💾 Archived View for gemi.dev › gemini-mailing-list › 000548.gmi captured on 2023-12-28 at 15:49:02. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2023-11-04)

🚧 View Differences

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

[tech] the duplo protocol

1. Petite Abeille (petite.abeille (a) gmail.com)

Duplo: We are from the planet Duplo, and we're here to destroy you. 
Emmet: Oh, man.
-- The LEGO Movie, 2014


Gemini is build around 4 major blocks: TLS, URI, MIME,  and UTF8. As 
standard as they come. The bread and butter of the modern internet.

Furthermore, the gemini protocol and the text/gemini format are meant to 
be easy to assemble -both in term of implementations, and content 
creation- given the stated familiarity of the building blocks. 

Finally, the entire construct is meant to be immutable and immovable. All 
i's are dotted and all t's are crossed. No change shall happen. Ever. This is the way.

Everything Is AWESOME!!!
-- The LEGO Movie, 2014


Turns out, each of the blocks have their own share of idiosyncrasies, 
ambiguities, and, well, complications:

TLS:  overhead, complexity, support, tofu, no tofu, how to tofu, what is 
tofu?, seitan?, noise? etc, etc, etc... 
URI: URL vs. URI vs. IRI vs. URN vs... data:... userinfo... segment... 
query... reserved characters... encoding... transliteration... dns... etc, etc, etc...
MIME: to size or not to size? extendable or not? parameters or not? Postel or not?
UTF8: Unicode! Normalization! Validation! Internationalization! 
Globalization! IRI! IDN! Emoji! Punycode! Would rather use ASCII in 
practice, considering. Be the rest of the world damned.

"In theory there is no difference between theory and practice - in practice there is" 
-- Yogi Berra, allegedly


Then there is the commune. Where contrarian/questioning views are frown 
upon, borderline rude, if not outright insulting.  This is meant as a safe 
space; with trigger warning, cancel culture, and all.

The sum of all these technical realities and social dynamics make it at 
time difficult to communicate constructively about gemini's truths.

Just saying.

Link to individual message.

2. Sean Conner (sean (a) conman.org)

It was thus said that the Great Petite Abeille once stated:
> 
> Turns out, each of the blocks have their own share of idiosyncrasies,
> ambiguities, and, well, complications:
> 
> TLS:  overhead, complexity, support, tofu, no tofu, how to tofu, what is
> tofu?, seitan?, noise? etc, etc, etc...

> URI: URL vs. URI vs. IRI vs. URN vs... data:... userinfo... segment...
> query... reserved characters... encoding... transliteration... dns... etc,
> etc, etc...

> MIME: to size or not to size? extendable or not? parameters or not? Postel
> or not?

> UTF8: Unicode! Normalization! Validation! Internationalization!
> Globalization! IRI! IDN! Emoji! Punycode! Would rather use ASCII in
> practice, considering. Be the rest of the world damned.
> 
> Just saying.

  Okay.  Now what?

  Remove TLS, URIs, MIME and UTF-8 and you are basically left with gopher
(that's down the hall to the right, by the way).

  -spc

Link to individual message.

3. Petite Abeille (petite.abeille (a) gmail.com)



> On Dec 20, 2020, at 07:01, Sean Conner <sean at conman.org> wrote:
> 
> (that's down the hall to the right, by the way).

where do you live? your place sounds awesome.

Link to individual message.

4. Sean Conner (sean (a) conman.org)

It was thus said that the Great Petite Abeille once stated:
> 
> 
> > On Dec 20, 2020, at 07:01, Sean Conner <sean at conman.org> wrote:
> > 
> > (that's down the hall to the right, by the way).
> 
> where do you live? your place sounds awesome.

  Right across the street from the Argument Clinic, which is next door to
the Ministry of Silly Walks.

  -spc (You can't miss it---there's a cheese shop on the corner, right next
	to a pet shop)

Link to individual message.

5. Petite Abeille (petite.abeille (a) gmail.com)



> On Dec 20, 2020, at 07:13, Sean Conner <sean at conman.org> wrote:
> 
>  -spc (You can't miss it---there's a cheese shop on the corner, right next
> 	to a pet shop)

:)

Link to individual message.

6. Petite Abeille (petite.abeille (a) gmail.com)



> On Dec 20, 2020, at 07:01, Sean Conner <sean at conman.org> wrote:
> 
>> Just saying.
> 
>  Okay.  Now what?

[NPR STYLE TRIGGER WARNING: THE FOLLOWING CONTENT MAY OFFEND SOME]

At a more practical level: gemini must man up. 

The specification is more akin to an animal rights activist manifesto than 
a formal protocol. Daniel Stenberg couldn't help but laugh at it in its 
current shape. Bad omen.

Stephane Bortzmeyer brought up good points about the sorry state of 
affairs of the protocol governance. Don't ignore him.

People are overwhelmed by all these discussions because the place is utter 
chaos. Organization wise. 

No amount of groupthink will get gemini out of this. A protocol is meant 
to be technical, not poetic..

Link to individual message.

7. ew.gemini (ew.gemini (a) nassur.net)

Hello,

Petite Abeille writes:

> ... Daniel Stenberg couldn't help but laugh at it [the gemini
> protocol specification] in its current shape. ...

Would you (or anyone else or several?) be able to /start/ writing
it down in a form that would make Daniel Stenberg smile?

Cheers,
Erich

-- 
Keep it simple!

Link to individual message.

8. Stephane Bortzmeyer (stephane (a) sources.org)

On Sun, Dec 20, 2020 at 07:42:04AM +0100,
 Petite Abeille <petite.abeille at gmail.com> wrote 
 a message of 26 lines which said:

> The specification is more akin to an animal rights activist
> manifesto than a formal protocol.

Could you be more specific? The specification is not perfect (as the
discussion on IDN and IRI demonstrated) but, as an author of a few
specs and a reader of many more, I can say that it is certainly not
ridiculous. I wrote a client just from it, without reading other
clients' source code, something which is not possible with other
specs.

> Daniel Stenberg couldn't help but laugh at it in its current shape.

Being a reader of the curl mailing list, I don't think it is a good
assessment. Daniel criticized the lack of registration of the scheme
but he mostly criticized actual code (TOFU by default, and the lack of
test cases), not the specification.

> Stephane Bortzmeyer brought up good points about the sorry state of
> affairs of the protocol governance. Don't ignore him.

But I can add that I don't have a ready-made solution. True, I wish
there were a clearer way of discussing, approving and rejecting
changes in the spec, but it is not obvious how to do so. I certainly
don't advocate a formal process ("Gemini Foundation" or "Gemini
Consortium") right now, the Gemini system being too recent an, IMHO,
not done yet.

> People are overwhelmed by all these discussions because the place is
> utter chaos.

Wait until you've been in an IETF working group :-)

> No amount of groupthink will get gemini out of this. A protocol is
> meant to be technical, not poetic..

But a protocol also have goals, and goals are political, not
technical. We already have HTTP, which works fine. If we spend time on
Gemini, it is not because HTTP is technically broken (it is not) but
because we have goals that we find important (simplicity, lightness,
power consumption, reject of tracking, no distractions, etc) and which
are not reachable with HTTP.

Link to individual message.

9. Petite Abeille (petite.abeille (a) gmail.com)



> On Dec 20, 2020, at 09:19, ew.gemini <ew.gemini at nassur.net> wrote:
> 
> Would you (or anyone else or several?) be able to /start/ writing
> it down in a form that would make Daniel Stenberg smile?

Actually, this has prompted Daniel to somewhat formalize what to expect 
from new protocols wishing to be included in curl:

https://curl.se/mail/lib-2020-11/0088.html
https://github.com/curl/curl/pull/6263

Link to individual message.

---

Previous Thread: [tech] /.well-known/capability ehlo

Next Thread: [Spec] Spec (un)freezes and the spec's future