The tilde.town ~ring:
tilde.town/~datagrok belongs to Michael F. Lamb.
My personal website is datagrok.org. I am also a member at protocol.club: http://protocol.club/~datagrok and I am @datagrok on Twitter.
This page style is an experiment in broadsheet-style web viewing, intended to avoid acres of margin around a single column of text. You must scroll right, not down. Unfortunately this is cumbersome for anyone:
In future revisions, I plan to address these problems by:
One piece of my Big Project, Opinions, implements a distributed network of metrics. In one sentence: imagine an OKCupid-style similarity-comparison tool, except it is not limited to personality-matching for dating, and it is decentralized for all to use, not locked up under a single domain.
With a distributed network of metrics across an open protocol, one may dynamically form interest groups of like-minded people without having to declare in what manner they are similar. Like the Advogato trust metric, one may calculate using Opinions data which sub-graphs are toxic and to be avoided.
Opinions is similar to the various twitter blockbots in this way, except we address the concerns up front at the protocol level:
Opinions provides a framework upon which one might build online communities that can simultaneously cater to diametrically opposed people and all of their friends without those groups ever having to be exposed to one another.
One hurdle with building such a peer-to-peer network of metrics is that to rely only on transitive weighted trust to measure the network’s opinion of some unknown subject means that people in disconnected sub-graphs might never be discovered.
In other words: your opinions might align very closely, and indeed you might become fast and best friends with someone on the other side of the country, but if you and they share no common friends you might never learn of their existence.
So, to connect disconnected ends of the nodegraph, we might take advantage of federated services. A federated service could provide many conveniences: connecting strangers, always-on presence, offload computation of network opinion, to name a few. These services might be valuable enough to charge a premium or sell advertising. However, since the underlying data and protocol is completely free and open, those services cannot perform “lock in” and capture your data, or prevent you from going to a competing service that performs better.
I wonder if the various tilde-sites might become the first hosting location for federated processing services for an Opinions distributed network of metrics?
I take for granted that tilde.town, tilde.club, and friends are ephemeral and may disappear at any time.
I treat my space as if it will certainly disappear someday, with neither warning nor backup. I ensure that any content that I place here, if I spend more than a few minutes to create it, has a canonical source, safe and secure in my archival storage at home, easily re-hostable elsewhere at a moment’s notice. It is a free, just-for-fun, temporary sandbox.
Do all tilde users have the same expectations?
I don’t assume the same about my Dreamhost account. I know they have incremental backups, and my ability to access to those backups is part of the reason their service costs money. I still keep local copies out of paranoia, but I expect that if their server crashes, except for incredibly rare and catastrophic situations, they can get my website back.
Might someone who is familiar with traditional paid webhosting incorrectly assume that it is possible to recover their tilde site in a disaster situation? (Disasters can include an accidental “delete everything” by an admin, an accidental decommission of the node in amazon, a simple hack or break-in, or a user accidentally deleting their own work.)
Might users not even think to consider this question until it becomes a problem?
Maybe not everyone is as emotionally invested in preservation of data as I am, or finds the irrevocable destruction of digital data as upsetting as I do. After all, in the age of Twitter and Facebook all our old thoughts just scroll away into nothingness. Twitter and Facebook could elect to stop providing archives, or limit them to a week, with no notice, if they felt like it.
I can’t find many assertions or warnings about this characteristic of tilde sites. On tilde.club/~ford: “There’s no plan, no guarantees.” Nothing to be found on tilde.town.
So, I wonder:
Do hobby micro-hosting maintainers need to do more to set the expectations of their users, with regard to data backup and data loss?
Could someone (me?) build an easy-to-apply puppet module that provides rudimentary server-side backup services, maybe only for users that request it?
Or, should such formal “hosting service”-style assurances be actively avoided, to prevent tilde sites from becoming real work for their generous maintainers?
I prefer hand-written <table>ed early HTML sites full of content and <3 to ghost town wordpress installs with one “i should blog more” post.— vilmibm shaksfrpease (@nate_smith) November 12, 2014
I have a long, multifaceted rant, festering for fourteen years, about why I hate blogs. I was reminded of one small piece of this rant today.
Around the year 2000 or so, the Internet seems to have decided as a whole that it was okay to give up on organization, give up treating our websites as gardens to be tended and instead begin treating them as dumps to be filled. Back then the tools available to allow Normal People to publish their thoughts online were painful, complicated, and opaque, so who could blame us? Gardening is hard work.
“I just want to share this little thought with my friends without interrupting them with an e-mail,” we all said in unison, and inserted the thought at the top of the existing HTML file that was there on our website already. Just as I have been with this page. The web log was born.
Then, websites became no longer art installations, archives, or gardens to be tended. They became periodicals, insisting to their authors that if a steady stream of fresh content were not supplied, they would rapidly decrease in value. They would be an embarrassment. One must keep talking, no matter how inane, no matter how disorganized, no matter how repetitive. I wrote about this before, I think? No matter, I can’t find it now.
A constant supply of new thoughts meant blogs grew very long. Old thoughts were moved to successive pages, or archive pages. Nobody has the time to step carefully through those, especially considering how uninteresting most of it is. We’re lucky if Google even sees those. They sublimate into the Internet æther.
Have a thought or an announcement that is a bit persistent, that you want more people to see? It must now be repeated, at different times of the day, to stay near the top of the stack.
An organized website is an investment, the currency of ideas into the real estate of knowledge. The more you invest, the better the home you build. A blog is renting a motel room and moving on the next day.
So: one reason why blogs are terrible, and why tilde sites are great: tilde sites are encouraging people to explore the notion that they can just write something, stick it on the internet, and leave it. Put other stuff up around it. Make a collage, or a garden, or a pile. Not insisting: you are a terrible person for not feeding the gaping maw of the blog frequently enough.
And if we didn’t have reason enough to hate them, realize: Twitter is a blog. Facebook is a blog. A disorganized, uncategorized serialization where older thoughts, no matter how relevant or timeless, erode into the infinite past (perhaps to be deleted–you don’t know!–at the company’s whim.)
How to create a blog: 1. collect all the pages in your carefully-organized website 2. sort by date 3. concatenate 4. delete originals— Michael F. Lamb (@datagrok) June 30, 2014
The tilde-site fad has shown that lots of people are motivated to try out weird old difficult technology to escape Facebook and Twitter. This makes me hopeful that they might try out my weird new difficult technology Dialogue to escape Facebook and Twitter.
I’m still trying to distill what it is about the tilde-site fad that makes it compelling, so that I might import some of those lessons into my Huge Important Project, Dialogue.
I wasn’t motivated to use the
irssi client on the tilde.town box to connect to its tiny internal-only (at the time) IRC server. Instead, I set up a tunnel to allow my comfortable native IRC client to connect to it. After having done so, I still wished for a more federated service, so that I might easily chat with similar groups on different tilde servers.
I hand-edited some HTML for maybe one day. Then, I set up a more convenient workflow, using the incredible pandoc tool on my local machine to build my tilde.town page from a local Markdown source file. That way, I won’t lose my words should the server someday spontaneously sublimate into the Internet æther. So much less painful, but now I’m missing the shared experience of “there’s
wall chats messing up my text editor.” Well, I wouln’t exactly say I’ve been missing it, Bob.
Evidently, I don’t have so much nostalgia for the early ’web that I’m willing to re-inflict the painful parts upon myself. But I still find myself motivated to participate. What should we keep, and what may we safely discard, from this re-examination of the early web?
Later, we got the #tildetown channel on freenode. This made the internal-only chatroom feel redundant, enough so that I stopped bothering with the tunnel, and just used irssi for a while. Suddenly, with two windows on my screen with entirely different interfaces dedicated to multiple communication channels for the same topic on top of the other channels I already haphazardly monitor, I found myself exceeding my cognitive limits and closed both. Don’t know what might be a solution for that. If I don’t keep up with the backlog in an channel I’m lurking in, I feel bad, like I’m losing important data. On the other hand I wouldn’t want to merge all channels into one, because it would become difficult to follow threads of conversation. User-defined client-side channel-merging?
Thinking about multi-tribe membership reminds me a bit of Google Plus “circles.”
People like gopher and yearn for its simplicity. Must ensure Dialogue, from some perspective, appears to be simple. Advertise it as “like gopher” but with new features?
One or two friends are enough reason to join a network.
~silver’s Pumpkin Spice anti-Facebook social network seems to assume that in order to be successful, it must be a silo buster–it must achieve feature parity with its competitors so that everyone on Facebook can move off of Facebook and defeat the network effect. I don’t think that’s necessarily the case. Facebook will probably cling to life for a very long time after better social networks arise; better social networks are likely to be populated–like tilde sites–by small groups of early adopters first.
Tribal membership (neighborhoods, sports teams, homeroom class) are fun and compelling, even if there’s no “real” shared commonality, even if there’s no real competition with other tribes.
As tribes get larger, it is harder to know all the members, so they become less compelling, or split into smaller tribes, or are abandoned for smaller tribes. Matters of consensus converge to an unpleasant median that marginalizes interesting minority or niche opinion. I’m not sure I would want to join the thousand folks over at tilde.club, especially since they seem to be welcoming all comers with not much vetting. What kind of weirdos could be secretly lurking, hiding in the crowd, in a group that large? Will tilde.town face the same troubles as its userbase grows?
This quote is something I can relate to, but at the same time it is a bit scary:
I don’t want more people in here because this space feels Safe to me, on a web that feels increasingly unsafe for women. I fear that will be lost if we have 10,000 tildes. –~megnut 2014-10-10
Decades on the Internet have developed in me an intuition that this is a false sense of security. If the tilde sites are safe spaces, it is by accident, not by design, and there’s nothing in place except poorly-scalable promises of moderation to ensure they stay that way.
I am hypercritical of Twitter and Facebook for not predicting and preemptively addressing the problem of toxic users. Am I giving tilde sites a pass? And is that okay, because they’re free experiments, not billion-dollar businesses?
For serious, how are engineers designing social networks without treating people you want to avoid as a first class issue?— Valerie Aurora (@vaurorapub) November 6, 2014
I said on Twitter,
I feel like http://t.co/TaEUhE3Doi and friends could be the thing that introduces multi-user Unix to a generation who grew up on Minecraft.— Michael F. Lamb (@datagrok) October 12, 2014
If you've gotten over someone accidentally setting your Minecraft forest on fire as a kid you're ready to graduate to a shared Unix box— Michael F. Lamb (@datagrok) October 12, 2014
Allen Tan (@tealtan) tracked down several examples of Minecraft as sysadmin-training:
I read this very helpful explanation of Windows vs Unix OSs THEN noticed it was on the Minecraft forum!? http://t.co/G0C56F3m— Dan Roddy (@danroddy) January 7, 2013
Some kid at airport: “I'm a moderator on a player server and destroying other people's houses isn't cool.” #nerdworldwins #Minecraft— Michelle Archer (@meesherbeans) April 24, 2013
You know your 9 year old kid is a nerd when you ask him how his day was & he explains how he setup a whitelist on his #minecraft server— Jim Rutherford (@jim_rutherford) May 15, 2012
! RT @diana_clarke: My kid brought home his year book. The back page is scribbled with Minecraft server IPs from his classmates. Grade two.— CoriDrew (@coridrew) June 22, 2014
My daughter moves, her friend makes a replica of her old house on her Minecraft server. Sometimes the 21st century is pretty cool.— Paul Mather (@paulmather007) August 1, 2014
My 10 year old daughter just announced she’s opening a head shop. On a Minecraft server. Where you can buy heads of famous YouTubers.— brian del vecchio (@Hybernaut) August 15, 2013
@dgwbirch My 12yo daughter asked why she needed Facebook when she had Minecraft! She meets her friends on her server.— Tom Standage (@tomstandage) April 10, 2013
Daughter is loving the power of being system op on minecraft server. I've created a monster…. out of an 8 yr old. #computercrack— The Motley Pugg (@Puggpaw) December 10, 2011
tilde.club is doing what I had hoped it would: spawning similar clones. They’re not carbon-copies, either: they use different webservers, operating systems, and cloud providers, but with the shared ideal of a fun multi-user environment for publishing. I hope that these done-on-a-whim toys will do something interesting and clever with each other to provide a tapestry of federated services.
The thing that I find most striking about this is that tilde.club and its ilk are not novel. Cheap shared Unix hosting has existed for a long time. I have a DreamHost account; I have a shell there; I can use
wall to send messages to others on the same system. Anybody can go set up their own AWS box, the same as the maintainers of these boxes have. Why haven’t all of these thousands of people built their little website on DreamHost, or GitHub Pages, or any of the many other hosting services on offer?
There must be a crucial, compelling, perceptive difference between that and this, and I’d love to identify explicitly what it is. The wait-list for an account at tilde.club is already up to 5000. This is something that people are eager to do and participate in, but in a way that all the other cheap hosting services that are out there are failing to provide.
Some tilde.club workalikes include my friend vilmibm’s tilde.town (where this page is hosted,) tilde.farm, and tilde.camp.
There seems to be an element of nostalgia to the “tilde” URL, but it is by no means a necessity. (Tildes don’t even URL-encode well.) With nginx, the mapping is specified by regex, so the tilde itself is somewhat arbitrary. If sites were to abandon the literal tilde, would people still want to play?
Is there something also enjoyable in a URL that is not standalone like
http://datagrok.org, but describes your little space as part of a larger community of spaces like
http://tilde.town/~datagrok? The very URL proclaims that you are a member of a larger tribe. (But again, if this is the case, why has it not been popular until now?)
It was not long ago at all that the various ISPs that I used offered web-space in a tilde-prefixed area under their own domain. That feels unlike tilde.club, because either there was no shell access or the other people on the server were professionals; there was no shared idea that “this place may be ephemeral and only exists for fun.”
I haven’t abandoned my concerns (mentioned in an earlier text) that this experiment isn’t doing enough to prepare ahead-of-time for toxic people, data loss, and being a single-point-of-failure. (Not that it intended to.) But I’m sure some who have joined this fad are doing so with the intention of escaping from Twitter or Facebook or their giant-corporation-hosted blog, to build something better. I like this experiment, and I think we can learn from it, but I still believe it is fundamentally flawed for building the “next” or “indie” web.
That said, I think this is tapping into something very interesting about the nature of fun and play and experiment in publishing. It’s providing an opportunity to go back and re-learn the lessons of the past two decades of web hosting, with foreknowledge of the kind of user-hostile systems that they can become if we let them (Twitter abuse, Facebook and Google nymwars). Maybe this will reinvigorate the discussion around hypertext and its uses, or inspire people to re-imagine what the web could be.
I have done some of this imagining myself; see my work-in-progress linked below.
I like the impetus behind tilde.club, and I’m happy to see how popular it is already.
ello is/was a cruel joke that preyed on our wanting to escape from the centralized, venture-capital funded identity markets like Facebook, while being simply more of the same. Albeit with a crappy (and thus indie-flavored) user interface.
Update (2015): When I first wrote this, I, like Aral, was incensed that Ello who positioned themselves as an anti-Facebook appeared to have sold their users for venture capital. But some time after that, Ello became a public benefit corporation, which greatly limits the evil they could possibly do (about as much as is possible without going full nonprofit.) Now, I believe their motivations are commendable, despite the service being centralized and lacking filters.
Tilde.club is a fun expression of our disenchantment, but if we begin our “rethink everything” movement with the same old:
then we’re heading full-speed down the same well-worn road already.
This isn’t a critique of tilde.club. I’m happy it exists and I like what people are doing with it. I just want more.
We may have nostalgia for the days before Blogger dot com when we all stayed up late to build a simple website, learning HTML by using “view source.” But we need to go back further, reexamine more of our assumptions, and solve problems with forethought, not afterthought. The solutions need to be baked in to the free and open protocols that underpin everything else we do, not bolted on or offered as a monthly service by a greedy profiteer.
There are many approaches. Here’s mine, unfinished: