% please me but I'm the only one to use :). So keep comments and criticism coming! I like this type of exchange and brain challenge about finding the right proposal for the bigger audience, and I won't be offended by any constructive criticism. I don't have an issue adding some code that would analyse URI parameters per say, I just thought it was maybe a bit against the gemini protocol concept.
Sean also wrote a simple proof of concept for this proposal in less than 20 lines of lua code[7], check it out if you are interested.
Martin had a different concept, and I thought the idea of parsing logs instead of having a CGI script interesting, specially if you are using a gemini server that does not handle CGI scripts. Check his post for more details[8]. Personally, I found it more complex to manage real time log parsing than a simple executable file that does it live, but that is because I'm using a server with CGI scripts and I can write simple code to make it works. I can understand Martin not wanting to change its gemini server just for gemini mention to work though.
I'd love to see more geminauts use any of these solutions, so feel free to respond on your gemlog or create a ticket on the RFC git repository on codeberg or send me an email.
This gemlog entry is (partly) a response to:
RE: It still surprises me what some find difficult to do (Conman)
RE: Gemlog responses - bacardi55's concept without CGI (Martin)
Footnotes
_________