💾 Archived View for gemi.dev › gemini-mailing-list › 000934.gmi captured on 2023-11-04 at 13:12:45. Gemini links have been rewritten to link to archived content

View Raw

More Information

➡️ Next capture (2023-12-28)

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

In Body URI references

Jonathan McHugh <indieterminacy (a) libre.brussels>

Im just wondering regarding in body URI references within content
bodies.

Does anybody have advanced policies to demarcate the citation and have the URL
below?

Is there a style like Bit.ly for url shortening? Which still respects
Gemini design principles?

For example:
 ```
This is url foobar, http://foobar.com/example.html
 ```
Becoming:
 ```
This is url foobar [QQQ63rteet63655346]

~~~~~

# References
http://foobar.com/example.html [QQQ63rteet63655346] 
 ```

Or, for terseness:
 ```
This is url foobar [QQQ1]

~~~~~

# References
http://foobar.com/example.html [QQQ1] [QQQ63rteet63655346] 
 ```

Is there something that integrates well with Tex style .bib reference
repositories?

Would it be interesting for there to be a (de)centralised site, which
provides?:

(re)interpretations.

Kind regards,


-- 
Jonathan McHugh
indieterminacy at libre.brussels

Link to individual message.

Stephane Bortzmeyer <stephane (a) sources.org>

On Sun, Jun 13, 2021 at 03:17:35PM +0200,
 Jonathan McHugh <indieterminacy at libre.brussels> wrote 
 a message of 48 lines which said:

> Does anybody have advanced policies to demarcate the citation and
> have the URL below?

This is probably a job for the CMS, not for Gemini itself.

> Is there a style like Bit.ly for url shortening? Which still
> respects Gemini design principles?

I don't think any URI redirector could respect Gemini design
principles. For instance, these redirectors are a privacy issue.

> Would it be interesting for there to be a (de)centralised site, which
> provides?:
> * A dictionary of url links
> * Appropriate hashes that can be used for publishing or client
> (re)interpretations.

WHat is exactly the problem you are trying to solve? I can imagine
some uses for such a site but I'm not sure which you had in mind.

Link to individual message.

Jonathan McHugh <indieterminacy (a) libre.brussels>

Hi Stephane,

Thanks for your interest.

Stephane Bortzmeyer <stephane at sources.org> writes:

> On Sun, Jun 13, 2021 at 03:17:35PM +0200,
>  Jonathan McHugh <indieterminacy at libre.brussels> wrote 
>> Is there a style like Bit.ly for url shortening? Which still
>> respects Gemini design principles?
>
> I don't think any URI redirector could respect Gemini design
> principles. For instance, these redirectors are a privacy issue.
>

I have no desire to redirect users.

Im thinking in terms of inline reference to unclutter a large paragraph
while still collecting the urls below a section or as a footer.

>> Would it be interesting for there to be a (de)centralised site, which
>> provides?:
>> * A dictionary of url links
>> * Appropriate hashes that can be used for publishing or client
>> (re)interpretations.
>
> WHat is exactly the problem you are trying to solve? I can imagine
> some uses for such a site but I'm not sure which you had in mind.

I would prefer to work to a standard equivalent to wikipedia
content. IMHO. Its a good expectation for managing content, irrespective
of site or service.

As an example, the HTTP wikipedia page for Gemini Protocol starts with this
paragraph
=> https://en.wikipedia.org/wiki/Gemini_(protocol)
 ```
The Gemini protocol is an application layer protocol for distributed 
hypertext information systems that provides access to simple, primarily 
textual documents in Gemini space. This is done with contemporary 
technologies such as TLS, thereby improving privacy and user agency 
relative to the Web. Servers run by default on port 1965. The protocol is 
being designed collaboratively and is not currently being standardized as 
an internet standard. 
 ```
Pasting this echews 6 uri links:
=> https://en.wikipedia.org/wiki/Application_layer
=> https://en.wikipedia.org/wiki/Hypertext
=> https://en.wikipedia.org/wiki/Gemini_space
=> https://en.wikipedia.org/wiki/Transport_Layer_Security
=> https://en.wikipedia.org/wiki/Internet_standard

Naturally these are all unternal links, which could be referenced if
needed. I expect serving ALL references as a footer bibliopgraphy is
equitable.

I do not know whether clients or users have policies for 'folding away'
cross references for obvious/queryable suggestions.

The next paragraph contains 2 internal navigation links and 2 internal citations:
 ```
The design is inspired by the Gopher protocol, but mandates the use of 
Transport Layer Security with trust on first use (TOFU)[1] and 
privacy-related features. It is not intended to replace Gopher or HTTP, 
but to co-exist with them.[2]
 ```
=> https://en.wikipedia.org/wiki/Internet_protocol_suite
=> https://en.wikipedia.org/wiki/Trust_on_first_use
=> https://en.wikipedia.org/wiki/Gemini_(protocol)#cite_note-1
=> https://en.wikipedia.org/wiki/Gemini_(protocol)#cite_note-faq-2

The use of [:digit:] doesnt seem to be stable for content
(intentionally) containing square brackets, such as in lua
 ```
x = foo[1]
 ```

The proxy Gopherpedia eschews internal links and citations for the
content it generates.

I have come across this repo for proxying wiki content with Gemini but I
havent come across a working service yet
=> https://github.com/pitr/wp

I dont know whether there are are successful CMS approaches for cleanly
managing links.


I have been reminising about the social bookmarking site Delicious. Id
be happy if by aggregate or by project I was capable of having more
solid references across documents and references.

I have been intrigued by the potential of Emacs' Hyperbole as a
mechanism for interpreting content and interfacing with it. Given that
it cooperate with (the Gemini client) Emacs' Elpher I would like to
think there could be utility
=> https://www.gnu.org/software/hyperbole/DEMO.html#Social-Media-Hashtags-and-Usernames

Approaches include
 ```
   [facebook|instagram|twitter][#@]<hashtag-or-username>
   [fb|in|tw][#@]<hashtag-or-username>
   gh#rswgnu/hyperbole/5ae3550 (if include user, must include project)
   github#hyperbole/5ae3550    (project can be given with user default)
   gh#5ae3550                  (user and project defaults are used)
   git#/hyperbole            (displays the top directory of the hyperbole repository)
   git#/hyperbole/55a1f0     (displays hyperbole git commit diff)
   git#=hactypes.el          (displays a git-versioned file regardless of directory)
   git#=master:kotl/kview.el (displays file in subdirectory from master branch)
   gt#55a1f0                 (when within a git repo, displays its commit diff)
 ```                   
As well as this
 ```
   A document id is used just like a reference citation in traditional 
publications but it
   actually links to the document that it references and the card catalog
   (index) entry for the document.  One can easily pass around doc ids to point
   people to appropriate documents.  For example, a mail message in response to
   a question might say, "See [Emacs-001] for examples of what Emacs can do."
 ```

Here is a Hyperbole mailing list thread, concerning my generation of
navigation links to anchored queries (based upon searching my
filesystem)
=> https://lists.gnu.org/archive/html/hyperbole-users/2021-05/msg00050.html

I would like to have content served to me in Elpher, with actionable
links that permit me to visit links with little fuss. If I can consume
(or create) content like wikipedia with the links but with minimal URI noise, it
would make me happy.

--
Jonathan McHugh
indieterminacy at libre.brussels

Link to individual message.

ew.gemini <ew.gemini (a) nassur.net>

Hello,

Jonathan McHugh <indieterminacy at libre.brussels> writes:

> Im just wondering regarding in body URI references within content
> bodies.
>
> Does anybody have advanced policies to demarcate the citation and have the URL
> below?

What I use and see is this:

$author writes:
> Quote text marked by a leading '>', possibly several such lines
=> URL for exactly this thing.


And in longer text I refence THING [a] and maybe another thing
[b] and collect these at the end of the document


=> URL [a] URL_or_text
=> URL [b] URL_or_text

There was someone suggesting the use of [characters] rather than
[numbers] for some reason I forgot. But it made sense at the time.

I do this manually.


If I need this in long text, then I would switch to a different
representation, like \LaTeX and prepare a .pdf Version for
download. I would not try to coerce this onto gemtext.


Cheers,
Erich



-- 
Keep it simple!

Link to individual message.

remyabel@tilde.team <remyabel (a) tilde.team>

I don't think there is a convention. Personally, I do the bracket
approach like so:

	This is some text.[1]

	=> https://example.com 1: This is a link.

If it's a footnote I just use f and render them all at the bottom of the
page.

	This is a footnote.[f1]

	## Footnotes

	=> https://example.com 1: This is a footnote.

Bombadillo already prepends an index before each link to allow you to
jump to it easily, so the number may seem redundant, but the number
allows you to easily group links per paragraph. Example:

	One[1].

	=> https://example.com 1: This is a link.

	One[1] two[2] three[3].

	=> https://example.com 1: This is a link.
	=> https://example.com 2: This is a link.
	=> https://example.com 3: This is a link.

Converting a wiki page to gemtext isn't too difficult, it's a matter of
finding an appropriate library and emitting the appropriate gemtext for
each HTML tag. For an example see:
https://tildegit.org/remyabel/tilde-wiki-gemini (note: not production
code.)

Link to individual message.

Alexis <flexibeast (a) gmail.com>


ew.gemini <ew.gemini at nassur.net> writes:

> There was someone suggesting the use of [characters] rather than
> [numbers] for some reason I forgot. But it made sense at the 
> time.

Maybe this post of mine?

    gemini://republic.circumlunar.space/users/flexibeast/gemlog/2020-09-30.gmi

"Given that some Gemini clients make links accessible via 
numbering - e.g. AV-98, and elpher with my proposed patches 
applied - i've decided to not use numbering for footnotes, and 
instead use lower-case letters[a]. i've personally found that 
gemtext pages which use numbering for footnotes can get a bit 
confusing in the presence of link numbering"


Alexis.

Link to individual message.

ew.gemini <ew.gemini (a) nassur.net>


Hello Alexis,

Alexis <flexibeast at gmail.com> writes:

> ew.gemini <ew.gemini at nassur.net> writes:
>
>> There was someone suggesting the use of [characters] rather than
>> [numbers] for some reason I forgot. But it made sense at the time.
>
> Maybe this post of mine?
>
>    gemini://republic.circumlunar.space/users/flexibeast/gemlog/2020-09-30.gmi
>
> "Given that some Gemini clients make links accessible via numbering -
> e.g. AV-98, and elpher with my proposed patches applied - i've decided
> to not use numbering for footnotes, and instead use lower-case
> letters[a]. i've personally found that gemtext pages which use
> numbering for footnotes can get a bit confusing in the presence of
> link numbering"

Yepp. That sounds "familiar" :-)

Thank you for providing the missing piece!

Cheers,
~ew


-- 
Keep it simple!

Link to individual message.

Rohan Kumar <seirdy (a) seirdy.one>

On Sun, Jun 13, 2021 at 06:09:43PM +0200, ew.gemini wrote:
>There was someone suggesting the use of [characters] rather than 
>[numbers] for some reason I forgot. But it made sense at the time.
>
>I do this manually.

I do something similar, but with Unicode superscripts. An example with 
lots of footnotes:

=> gemini://seirdy.one/2021/01/12/password-strength.gmi

Excerpt:

>A good value for T would be the average temperature of the entire 
observable universe. The universe is mostly empty; T is around the 
temperature of cosmic background radiation in space. The lowest reasonable 
estimate for this temperature is 2.7 degrees Kelvin.? A lower temperature 
means less energy usage, less energy usage allows more computations, and 
more computations raises the upper limit on password strength.

>[...]

>## Citations and Footnotes

>? James Massey (1994). "Guessing and entropy" (PDF). Proceedings of 1994 
IEEE. International Symposium on Information Theory. IEEE. p.?204.

>? Assis, A. K. T.; Neves, M. C. D. (3 July 1995). "History of the 2.7 K. 
Temperature Prior to Penzias and Wilson"

I borrowed this convention from a couple other capsules already using 
it.

-- 
/Seirdy

Link to individual message.

---

Previous Thread: MoonGem 1.4 + Docker Image

Next Thread: A proposal to fund the Gemini project