[tech] [spec] On extending gemini

Drew DeVault sir at cmpwn.com
Mon Feb 22 17:13:52 GMT 2021

On Mon Feb 22, 2021 at 12:07 PM EST, Pierre-Jean Baraud wrote:
> I am very new to the game (this is my first mail in this mailing list)
> and I never really participated in any "protocol ecosystems" in
> making.  Nevertheless, is there not a risk in that game that's the
> most popular implementation defines the rule and hence, the game is
> not being right, but being popular?

Yes, that's a risk indeed. I think it's a well-managed risk in Gemini
right now, we have a huge selection of client and server softwares to
choose from and no single actor has a dominant role. And unlike web
browsers, new Gemini software can be written and useful in the span of a
weekend, providing an easy escape hatch - but this will change if we're
not strict in limiting the scope of the protocol.

> That is very nice to clarify that. But it still raises the concern of
> such risk in future by an actor of similar size: will the stick not
> overcome any rational debate?

I do realize that srht.site has the potential to become sizable enough
to carry substantial influence on the direction of Gemini. However, I
don't want it to do so: srht.site should only be a boon to the
community, not a risk or a political tool. I pledge to not use it as a
platform for pushing any opinionated decisions which are not already
representative of the community consensus.

I can't say this for the next service that comes along, but hopefully it
helps to set the tone.

> Also, unrelated, but I have learnt a lot from this mailing list
> through these different debates and I think I start to have a better
> idea now of Gemini's philosophy thanks to you all (and I'm sure many
> of my first intuitions would have first been considered as
> anti-pattern / recreating the web)· This learning has been very
> organic, almost chaotic and only thanks to the people posting "wrong"
> initiatives here as the answer they receive served for me as
> guidelines as "Gemini philosophy".  Is there anywhere a more
> structured document, the equivalent of "unix philosophy" but for
> Gemini that would express the intention behind some of the
> "limitation"/restrictions of the platform that a newbie like me could
> refer to instead of learning "by chance" this through some response
> here and there (and sorry if it is an obvious link that I am not aware
> about) ?

The FAQ helps a lot:


But I think a lot of people don't read it, or perhaps forget about it,
because many of the proposals oft floated are clearly not in the spirit
of Gemini as laid down by the upstream documentation.

More information about the Gemini mailing list