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