Hahaha, I built something just like this site once, just to see what would happen.
Wouldn't this be fun? No auth, no rules, no permanent database, just plain text that hangs out for a while (though people could reply to it - ours was more like a forum). The ones with replies jump to the top, everything else slowly falls off into the abyss where it can never be found again.
Literally had people posting ASCII-rendered child pornography within 24 hours
Apart from the last line, that site design sounds like a lot of fun. I'm guessing you went "okay this isn't going to work" and killed it - or is it still quietly buried somewhere?
Also, did you keep the code? Might be a fun startpoint for others to springboard off of.
We (and by we, I mean, I) launched "Oh By"[1] last year which is very, very much like what is going on here, but with deliberately human-readable and recognizable codes ("Oh By Codes") ....
Do you have any comments ?
Of course we had pastebin very much in mind - not as a model to follow but as an analogous service in some ways. Our message is limited to 4096 characters so it isn't quite as useful for large docs ... but it's much easier to pass around the "code" chalked on the street ...
I don't think I have that many interesting stories, it was just the time involved with dealing with emails from irate individuals, companies, and (occasionally) law enforcement agencies!
But in an AMA, it's the audience that asks interesting questions, with you discovering that your stories are more interesting than you originally thought they would be!
God i feel you. Everytime a pastebin close we get a surge of dox / warez / child porn activity on http://0bin.net. Since by design we can't monitor it we just get notified after the fact by concerned citizens mailing us. One month our mailbox is empty, the other one it's report day every day.
How confident are you in your legal shield that, although you know your service is being used for child porn, you're claiming not to be responsible for it?
Edit: I see you're behind whoisguard but hosted in Bulgaria, which is probably uninterested enough in chasing you.
I can confirm that. I made https://pastery.net, a pastebin (THE BEST ONE), and I looked a few weeks later to find that it had a few hundred thousand pastes. I was very happy to see that people were using it that much, until I realized that most of it was spam.
I wrote some heuristics to delete spam pastes, and only about 1,000 pastes were left, about half of which were spam still :(
I don't even know why they do this. They just do it.
Require some complex js codes workflow to post using websocket or something and most automated spam goes away. They are mass mindless bot, not targeted spammers.
That's a good idea, but most of the functionality of the site comes from its API. Then again, it's easy to require a key and a captcha to sign up, so your idea is good, thank you.
Cheaper for their scripts to indiscriminately spam than them manually work out which forms and public sites are indexed and will actually bring them value.
Nofollow links seem to help with getting the target indexed. So, the spammers create their second and third tiers of actual content on sites that do not use nofollow. However, they'll use anything with any kind of links to get those tiers crawled and indexed.
i guess thats what they did. funny, it seems they got penalized regardless somehow, because i started getting emails asking to please remove the spam they posted.
Edit: It seems some of them were trying to damage competitor's SEO ranking by posting irregular links. SEO is wild west.
If you're implying that the technically literate are either better or worse in this respect (or even just differently distributed), you might want to back that up with some arguments. Maybe you tried to say something else but otherwise I don't see the point of this remark.
Hi, nice idea! It went down though so I was thinking maybe you could build it strong but on the cheap too without needing a database and very little coding.
There is an approach I've used for years and my site never goes down with high traffic and I didn't use load balancers or anything fancy. Its very cheap to run and maintain.
1) Use Amazon S3 and Cloud Front for all your web pages, including your web app as well along with storing anything such as file uploads, etc.
2) Have a simple cheap $5 Digital Ocean Droplet (Node or Go) used only as an API.
3) JavaScript is hosted on your website front end that upon save for content then hits your API server just once to get a Just-In-Time signing request that authorizes the web browser to make an upload directly to S3 temporarily and signed to a specific namespace/bucket. This type of traffic happens very less than say viewing a web page along with it taking up very little processing and bandwidth so you really don't need to scale this.
4) The web browser's javascript literally creates the entire html page and uploads it directly to S3.
You can easily do image uploads as well or video too using this method and use the just-in-time signing method for those too.
Yes it is not secure if there is no signup etc but you can do crypto.generateRandom to put people in specific namespace path then show their URL after, controlling limited permissions and the location of where things are saved from your cheap API server.
Sure you could. The funny thing is, I've had this product on the market using this very concept of JIT signing, using the browser (and apps) to do all the work, didn't really put any effort into, only used for my clients as a CMS-as-a-service or CMSSAAS WTF? called kitgui.com, been on some department store brands using it for ecommece sites that i've put as a consequence of needing this on multiple backend platforms. Any hoo, never goes down, never had to do jack shit scaling or worry about it but I've completely neglected it so it looks like 2010 crap themes and diy web standards. Anybody want to partner up on it and make a bad-ass json cms that spreads to all devices? I'll gladly divvy % on current revenue.
For the record, firefoxd's static assets were offloaded to another server for an additional $10/month. $15/month is still not bad, but I feel like this comment is a bit misleading.
Too many publishing platforms sunrise and sunset. To use one that's not self-hosted, I really think one either needs to (1) charge money, and/or (2) offer guarantees like Posthaven or Svbtle, if you want users to publish good content on your platform.
Well, yes. But it has already proven staying power that puts it in rather a different class to "somebody's neat side project" in terms of expectation of survival.
I'd imagine for the time being you could use DO's resizing features to vertically scale your droplet to handle the load. Then at a later point, horizontally scale.
Yes, I'm just editorial (and until recently design). Ken is our sysadmin and, as it happens, got txt.fyi back on its feet today when HN blew it down! He's brilliant!
I HATE gray text. I haven't had a single book with gray on white, not a single newspaper with gray on white, what is this gray-on-white trend in the internet, I don't understand it. But that's illegible, that's an eyesore, and if I were to have to read it, I'd first go to the dev console and change the foreground colour.
>> Black on white? How often do you see that kind of contrast in real life?
I mean where these people live? Don't they read anything on paper? Don't they look ad ads on the road, ads on TV. God, is this a plot on our eyes?.
If it's well done, you don't immediately realize the text isn't exactly black. You'd have to have a closer look to find that out.
The paper color of good books isn't glaring white, either, but has a slight yellowish tint. FWIW, I think the background color of HN might be a nod to that.
I nearly snorted my morning coffee when I read:
> Shit's legible and gets your f...ing point across (if you had one instead of just 5mb pics of hipsters drinking coffee)
So good. For the record I'm a programmer, not a hipster. ;)
The problem with these sites is, they're hard to discover. Telegra.ph is the most awesome (rich text, widgets for youtube & co...), but can't be found when you type "richtext pastebin" on Google, because the welcome page is too simple, obvious, perfectly ergonomic, ultimately simplistic...
Aren't people who visit Hacker News already familiar with a ton of ways to publish anything they want instead of soon-to-cease-to-exist service that only serves raw text?
I say this not to be mean -- I understand where you're coming from, I myself have written a "pure plain text" service in the past and thought it would be revolutionary due to its simplicity.
I don't think its author thinks its revolutionary (at least it doesn't send off that kind of vibe). Just a neat personal project that some people might find useful.
They definitely are, but it's still an interesting, useful project.
As someone also building and running a "plain text" service, I know there's a lot you can do with such a simple product, if you care to. [0][1] It's great to have more services like this -- it's what the web is all about.
Really. Is THIS what the web is about? Just sharing text? Of course sharing text is part of what the web is about, but hey, how can we find this text? How can the author know what he has written? How can we be sure of who wrote it? How can we make sense of what text refers to what? Which of these texts is more worth reading?
Just in the field of text-sharing, there are tons of problems which are not being solved by these pure plain text services, while these are not new problems and other services are already trying to solve them in multiple different ways.
I meant "this" as in people creating simple tools and services for everyone. I meant your pure plain text service would've been a nice addition to the web. Everyone who creates things online brings their own ideas to the table, and that's what the web is about. (And of course it's good for other things.)
And certain services (like Write.as) keep track of all the posts you've written, and they can optionally can be tied to a pseudonym. I don't think that an assurance on who wrote what is always necessary. If the writing is good, the author doesn't necessarily matter. And anecdotally I think people who abuse anonymity online are in the minority -- most people use anonymity to share more candidly where they otherwise feel repressed.
And I think "worth" is best decided by the reader, and a human touch continues to be superior to any technical solutions we've come up with. Platforms who try to uniformly enforce some standard of "worth," in my opinion, risk alienating a large amount of people, and usually miss the entire standard they were going for in the first place.
But what other problems do you think exist with text publishing services? Or what do you think needs to change? How do we solve those things?
> I don't think that an assurance on who wrote what is always necessary
I'm not saying it is always necessary, but that it is a recurring problem that occur but txt.fyi doesn't care about.
> most people use anonymity to share more candidly where they otherwise feel repressed.
It doesn't matter if you share something anonymously in txt.fyi if nobody is ever going to find it.
> And I think "worth" is best decided by the reader, and a human touch continues to be superior to any technical solutions we've come up with.
You're proposing that we go read everything on the internet to pass our judgments on what is worth. That was the problem in first place. Categorization, ratings and reputation are some of the solutions already tried. I don't know if they've solved everything or if there is room for better solutions. What I know is that txt.fyi doesn't care about this problem.
> But what other problems do you think exist with text publishing services? Or what do you think needs to change? How do we solve those things?
I don't know. I just imagine that there are many problems we know of (such as the ones I listed) and the author of txt.fyi is not trying to solve any of these. Other problems may exist or come into existence, and it is up to platform developers to try to solve them too. There is of course room for pure plain text published anonymously, but the supply is already immense there. I don't think we should incentive new developers to write one more of these services (and worse: keep running it an improving it technically for a long time).
> Really. Is THIS what the web is about? Just sharing text?
Text is communication. Anything else is an increase of human efficiency at a cost of data. I wouldn't say the web's about text...but I would say that it's about communication, and that services like this clear away the extra frills to get to the essence and remind us all.
Looks great, but it's sort of unfortunate that the author doesn't seem to have seen the work on microformats[1] that the IndieWeb community has been working on.
Being able to pull microformats out of the posts to reconstruct feeds, automatically embed txt.fyi post summaries, syndicate to facebook/twitter/etc (with comments backfed in to it) is super nice for low-bandwidth sites like these, you can add a lot of useful value for little more a few extra HTML classes. The core of a lot of this, the "webmention" standard is a W3C recommendation, even.[2]
Maybe I'm missing something, but after a quick look through the first two links I have no idea what microformats are.
Even the about pages appear to be nothing but buzzwords and acronyms.
The idea of microformats is giving some known structure to much of the unstructured data used in pages across the web. The idea has a lot of potential, but I don't know how much they're used in practice. There are a few formats and microdata can be in the body of the html itself as attributes or separate as JSON-LD. Google supports microformats for use cases like contact information and restaurant reviews [1][2].
If they were used ubiquitously, there could be an obscene amount of power in essentially being able to run one SQL query across different data sets on multiple websites.
Maybe "vanilla as possible" is not desirable if it means that your markup is non-semantic. I like that Txt.fyi tries to be as simple as possible, but there are some UI anti-patterns like the textarea placeholder instead of a matching label.
It would be pretty easy to add the proper HTML elements like <header> and microformats.
I like the idea. I've used services like this before. I do wonder if monospaced font is really the best route. It makes anything longer than a couple hundreds words a pain to read.
Never realized how uncomfortable it is to read a paragraph of monospaced font text, even though I read 1000s of lines of code in monospaced font without any issues.
Maybe I'm in the minority, but I enjoy reading in monospace. And I certainly prefer it over a sans-serif font with font-weight set to 0.01 and poor contrast!
Incorrect. This uses storage on a server to store post data. When I first read that I thought this would be a distributed p2p platform of some sort.
The p2p web platform being worked on at OFTN makes it very easy to create a truly static version of this website (100% client-side p2p logic & data distribution) without a server database.
You will be liable for illegal content uploaded to your server (when notified of its existence), so you will be required to remove content from your server when you receive court orders/DMCA takedown requests/etc. With a proper p2p tracker-based system, you can simply forward legal proceedings to the ISP of infringing users, and avoid the legal time sink involved with this kind of site.
I see how you can call this independent (eg. not a big hosting company), but for me the independent web is people hosting their content and keeping control on it. The ideal case is everyone hosting on their own machine (the old WWW?)
The server at txt.fyi is taking too long to respond.
The site could be temporarily unavailable or too busy. Try again in a few moments.
If you are unable to load any pages, check your computer’s network connection.
If your computer or network is protected by a firewall or proxy, make sure that Firefox is permitted to access the Web.
I think half the "pastebin problem" has been solved by just avoiding setting up your own backing store (and frequently your own input step), and instead being a presentation-proxy for a more "raw" content-editing service, like Github Gists. For example: http://gist.io
There are a few services like that; what I'm still yearning for, though, is one that either
1. has a real design team focused on making its pages "read" well, like Medium does; or
2. does have accounts, to let you set up a custom theme for your pages. (Though you can be creative about this, for example by letting the Gist include a JSON/YAML/TOML/whatever properties file that specifies the theme, and by making the themes Github repos that the service just pulls and caches on first use rather than needing to own itself.)
A service that offered either of these would finally fulfill the promise of letting me separate "writing" from "blogging", such that I wouldn't really need a "blog", just a microblog containing links to my own posts.
So I had a go of this. I pasted in Yeats' Second Coming [0] (feels somewhat salient these days) to see how it looked [1]. For the most part it looks okay, but visually it adds too much space between lines and the gap between the two verses has been eliminated.
I guess what's happening here is the usual HTML thing of compressing all whitespaces. I would just be concerned that if a plain-text medium is being used then whitespace becomes more important for formatting and it should perhaps be preserved.
Legible is subjective. I personally find the font and spacing quite difficult to read actually. Conceptually reminds me of an old writing platform I used to use called QuietWrite. Loved it while it lasted.
These types are becoming more popular for people who don't really care to blog and just want to get something up quickly. I think pen.io were the ones who pioneered the idea of it.
So many others followed and I wrote something similar: https://mypost.io/
It allows for much more user control including easy HTML and CSS usage with the ability to set your own URL. It hasn't receive as much thrill as your product has on Hacker News, but it is being used worldwide.
I just want to chime in and say that I've had a look and a quick play with mypost.io and think it looks awesome. Like many of the links on this HN thread, but with way more feature.
As you are allowing HTML entry, how are you protecting against Javascript inside user created pages?
I appreciate your great feedback. I barely had any influence when writing it, but all I knew was that I wanted to be able to write HTML and CSS -- and help my visitors with learning the basics, without having to even sign up for an account. All HTML is allowed though iframes seem to be so/so -- sometimes they work, sometimes not.
There is certain Javascript that is allowed, but others I've had to disallow. The way I've been able to allow it and monitor it is through BBCode. Basically writing [script src=myjsfile.js][/script] or something like [script]alert('hey there');[/script] is how you can get Javascript on a post but it does go through a filter beforehand to scan for any potential harmful code. Trying to write javascript the original way doesn't usually work.
At one point, I had Javascript as a main feature in the Advanced Options section, but I've since removed it, as I saw its potential use for abuse. As more and more users are using it from around the world, it's kind of interesting to see the clever "hack codes" people come up with. For the most part, I've mostly filtered out all "dangerous" code, but it still arises occasionally as the Internet evolves and more people are clever.
Mainly with the hacks I've seen: people using it to redirect directly to their own spammy websites. People creating divs that block MyPost but show a message on top and you can't do anything but go to the person's website or link. I've been able to filter these types of codes and prevent them from being entered into a post. The thing with Javascript though, there are dozens of ways to write the same code.
I even had to build my own captcha as people learned how to automate the creation of mass-posts. Some Russian guy emailed me (it was in Russian, but the translator basically told me he was pissed off because I added the captcha), but I knew it was him who created about 2000 posts in less than an hour. They then learned how to somehow bypass the Google Recaptcha and so I ended up building my own, which fortunately, at the moment, has successfully stopped bots from being able to automatically post thousands of posts at a time.
I'd rather people use it for its true purpose: getting webpages up on the Internet in seconds; learning how easy it is to code, etc. than to have hundreds of thousands of "spam" posts on the website. So those have been my battles since creating it: fighting bots and fighting people who are coming up with clever ways to "hack it".
> and know the form of your voice is out there forever.
Lots of these types of sites usually have some limit. That is, if the page hasn't been visited in 6 months - out it goes. Curious what the time limits are.
Also, unable to test (site under heavy load) but is there an edit link that's made available after making a post? From the comments, that doesn't appear to be the case. So if I wanted to use it as a knowledge base for something, I wouldn't be able to keep updating it. I think this is the missing (basic) feature.
That's really the point, to be as minimalistic as possible. The less features something has the less points of failure are possible. If you need a certain feature, use something else.
Tried some ASCII art, but it appears multiple adjacent spaces get collapsed into a single space, meaning this isn't the platform I was hoping for for ASCII needs.
Plus, the line spacing eliminated the possibility for sane multiple-line ASCII.
Definitely hipster not true oldskool. I will live and dream...
I promise that so long as I can pay my domain name dues and maintain backups, the stuff posted to it will remain online e forever, even if posting new stuff has to end for whatever reason
You POST your uploaded doc to 'do.php', which renders into an html file and puts it on disk, which is served by Apache. That's about as simple as you can get. Let the web server do what it does best (serve files).
Since the site is already experiencing the HN hug of death, it makes me wonder how doable it would be to make something like this but backed by a P2P store, possibly verified by blockchain
Some markdown features aren't working, like code blocks and newlines (by ending a line in double-spaces). Also, unknown routes don't 404 nicely; they just hit the Apache default.
Seems like a cool idea, and the creator is pretty responsive on Twitter. I offered to assist with the infrastructure as a way of supporting the project.
