💾 Archived View for envs.net › ~pen › gemini captured on 2021-12-05 at 23:47:19. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2021-12-04)
-=-=-=-=-=-=-
Some ramblings when initially encountered Gemini end of 2020 and beyond.
… as he's not into that brevity thing.
https://lists.orbitalfox.eu/archives/gemini/2021/007399.html
Additional capacities in the gemtext format are not necessary. That's not just, like, my opinion, man, that's an empirical fact. Geminispace is there. It's *exactly* the kind of space I originally envisaged.
[original emphasis] Bit *funny*, though 😏
Hm, should the git tracker be deleted?
both on this list and in the git trackers that sprung up once I delegated spec finalisation
Further, nothing about Usenet, so I guess we're on our own there.
⚔ The Gem is mightier than the sword.
Edited to add: That should probably read "The Gum is mightier …"
Don't worry, I've got plenty of gum.
Big-8 Management Board has decided by 3/3 consensus to create unmoderated newsgroup comp.infosystems.gemini
However, this is the first "Big-8" newsgroup to be created in 8 years so things might be a bit rusty.
https://lists.orbitalfox.eu/archives/gemini/2021/007388.html
Gemini bridging past and present.
But sometimes stuff disappears because the author didn't want it up anymore and willingly took it down, which feels like a reasonable thing for authors to be able to do if they like.
[…]
Basically a “unforgeable append-only” system is similar to a legal set of contracts and not at all like conversations in real life". He's well aware that it's impossible to guarantee all participants in a distributed system will comply with a request to delete something, but still thinks it's better to build systems which at least *try* to let us undo. I'm still not 100% sure I agree, but I totally understand and respect the perspective.
gemini://alexschroeder.ch/page/2018-06-29_No_Take_Back
Usenet has Cancel messages that let authors cancel their previous posts, but claiming authorship is not very solid:
A cancel message requests the deletion of a specific article. The body of the Control field contains one argument, the Message-ID of the article to delete.
[…]
According to RFC 1036 only the author of the target message or the local news administrator is allowed to send a cancel (cancels not meeting this condition are called "rogue cancels"). To verify authorization the From: line (or Sender: line, if it exists) of the cancel message must match the target article. This verification does not work well in modern day Usenet and is rarely used.
RFC 850 uses the term "local super user" instead of "local news administrator". > Son-of-RFC-1036 (this is the colloquial name of an Internet Draft written by Henry Spencer) drops the administrator's cancel altogether. The problem with the verification scheme is that the From: line is trivial to forge and with cancelbots the cancel message often arrives before the target article.
https://en.wikipedia.org/wiki/Cancel_message
With OpenPGP claiming authorship should be feasible.
Actually German Wikipedia has more information on cancelling Usenet posts, e. g. also mentioning "NoCeM".
NoCeM – pronounced: [nou'si:əm] or as if No See 'Em – is a protocol enabling authenticated third parties to issue notices which site administrators can use to delete unwanted articles from their news spool, or which end users can use as a 'third-party killfile'. It is intended to eventually replace the protocol for third-party cancel messages. Other kinds of application are also possible.
https://rosalind.home.xs4all.nl/nocemreg/faq-reg.html#about-nocem
https://de.wikipedia.org/wiki/Canceln
As has been pointed out on the mailinglist, the Gemini community should not rush things.
Solderpunk still seems to be the perfect BDFL here, so knowledgable and calm.
gemini://gemini.circumlunar.space/users/solderpunk/gemlog/been-an-even-longer-time.gmi
There has been some fruitful discussions about establishing a newsgroup specifically for Gemini.
I liked the idea of changing the existing group comp.infosystems.gopher to something like comp.infosystems.small(net) or just using comp.infosystems.hypertext. Well, as was noted on the mailing list a name with actually Gemini in it helps to find and adopt the group.
Free newsgroup service providers include Eternal September, Solani, and AIOE, the latter restricted to posting forty own articles per day and IP but no need to create an account. Forty is plenty, isn't it?
https://www.eternal-september.org
Was Eternal September's previous name Motzarella? Well, sounds like one of my old puns, actually.
Motzarella.org (Last update 05/05/09 Soon to be Eternal September)
http://groupsearch.alt-config.net
Gemtext paragraphs spanning multiple sentences must reside in a single line to be rendered accordingly. Compared to editing Markdown this asks the editor to be more knowledgable about keying in their textual content. Thus, Geminauts propose to configure vim with softwrapping long lines along word boundaries – maybe at given column width instead of the (terminal) window width:
:set wrap :set linebreak :set columns=80
Also navigating not just using keys hjkl but prefixed by g to only jump visual/soft lines. Or use these settings:
noremap <silent> <expr> j (v:count == 0 ? 'gj' : 'j') noremap <silent> <expr> k (v:count == 0 ? 'gk' : 'k')
Thanks for the tips, people of Gemini, some pointers:
https://gerikson.com/blog/comm/Gemini-misaligned-incentives.html
https://lists.orbitalfox.eu/archives/gemini/2020/003315.html
https://lists.orbitalfox.eu/archives/gemini/2020/003311.html
Gemini has its mailing list. End of 2020 Solderpunk established topics like announce, users, spec to be used, so that subscribers could choose to receive less posts. Recently rules got more strict, as I understand, e. g. not allowing to answer posts without being subscribed (which isn't straightforward anyways to get the Reply-To header right, I guess).
Some people understandably prefer not to subscribe to the list, possibly cluttering their inbox or triggering notifications, due to lack of filtering possibilities in their email hosting. Personally I generally do not like the concept of mailing lists for this and other reasons, e. g. enjoying Pull more than Push. It's debatable but I'd consider newsgroup messages Pull whereas email messages Push, but of course it depends on the mode employed by each user in their clients.
For discussions about the spec there is now a repository on GitLab. This seems great for focussing on individual issues. To take part you need to register with GitLab, registration seems to be a hurdle with some browsers, though. Thus, alternatives were suggested, e. g. Codeberg and Sourcehut, with the letter being able to take part also via email only.
There's an inofficial IRC channel, and maybe some others.
Possibly lots of good thoughts scattered among various capsules.
Some notable people involved, without much of any aim for comple...
Personally, I think a community wiki could be good to maintain collaboratively pointers and excerpts organized by topics, e. g. TLS, TOFU, accessibility, gemlog structures, metadata, etc.
Once again I restructured my sandbox-like content in this capsule.
It seems I like a more topic based content organization, where I could come back and edit things later on. For transparency reasons edits, at least edits changing the gist, should be accompanied by metadata denoting such change, though. Something like:
Edited on [date] to add ...
Also, if edits are substantial, change the date heading and denote the history of the post?
So not TOFU up?
gemini://tilde.team/~khuxkm/consciousness/certinfo.gmi
gemini://drewdevault.com/2020/09/21/Gemini-TOFU.gmi
gemini://gemini.circumlunar.space/docs/specification.gmi
Quiet speakers' corner?
top or bottom posting
what sort order hmm
To make subscriptions possible via the homepage but nevertheless to only have a single file per year, just also update the `index.gmi` with the latest date stamp linking to the `{year}/index.gmi`. Like so:
# My Index => 2021/index.gmi 2021-02-27 Gemsub Test
Does that make sense?
Gladly gemlogs often feature feeds (RSS, Atom – not seen JSON, yet). Now gemini clients add feed subscription capabilities.
Not only when discovering twtxt back some years ago and also JSONfeed and now Gemini I tend to think that it would be nice to simply have all the content of a site/capsule in one single feed available in your favourite feed reader – without really the need for further clients, as the feed reader displays everything what's there, not just a summary or teaser.
Is feed size and repeated fetches a concern? … could be split by year or cf. "next_url" of https://jsonfeed.org/version/1.1
Unfortunately newsboat does not yet support JSONFeed as of 2020-12.
As finally I adapted the themes in Amfora client to match my usual terminal colours, I really enjoy looking at pages of Gemini content in it.
Nevertheless I wonder, whether monospace fonts as widely used in terminal emulators, is not too much of a draw back in terms of readability as opposed to proportional (and serif?) typefaces.
By the way, this collection of Gemini software and related projects is awesome: https://github.com/kr1sp1n/awesome-gemini
I'm a fan of "clean" URLs, i. e. those without a file suffix/extension, like ".gmi".
As for the gemlog, I first thought that a directory for each day is too much of a structure and I'll save the Gemini content in files without extension.
Turns out the used Gemini server does not serve the content as Gemini that way.
So let's script the mkdir part a bit.
Also, we might want some meaningful pieces of information coded in the URL, hmm …
Is Gemini too simple for typographic emphasis?
I guess I'd enjoy seeing Gemini clients optionally offer rendering emphasized words differently, e. g. *bold*, /italic/, _underline_, optionally with(out) deleting the control characters.
On the other hand, we still are used to see ones names written without diacritics and/or transliterated (e. g. on id/credit cards), so let's keep Gemini as s i m p l e as it is. With UTF-8 there's probably enough on offer, e. g. narrow non-break space (U+202F).
I found the idea compelling that Gemini clients would show the user, which external Gemini pages link to the current one. How should the client or the server know this? The idea was that clients send a referer when visiting a page via link, the server collects these and delivers the list of referers when requestet by the client for the current page. Let's be glad for privacy reasons we do not send referers. Alternatively, a large central database like the gus.guru search engine provides the ability to search for these so-called backlinks.
Is there maybe something like webmentions?