💾 Archived View for zaibatsu.circumlunar.space › ~visiblink › phlog › 20190617 captured on 2024-07-09 at 04:24:03.

View Raw

More Information

⬅️ Previous capture (2021-12-03)

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

I was scouring a cement wall with a 
wire brush yesterday. You have a lot 
of time to think when you're getting 
concrete ready for painting.

So I've been thinking about Gopher 2.0 
/ HTML 0.5. If I understand correctly, 
the consensus among the proponents of 
the liminal protocol seems to be that 
the it would have lesser capabilities 
than the web, but more than gopher. I 
think it's reasonable to state that 
this seems to be a universal 
assumption.

If that's the case, then the new 
protocol might be viewed as both:

1. Gopher++: an enhanced version of 
gopher with links, embedded media, 
other new features, and improved 
syntax.

2. Web-: a stripped down version of 
the web, designed to prevent bad 
practices, like tracking and traffic 
filtering/shaping (just to name a 
couple).

I know that seems obvious, but I think 
how you conceptualize the proposed 
alternative impacts how you react (and 
some people are feeling it kind of 
viscerally!).

At first, because of the nature of 
some of the discussion, I considered 
the proposals to be focussed on the 
creation of a kind of Gopher++ 
(Gopher+ is already taken, so I'm 
ad-libbing here). I hate that idea 
because I like Gopher (it's lovely to 
inhabit an online world without links 
and images), and because the idea of 
Gopher++ holds out the possibility of 
dividing the existing community.

However, when I think about the 
proposal as Web-, I like it a lot. 
From that perspective, it's a 
replacement for the web, not gopher. 
Yet if viewed as web-, we do have to 
deal with cat's contention that we can 
do all of the things we want to with 
the web already. That's true. The 
problem is that we cannot -- as a 
community -- constrain the overall 
impact of the practices we don't want 
imposed on us on the web.

As users, we can block the 
anti-features of the web, and many of 
us do. But when the majority around us 
are exposed to those anti-features, 
they pattern the collective mentality 
and the social interactions that 
predominate in the environment. A 
simplistic example might involve me 
visiting reddit with images turned 
off. Yet everyone else is exposed to 
and considering those images. 
Likewise, I can peruse the web with a 
pi-hole and ad-blocker, but the entire 
web has been shaped by what I'm 
avoiding, including my own experience 
of it.

A new protocol -- if it is to be 
designed -- should start from the 
premise that it is a 'social purpose' 
protocol, and define standards based 
on a shared vision of a good online 
environment rather than technical 
possibilities or individual gain.

One of the real problems with the web 
is that its capabilities are being 
(and have long been) reshaped by 
private corporations. Their 
non-standards-based 'features' are 
added into browsers as unique 
additions to html, and then frequently 
adopted into the html standard 
afterwards. Some of you may remember 
the "Best viewed in Netscape" and 
"Best viewed in IE" labels on sites in 
the 90s/00s. If not, you surely notice 
the "works best with Chrome" messages 
peppered across Google's sites these 
days.

Could Web- and it's RFC remain in the 
hands of a community? I think so -- 
and I think it would be an interesting 
exercise to design the standards. 
Agreeing on limits, however, would not 
be easy. Many of the abused features 
of the web undoubtedly started out as 
useful enhancements.

I can imagine the designers of the 
cookie thinking that it would be a 
wonderful thing to tailor a website to 
each visitor -- without considering 
that the stored information could be 
exploited in nefarious ways. Mind you, 
I know nothing of the history of 
cookies. I could be completely wrong 
about their origins. But it's 
conceivable that they might have been 
created to serve a useful, 
non-exploitative purpose.

In any case, I think it very likely 
that there would be great debates 
between the enhancers and privacy 
advocates in defining the protocol, 
but the key thing would be to have the 
protocol remain in the hands of the 
community, with it's purpose being 
shaped by a set of elucidated 
principles respecting user interests.

That being said, I have a 
recommendation for the new protocol, 
intended to prevent private entities 
or random individuals from introducing 
non-standards-based feature-creep:

It should block all clients that add 
non-protocol features. 

I don't know how this would be done on 
a technical basis, but I think it 
should be done, given what we know 
about how standards become abused and 
broken.