Link syntax, stagnation
-----------------------

Quick update before I go without any access to my gopherspace for 10
days:

Tomasino (now in Iceland!) has commented[1] in favour of the new =>
link syntax, and votes for the URL first ordering.  Germinal author
JFM/gcupc however, is strongly in favour[2] of URL last ordering.

Making the final decision on this is ridiculously difficult for me.
Having something as fundamental as the linking syntax being up in the
air is a big barrier to implementation and adoption of the protocol
and I feel like it should be settled soon.  My natural inclination is
to try to please everybody, but this is obviously going to be
impossible.  There are compelling arguments in favour of both
options.  I think the balance is slightly in favour of URL first, but
honestly I may just end up tossing a coin.  Neither option is
horrible, and I'm sure we'll all get used to whatever is chosen after
using it for a week.

JFM also comments on an apparent slowdown in Gemini activity.  It's
certainly true that *I've* slowed down a bit because I'm travelling a
lot this month, but in my occasional check ins I'm still perceiving an
amount of interest which surprises me and makes me nervous.  A Gemini
client library has been written in Guile Scheme[3], and people are
still occasionally asking me "What is Gemini???" on Mastodon.

To be honest, I think a little bit of slowdown at this point makes a
lot of sense - fundamental details of the protocol are still under
discussion and I think a lot of sensible people observing the whole
thing are likely to hold off on implementing things or setting up
servers until this stops being the case.  I know sloum is planning
to write a server once we have the details ironed out.  I also plan to
write a lot of stuff eventually which should help drum up interest.
I'm not bothered at all by people holding off on setting things up for
now.  The more things which are put up using the current speculative
specification, the greater the risk that they don't all get updated
once things are finalised, resulting in a fractured and incompatible
Geminispace.  That would be a bad thing.

Things have slowed down more than I expected them to when I wrote my
"slowing down" post, but there's nothing to be done about it, it's
just bad timing.  As I've said, I really didn't expect so much
interest in Gemini when I gave the thing a name.  At any rate, I
really hope that by the end of July there is a solid "core spec", such
that all the basics are locked in and people can safely begin to
really experiment.  I'll do my best to stick to that schedule.  I hope
people can be patient until then.

I will continue to solicit feedback and to try to build consensus, but
it's increasingly apparent to me that it will be impossible to please
everybody and that I probably need to be more assertive in just making
the decisions that I think are best.  A long, pseudo-bureaucratic
process that drags on forever is probably a bigger threat to interest
and activity around Gemini than is a finalised protocol that not
everybody agrees with in every small detail.  It's very clear that
lots of people want something *like* Gemini, and right now there is
just nothing else out there.  Anything non-awful in this space is
likely to see some use.  If it becomes apparent two years later that
we made some mistakes and something else comes along to replace it,
Gemini will still have been an extremely worthwhile undertaking.

[1] gopher://gopher.black:70/1/phlog/20190705-some-replies-from-iceland
[2] gemini://carcosa.net:1965//journal/20190705-link-syntax.gmi
[3] https://paste.sr.ht/%7Ejakob/a31b19708779e6b797fb9f00694be27f56d2e226