💾 Archived View for rawtext.club › ~sloum › geminilist › 001838.gmi captured on 2020-11-07 at 02:30:17. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2020-09-24)

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

<-- back to the mailing list

[ANN] List of Gemini web mirrors

Peter Vernigorov pitr.vern at gmail.com

Mon Jun 22 14:29:33 BST 2020

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

Thanks for feedback, Luke. I initially went with the [N] syntax. WhatI found out was that "url [1]" does not work because of things like"[constellation]s" becoming "constellation [1]s". "[1] url" is notideal either, as some clients, like av98, would already have a linkcounter, and since I have a 2 links hardcoded at the top of the pageto search and home, I ended up needing to increment link counter by 2,and even then it would be rendered as "[22] [22] name". In the end Ichanged it to "... [url] ..." so the reader knows that they can getmore info by finding the link at the end of the paragraph. Hopefullyit's a good compromise.

On Sat, Jun 20, 2020 at 10:19 AM Luke Emmet <luke at marmaladefoo.com> wrote:

Hi Peter
That looks great and is worth keeping if you are happy to host it.
The question of what is the equivalent of in-line links in Gemini comes
up from time to time, either as it is a natural thing that authors want
to do and already do, or when re-presenting existing markdown or html
for gemini.
The most common idiom I have seen is the use of square brackets to
indicate the placemarker in the text line as a citation, followed by a
link having the reference. For example like this [1] that would be one
of the subsequent links, or another one [2] that goes to the second one.
It is a common form seen in many places, such as academic papers with
footnotes and references.
=
url display text with matching item at end [1]
=
url2 [2] display text with match at beginning
Personally I find this better than simply having the text without any
clear boundary, as it is clearer where the citation is made. For example
if you just use a single word it is unclear which usage of it is the
link anchor. There could even be multiple words that match, and you
don't want them all implicitly referencing the link. So it is more
specific this way.
My personal view is that this type of re-wiring-up, to put back the
links into the text could be a client option, and user choice. Then the
hotspots in the text could be reinstated. The criteria would be:
1. Link anchor uses the defined pattern e.g.
- [n] as the first or last item in the display text
- or more adventurously, matching text in a square bracket as the
whole link line display text like this: [the thing]
2. After the line is a list of links, and there is a match as the first
or last word in the display text. Or maybe just in the following content
(like a list of references at the end of the page)
3. Then the link anchor is wired up to the target
- optional) the link line is optionally hidden (again user and client
choice)
It would be nice if the mirroring tools adopted a common convention on
this, as then clients can do more work to improve the UI for users.
And anyway, this all gracefully degrades and is just a client nicety. It
could be noted as part of our collective common practice.
Best wishes
- Luke
On 19-Jun-2020 23:46, Peter Vernigorov wrote:
Wikipedia on Gemini by popular demand:
gemini://wp.pitr.ca/en/Gemini_(constellation)
This is not strictly a mirror, as a dump of wikipedia is quite large -
https://dumps.wikimedia.org/enwiki/20200520/ - but rather a proxy that
makes request to wikipedia API, parses wikitext into an HTML DOM
(since wikitext supports HTML tags) which is then simplified into
text/gemini. This process is not perfect and there are quite a few
imperfections. But before I sink more time into this, I wonder if
text/gemini is indeed the best format for wikipedia articles. Pages
are usually huge, with lots of links. And this is even before
considerations for how best to handle images, tables, special symbols,
special tags, etc. From trying to browse it myself, my first
impression is that wikipedia format can't/shouldn't be simplified any
further than it already is on the web. What do others think?
On Wed, Jun 17, 2020 at 12:45 PM defdefred<defdefred at protonmail.com> wrote:
Soon a wikipedia mirror?
:-)
‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Wednesday 17 June 2020 12:24,<paper at tilde.institute> wrote:
On Wed, Jun 17, 2020 at 10:17:58AM +0000, solderpunk wrote:
On Wed, Jun 17, 2020 at 05:56:57AM -0400, paper at tilde.institute wrote:
Good job with the "Why?" page! You convinced me quickly.
Cheers,
Solderpunk
I wasn't sure the Why page was good enough, thanks alot.
Paper