💾 Archived View for rawtext.club › ~sloum › geminilist › 006992.gmi captured on 2024-02-05 at 10:53:46. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2021-11-30)

-=-=-=-=-=-=-

<-- back to the mailing list

[Discussion][Server Script Idea] Gemlog pages that automatically add links to each other

Andrew Singleton singletona082 at gmail.com

Tue Aug 10 16:18:48 BST 2021

- - - - - - - - - - - - - - - - - - - 

This started by me going through earlier discussions (I still have noidea how to condense the search down so it isn't blindly scrolling) ondifferent methods of replicating discussion/forum and the one I likedmost actually isn't a really discussion board/thread type of feel.

I don't have the archive in front of me, so forgive me if I havemisinterpreted the original idea. However this is my proposal eventhough I don't know how hard it is to implement technically, or ifthere is any kind of want. This is just me seeing several threadscovering 'we want more discussion that isn't retreating to BBS's, orcrawling back to the web' and having a brainwave that 'maybe thiscould help solve the problem of how to find new content since thesearch engines don't seem to be doing so well (or maybe it's just me.)

File this under 'MOL' to keep with the naming convention of the protocol.=

https://en.wikipedia.org/wiki/Manned_Orbiting_Laboratory

You would give your server a list of gemini pages (specific singularpages or whole domains) to look through hourly/daily/weekly (I wouldsuggest daily as a default to avoid spam) and if a specificsubheading comes up (example if the line after H1, or closingunformatted text (in the case of pages starting with ascii art) is aspecific string (Example: Banana) it would send you an emailnotification that 'banana' has come up and give you the link/s. Thisis Stage One. I am sure you're going 'well that sounds like RSS/Atomyou're reinventing the wheel.'

Stage Two involves the same server backend searching your OWN gemlogand if 'banana' is found because you read the inload and want to writea page either responding or linking to that page the followinghappens.:* Server takes the links and timestamps of the latest sweep it did for 'banana'* Server then adds at the end of your 'banana' gmi file these links.Exact formatting could be user editable, but by default I want it tohave an H3 heading [Most Recent uses of 'Banana'], right below thatheading [Terms searched from select Gemini capsules from $DATE to$DATE].If the update interval is longer than one day. If update interval isshorter than one day: [Terms searched from select Gemini capsules on$DATE : $LOCAL_TIME]* List is populated with links.* Footer text is applied if applicable. This text can include a seriesof separator characters followed by navigation links or a simple EOLstatement.

Potential applications:* Direct discussion between differing gemini users.* Pooling information on a given subject or theme so that the endreader can immediately jump between tasks* Ease the creation of Web-Ring like site chains.* Experimentation.

Potential Issues:* The fact this assumes a known site list greatly limits propagationas it doesn't matter if several user clusters adopt this if none ofthem know the others exist for the sake of link creation. At the sametime, though, this could also be seen as a positive depending on ifthe server admin wishes to limit who and what they wish to interactwith.*Low Adoption Rate: There are already several server setups out there.While this feature could be seen as useful, given it must compete withsimpler solutions for those that 'just want a stinking capsule' alongwith options like Flounder that are currently feature locked due tocode maintainer/s not having the time or interest to do more thandebug. This is one of those features where you either need areasonably high adoption rate, or it dies out. * Inability for the proposal writer to properly articulate theconcept framework.

I am confident those with a better grasp of server operation andprogramming could come up with a better system or at least optimizethe flow of operation. If nothing else, having a specific prefixcharacter to denote tags '+' or '@' as examples may help to allow forpages that users want to appear with multiple search terms and orallow strings that have more than a single word to them.

This is not intended to emulate discussion boards or mailing lists,though I am sure there will be those that will treat it as such. Thisis intended more for the sake of allowing document sharing betweenseveral servers.