💾 Archived View for gemi.dev › gemini-mailing-list › 000088.gmi captured on 2023-11-04 at 12:24:34. Gemini links have been rewritten to link to archived content

View Raw

More Information

➡️ Next capture (2023-12-28)

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

Wikipedia mirror over gemini "proxy"

Travis Briggs <audiodude (a) gmail.com>

Hello,

I have an idea for quickly and vastly expanding the amount of content that
is available in Geminispace. What about a Wikipedia mirror/proxy? I put
proxy in quotes because it would modify the documents it returned in order
to be largely text based.

The main thing that's tripping
<https://en.wikipedia.org/wiki/Psychedelic_experience> me up is the fact
that wikipedia <https://en.wikipedia.org> articles are written
<https://en.wikipedia.org/wiki/Writing> with a lot of inline links. And
navigating those links is a key part of the Wikipedia browsing experience.

My naive read of the text/gemini format is that I would have to do
something like:

The main thing that's
=> gemini://gem.wiki-mirror.pizza/article.cgi?Tripping tripping
me up is the fact that
=> gemini://gem.wiki-mirror.pizza/article.cgi?Wikipedia wikipedia
articles are
=> gemini://gem.wiki-mirror.pizza/article.cgi?Writing written
with a lot of inline links.

Which seems really clunky. The other option I'm considering is serving
text/markdown, which I assume most Gemini clients would display as plain
text. The problem then is that [links won't be clickable](http://dev/null).
But maybe the mere existence of so much markdown content would incentivize
some client author to add native support?

Feedback is appreciated, thanks!
-Travis

Link to individual message.

colecmac@protonmail.com <colecmac (a) protonmail.com>

Personally, I think markdown would make the most sense in this usecase.

> I have an idea for quickly and vastly expanding the amount of content
> that is available in Geminispace. What about a Wikipedia mirror/proxy?

I think this idea is kinda flawed. I can only speak for myself,
but I'm guessing that most people here want the Geminispace to expand
naturally, and for interesting and unique content to come up. There's
nothing wrong with a Wikipedia proxy, but I feel like it misses the
point a bit. If I want to read Wikipedia, I would just use Firefox.

makeworld

??????? Original Message ???????
On Friday, May 15, 2020 6:10 PM, Travis Briggs <audiodude at gmail.com> wrote:

> Hello,
>
> I have an idea for quickly and vastly expanding the amount of content 
that is available in Geminispace. What about a Wikipedia mirror/proxy? I 
put proxy in quotes because it would modify the documents it returned in 
order to be largely text based.
>
> The main thing that's tripping me up is the fact that wikipedia articles 
are written with a lot of inline links. And navigating those links is a 
key part of the Wikipedia browsing experience.
>
> My naive read of the text/gemini format is that I would have to do something?like:
>
> The main thing that's
> =>?gemini://gem.wiki-mirror.pizza/article.cgi?Tripping?tripping
> me up is the fact that
> => gemini://gem.wiki-mirror.pizza/article.cgi?Wikipedia wikipedia
> articles are
> => gemini://gem.wiki-mirror.pizza/article.cgi?Writing written
> with a lot of inline links.
>
> Which seems really clunky. The other option I'm considering is serving 
text/markdown, which I assume most Gemini clients would display as plain 
text. The problem then is that [links won't be 
clickable](http://dev/null). But maybe the mere existence of so much 
markdown content would incentivize some client author to add native support?
>
> Feedback is appreciated, thanks!
> -Travis

Link to individual message.

Sean Conner <sean (a) conman.org>

It was thus said that the Great Travis Briggs once stated:
> Hello,
> 
> I have an idea for quickly and vastly expanding the amount of content that
> is available in Geminispace. What about a Wikipedia mirror/proxy? I put
> proxy in quotes because it would modify the documents it returned in order
> to be largely text based.
> 
> The main thing that's tripping
> <https://en.wikipedia.org/wiki/Psychedelic_experience> me up is the fact
> that wikipedia <https://en.wikipedia.org> articles are written
> <https://en.wikipedia.org/wiki/Writing> with a lot of inline links. And
> navigating those links is a key part of the Wikipedia browsing experience.
> 
> My naive read of the text/gemini format is that I would have to do
> something like:
> 
> The main thing that's
> => gemini://gem.wiki-mirror.pizza/article.cgi?Tripping tripping
> me up is the fact that
> => gemini://gem.wiki-mirror.pizza/article.cgi?Wikipedia wikipedia
> articles are
> => gemini://gem.wiki-mirror.pizza/article.cgi?Writing written
> with a lot of inline links.
> 
> Which seems really clunky. The other option I'm considering is serving
> text/markdown, which I assume most Gemini clients would display as plain
> text. The problem then is that [links won't be clickable](http://dev/null).
> But maybe the mere existence of so much markdown content would incentivize
> some client author to add native support?

  Okay, here's a blog post I made a few days ago [1] and how it could be
rendered into a text/gemini.

  -spc

[1]	http://boston.conman.org/2020/04/23.1

===

Of course, after I point the finger to LibreSSL [1] for the memory leak, I
find the leak ? in my own code.

=> gemini://boston.conman.org/2020/04/22.1 [1]

Sigh.

Not knowing what else to do, I thought I would go through my TLS Lua module
[2] to make sure I didn't miss anything. That's when I noticed that I was
keeping a reference to a connection so that I can deal with the callbacks
[3] from libtls. I was expecting the __gc() method to clean things up, but
with a (non-weak) reference, that was never going to happen.

=> https://github.com/spc476/lua-conmanorg/blob/master/src/tls.c [2]
=> https://github.com/spc476/lua-conmanorg/blob/master/src/tls.c#L107 [3]

Yes, just because you are using a garbage collected language doesn't mean
you can't still have memory leaks.

I verified that, yes indeed, the references were being kept around after the
request was finished. It was then straightforward to fix the issue.

That's not to say that libtls still isn't leaking memory?it is, but (it
seems) only when you initialize it (which means it's not as bad). But I'll
know in a day or two if I fixed the leak. I hope that was it.

Link to individual message.

Travis Briggs <audiodude (a) gmail.com>

On Fri, May 15, 2020 at 3:23 PM Sean Conner <sean at conman.org> wrote:

>   Okay, here's a blog post I made a few days ago [1] and how it could be
> rendered into a text/gemini.
>

This looks like a really nice idea. Do you think it's a problem that for
long Wikipedia articles, the refs could go into the hundreds or even
thousands [1345]?

[1345] //dev/nul

-Travis

Link to individual message.

Ben <benulo (a) systemli.org>

I was actually thinking of this problem today with respect to converting 
content with in-line linking to Gemini format.

So far the best idea I could imagine is that during the conversion 
process the linked items are marked (for example, number in superscript) 
in reference to a list of links provided below, basically as a footnote 
or appendix either to the paragraph, section, or the entire page.

Maybe not ideal, but with a little experimentation this method could 
result in a format that's pleasing/readable and also navigable 
especially if the user can manipulate their client to jump to the link 
they want... for example, by numbers or tabbing through them or some 
other jump mechanism.

Or maybe Gemini should support in-line links of its own? It's already 
got some markdown elements in other aspects of page formatting.

On 5/16/20 2:40 AM, Travis Briggs wrote:
> The main thing that's tripping 
> <https://en.wikipedia.org/wiki/Psychedelic_experience> me up is the 
> fact that wikipedia <https://en.wikipedia.org> articles are written 
> <https://en.wikipedia.org/wiki/Writing> with a lot of inline links.

Link to individual message.

Ben <benulo (a) systemli.org>

We need to make our own wiki ;)

On 5/16/20 2:49 AM, colecmac at protonmail.com wrote:
> I'm guessing that most people here want the Geminispace to expand
> naturally

Link to individual message.

Sean Conner <sean (a) conman.org>

It was thus said that the Great Travis Briggs once stated:
> On Fri, May 15, 2020 at 3:23 PM Sean Conner <sean at conman.org> wrote:
> 
> >   Okay, here's a blog post I made a few days ago [1] and how it could be
> > rendered into a text/gemini.
> 
> This looks like a really nice idea. Do you think it's a problem that for
> long Wikipedia articles, the refs could go into the hundreds or even
> thousands [1345]?

  Don't know until I see it.  But I don't think it will be that bad, given
the links appear below the paragraph they're found in.

  -spc (Give it a try if you can ... )

Link to individual message.

Sean Conner <sean (a) conman.org>

It was thus said that the Great Ben once stated:
> Or maybe Gemini should support in-line links of its own? It's already 
> got some markdown elements in other aspects of page formatting.

  As I said before, check the archives [1] for previous discussions.  It
won't be hard to find them, as they make up about half the posts so far.

  -spc (It's a popular bike shedding topic ... )

[1]	https://lists.orbitalfox.eu/archives/gemini/

Link to individual message.

Brian Evans <b__m__e (a) mailfence.com>

This was done for gopher and can be found here:

gopher://gopherpedia.com:70/1

You may be able to pick up some tricks from what they did (and possibly improve upon it).
I have used gopherpedia a good amount and would probably use a gemini version. That said,
I do think it would be nice to build out original (non-web-mirroring) 
content. But that certainly
doesnt mean that we cant or should have some degree of both!


--?
Sent with https://mailfence.com
Secure and private email

Link to individual message.

Luke Emmet <luke.emmet (a) gmail.com>

Hmm this convention of using square bracket note-citations inline, 
followed by the URL. It seems nice and gemini-friendly.

How widely used is that? Can it be "recommended" as a good practice if 
you want to do inline links?

I'm thinking if it is a sufficiently established practice, clients can 
look for this pattern. And then when a suitably designed client if it 
sees this, could reverse the process to provide an inline clickable 
link, and then no need to duplicate the URL in the displayed content

e.g. Source of

"some text with a citation [1] more text etc

=> gemini://domain/path [1]

next paragraph"

is then simply rendered as

"some text with a citation [1] more text etc

next paragraph"

and when rendered by the client, the "[1]" is an active element.

Regards

  - Luke

On 15-May-2020 23:41, Sean Conner wrote:
> It was thus said that the Great Ben once stated:
>> Or maybe Gemini should support in-line links of its own? It's already
>> got some markdown elements in other aspects of page formatting.
>    As I said before, check the archives [1] for previous discussions.  It
> won't be hard to find them, as they make up about half the posts so far.
>
>    -spc (It's a popular bike shedding topic ... )
>
> [1]	https://lists.orbitalfox.eu/archives/gemini/
>

Link to individual message.

Travis Briggs <audiodude (a) gmail.com>

Thanks for the pointer, that's really helpful and cool and basically
exactly what I was hoping to accomplish.

Basically a "Simpsons did it" moment for me. Ah well.

But I've contacted the author and maybe I can get some insight into what
could be done differently the second time around.

Cheers,
-Travis

On Fri, May 15, 2020 at 3:55 PM Brian Evans <b__m__e at mailfence.com> wrote:

> This was done for gopher and can be found here:
>
> gopher://gopherpedia.com:70/1
>
> You may be able to pick up some tricks from what they did (and possibly
> improve upon it).
> I have used gopherpedia a good amount and would probably use a gemini
> version. That said,
> I do think it would be nice to build out original (non-web-mirroring)
> content. But that certainly
> doesnt mean that we cant or should have some degree of both!
>
>
> --
> Sent with https://mailfence.com
> Secure and private email
>

Link to individual message.

solderpunk <solderpunk (a) SDF.ORG>

On Sat, May 16, 2020 at 02:55:25AM +0430, Ben wrote:
 
> Or maybe Gemini should support in-line links of its own? It's already got
> some markdown elements in other aspects of page formatting.

I'm really not keen on adding in-line links.  It would significantly
increase the effort involved in parsing text/gemini documents - instead
of peeking at the first 3 characters of a line and then being able to
handle it, every line has to be split up into words and each word tested
individually.

Yes, there are e.g. Markdown libraries out there which will do this, but
they're only usable if we adopt Markdown wholesale, which brings in a
whole load of other problems.  Not to mention that almost none of those
libraries do anything other than convert to HTML, etc.).

The current text/gemini definition was very carefully and painstakingly
arrived at after a lot of discussion.  I'm not saying it's perfect, but
I think any future changes to it will be very minor, and something like
a switch to in-line links is pretty well off-the-table at this point.

I acknowledge that when you are used to being able to link in-line, not
being able to feels weird and different and like you are forced into
clunky ways of formatting your text.  But the one-line-per-link format
has real advantages, aside from extreme simplicity of parsing of
presentation.  It makes links extremely easy to find, and it encourages
neat, well-organised layout where links are put into related lists.  It
works very well in particular for "landing pages" - text/gemini as a
whole does, actually - the combination of headers, subheaders and
sub-subheaders via #, ## and ###, with one-line-per-link, strongly
encourages you to structure your site neatly and to layout the entire
structure on the front page so that navigation is an absolute breeze.
Chalk and cheese, when compared to modern websites.

For content it doesn't always work as well, but before deciding that it
just cannot work I propose we make a concerted effort to develop new
writing habits/conventions which work with it as well as possible.

Cheers,
Solderpunk

Link to individual message.

solderpunk <solderpunk (a) SDF.ORG>

On Sat, May 16, 2020 at 12:09:05AM +0100, Luke Emmet wrote:
 
> How widely used is that? Can it be "recommended" as a good practice if you
> want to do inline links?

It's extremely widely used on Gopher, where you can't do any actual
linking from within content pages (without resorting to ugly hacks
that work well in practice but are definitely a violation of the spirit
of the protocol).

Cheers,
Solderpunk

Link to individual message.

Luke Emmet <luke.emmet (a) gmail.com>


On 17-May-2020 09:40, solderpunk wrote:
> On Sat, May 16, 2020 at 02:55:25AM +0430, Ben wrote:
>
>> Or maybe Gemini should support in-line links of its own? It's already got
>> some markdown elements in other aspects of page formatting.
> I'm really not keen on adding in-line links.  It would significantly
> increase the effort involved in parsing text/gemini documents - instead
> of peeking at the first 3 characters of a line and then being able to
> handle it, every line has to be split up into words and each word tested
> individually.

I think certain common idioms can be implemented by a client as a UI 
pretification, for example the frequently seen square bracket citations 
could be expanded [like this] or this [1]. The community could encourage 
this idiom as it is very gemini-friendly.

=> scheme://domain/path1 [like this]
=> scheme://domain/path2 [1]

I'm working on a UI playground to explore some of these ideas, among 
others, hope to have something to show soon.

  - Luke

Link to individual message.

Jason McBrayer <jmcbray (a) carcosa.net>

Luke Emmet <luke.emmet at gmail.com> writes:

> I think certain common idioms can be implemented by a client as a UI
> pretification, for example the frequently seen square bracket
> citations could be expanded [like this] or this [1]. The community
> could encourage this idiom as it is very gemini-friendly.
>
> => scheme://domain/path1 [like this]
> => scheme://domain/path2 [1]

I think something like this as a "best practice" is our best substitute
for inline links. I'd add a short link description after the bracketed
expression in the link line, like this[3]

=> gemini://domain/path3     [1] Further discussion of link formats

I had tried things like *, ?, ?, as in print, but they're not that great
visually, especially without the ability to superscript them.

-- 
+-----------------------------------------------------------+  
| Jason F. McBrayer                    jmcbray at carcosa.net  |  
| If someone conquers a thousand times a thousand others in |  
| battle, and someone else conquers himself, the latter one |  
| is the greatest of all conquerors.  --- The Dhammapada    |

Link to individual message.

jan6@tilde.ninja <jan6 (a) tilde.ninja>

May 18, 2020 12:41 AM, "Jason McBrayer" <jmcbray at carcosa.net> wrote:
> => gemini://domain/path3 [1] Further discussion of link formats
> 
> I had tried things like *, ?, ?, as in print, but they're not that great
> visually, especially without the ability to superscript them.

technically you can use unicode with superscript block, like: x???? and whatnot,
but that's...probably not the best way to go about it

I like your format the best of the ones I've seen so far

Link to individual message.

Ben <benulo (a) systemli.org>

I for one am Unicode-positive... it's actually one of the main reasons I 
chose to start using Gemini in the first place, so I could create 
pages/documents that use it. They might not render in old Unix 
terminals, but nowadays even my Linux desktop has support for crazy 
things like emoji that I now see even while I'm using irssi. We just 
have to embrace it.

On 5/18/20 2:18 AM, jan6 at tilde.ninja wrote:
> but that's...probably not the best way to go about it

-- 
gemini://kwiecien.us/

Link to individual message.

---

Previous Thread: ANN: SecretShop 1.0.0 and saintnet.tech

Next Thread: Elpher 2.7.0: the text/gemini update