Wikimedia telnet interface
meta.wikimedia.orgThe revision history of this page indicates this is an April fool's thing. Besides images, I might actually totally love using Wikipedia from the command line. It would fit in well with my general command line usage. The Tor feature is particularly cool.
(Just to be clear, this post is not a joke. I actually use a terminal all day for various tasks so it might fit in well.)
It might have been set up as an April Fool's joke, but it is also a real thing that works.
A lot of viable products came out that way. Gmail, I think.
Edit: Looks like it wasn't intended as a joke, but was initially received that way because of the timing: https://en.wikipedia.org/wiki/History_of_Gmail#Public_releas...
There are so many sites that I would love to see implement cli interfaces for. HN is one. tldp.org, die.net, etc.
I'm sure many of them have a lynx-useable page, but having a standardized cli interface to these would be super-useful to me as well.
No need to reinvent the wheel — topic/conversation sites like HN should just support RFC3977, while things like wikipedia can use RFC1436 (although a MIME typing extension is sorely needed).
You could look into links/elinks for a command line browser. Or WikipediaFS (http://wikipediafs.sourceforge.net)
Why not for images too? There are lots of good image-to-ascii art converters.
"We don't just support thin clients; we support anorexic ones!"
Yeah, kind of sad that this is an April Fool's joke and not a serious interface. I would love to be able to quickly look stuff up from the terminal.
>>> import requests, bs4 >>> def summary(name): ... resp = requests.get("https://en.wikipedia.org/wiki/{0}".format(name)) ... return bs4.BeautifulSoup(resp.text).p.get_text() ... >>> summary("Hacker news") 'Hacker News is a social news website focusing on computer science and entrepreneurship. It is run by Paul Graham\'s investment fund and startup incubator, Y Combinator. In general, content that can be submitted is defined as "anything that gratifies one\'s intellectual curiosity".[1]'It may be a joke, but it actually works:
$ telnet telnet.wmflabs.org Trying 208.80.155.160... Connected to telnet.wmflabs.org. Escape character is '^]'.
(I had some more text pasted here, but it didn't format well on HN)
I've tried a few queries, and it seems to work. Though I did just get an article in French for some reason...
I was hoping that telnet was making a comeback. It feels like an old friend.
Its called ssh. Telnet is pretty much dead to me.
Yes and no.
Telnet is perhaps the internet's debugger.
While the most common use was to access a CLI Somewhere, most of the "traditional" internet protocols can be operated via telnet.
Note btw that there is no login needed to access this "joke". Just enter the address and presto.
SSH is a single use tool in comparison.
html2text and a bash function seem to do okay.
Edit: Replace spaces with underscores for multi word lookups, like George_Washington. Or fix the function.$ sudo apt-get install html2text $ function lookup() {curl -sL http://wikipedia.com/wiki/$1|html2text -ascii|less;} $ lookup node.js
What's funny to me is how incredibly much faster the joke telnet interface is than using a web browser. Well, 'funny': it's actually pretty sad that we slow down our network usage so badly.
It seems that as hardware improves, software degenerates.
This in most part because software keeps being abstracted from the "hardware".
Also, graphics. As bandwidth has gone up we have gone from newline encoded ASCII to video at 1080p or higher.
I "love" it whenever i want to look up something and what i get pointed at is not a simple text document, but a 30+ minute video on Youtube of someone basically reading the same information.
> It seems that as hardware improves, software degenerates.
Spot on. I find myself frequently and increasingly frustrated with how slow most software is nowadays. Every app in my phone takes 2-3 seconds to load. Lots of websites/webapps take several seconds to load before they are readable/usable (Google Sheets is particularly offensive here).
Years ago I had a lot less hardware than today, and the software I use had less features, but I remember response times being much faster. I hope we're able to reverse this situation soon.
Graphics are no longer the issue. The issue these days is that your browser has to download an entire is apps with all it's dependencies before it can start rendering the few KB of text.
While i agree on the "app" thing, and something i tried to capture with the comment about abstraction, i have seen way too many sites default to massive images that they expect the client to scale as needed, taking up both bandwidth and client hardware resources.
I know people (mostly under 25) who look at a text document like that and complain "why isn't this a YouTube? I don't want to read shit"
Ah youth, and their fervent belief that they can multitask...
Cool. I would use this quite often.
I hope https://github.com/cscott/wikipedia-telnet gets more contributor love now. Especially a pager would dearly be needed.
I vividly remember the telnet interface to the library system at my university. It was easy to use and fast. There were even some old terminals you could use (my only real exposure to real terminals).
Then they replaced it with a slow html page. :-(
Joke or not, this is awesome! I could imagine using quite a few apps in the terminal like time tracking or chat. It feels so much quicker and quite frankly when browsing the web a lot of pages are visually heavy and chaotic that I fall back to the Safari reader mode very often anyway.
Well, what you're looking for is Emacs
Never got along with emacs. More of a vim person. I might misjudge of ignorance but emacs seems too heavy and complicated and this do it all attitude hits my as not very unixy.
That's because it's lispy!
Great stuff! A bit like the old BBS days albeit not as great an interface. Someone with a sniffer should see what traffic size is during article load on Telnet vs web version w/ cache cleared. I'm curious if it's more or less efficient.
On a related note, I was digging up info on old systems. OpenVMS this time. One company that still supports terminal ("green screen") and web apps had interesting things to say.
http://www3.sympatico.ca/n.rieck/docs/openvms_notes_my_OpenV...
In paradigm changes section, they pointed out they had been doing textual apps because they were easy, worked, and ran really fast. They apparently supported tons of users on a few VMS boxes that way. Hardly any company was interested unless they had a web interface. Switching to Web made the services 2-5x slower, necessitating hardware and software upgrades. They also had security troubles. Business is booming, though.
Lots of lessons to be learned. Old way was fast but harder to use and inflexible. Web is easier interface but slow and insecure. I still think client-server w/ minimal GUI's (eg REBOL) w/ efficient protocol is best middle ground.
I keep reposting this comment, but whenever telnet is mentioned (fun simulation of 1980s):
telnet telehack.com .... . usenet
I with they had trn installed.
Naturally I did some googling on "telehack" before typing it into the terminal. ;) So, I run it and see some familiar things. Talking to Eliza is a trap. I try StarWars expecting a banner or game. What I find... the movie rendered as text art... is awesome. The level of detail and compromises they made were great.
Thanks for the link as I'm sure I'll find more interesting stuff on here. :)
Huge telehack fan here, glad to see it get mentioned -- this immediately reminded me of TH.
It's case sensitive :-(
That's worse than it sounds thanks to Wikipedia's archaic capitalization constraints, so it's neither going to be the "official" case-sensitive name of the page nor the Title case, you could very well need to do SomeTHING like_THIS to read an article.... and then it gave me the German version for some reason?
I wish it had search.
>>> easybcd
easybcd
Sorry! Could not fetch "easybcd" for you.
No worries. There are lots of other pages to read.
Pick a different title.
>>> Easybcd
Easybcd
Sorry! Could not fetch "Easybcd" for you.
No worries. There are lots of other pages to read.
Pick a different title.
>>> EasyBCD
EasyBCD
EasyBCD
EasyBCD ist ein Programm, das von NeoSmart Technologies entwickelt wurde.
Es wird zum Konfigurieren und Anpassen des von Microsoft entwickelten
Bootloaders Bootmgr verwendet, der Teil der Boot Configuration Data (BCD)
der Windows-Versionen Windows Vista und jünger ist. EasyBCD kann benutzt
werden, um eine Multi-Boot-Konfiguration zwischen diesen und vorhergehenden
Versionen von Windows, sowie Linux, BSD und Mac OS X zu erstellen.
....Interestingly, the first character doesn't seem to matter. "easyBCD" and "EasyBCD" return the same results. And for some things all lower case appears to work:
"USB", "uSB", "usb", "Usb" all work, "usB" does not. I agree that case sensitivity is pointless here.
> and then it gave me the German version for some reason?
It probably associated your location with Germany. It gave me the English version.
Try:
:use en.wikipedia.org
> I wish it had search.
My first lookup (which I've already forgotten) gave me "results" but it could have been a disambiguation page.
> It's case sensitive :-(
Also IP-protocol sensitive :(
No AAAA RR.$telnet -6 telnet.wmflabs.org telnet: could not resolve telnet.wmflabs.org/telnet: Name or service not knownSadly, this does seem to be a limitation of our labs infrastructure. :(
Fixed!
somewhat surprisingly easyBCD works
I connected and spent a few minutes poking around to try and understand the usefulness of this. Something for those in countries where web traffic is blocked? Is this a power tool just for Wikipedia editors?
Oh, wait, ugh... is this an April Fool's joke?
Accessing Wikimedia content while being Really Really certain that there is no downloadable code being executed, or browser exploits being inserted. Browsing for the paranoid.
Here's a bit more: https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2...
I'm actually incredibly disappointed that this isn't a thing. I had a screensaver that would curl a random wikipedia page and display it - but it always looked like hot garbage, due to a bunch of extraneous stuff that I didn't want to bother filtering out. This would have been pretty great. :/
Have you not tried it? It's working great!
EDIT: With the exception of the connection being closed on the Wikipedia end relatively often.
Sure, but will it be up on April 2nd?
If you use it, we will keep it up.
Turns out, I don't know how to do that with telnet as a one-liner :/
Wow! This even has auto-completion and other shell goodies.
However, there seems to be a bug in the auto-completion:
When pressing TAB immediately after the prompt appears, the whole telnet session hangs and does not respond anymore.
Fixed!
April fools joke aside, I'd actually use this, but it doesn't seem to fully work. Most articles would not resolve for me and welcome page was joke content.
Try typing "use en.wikipedia.org" first.
Yeah, the domain setting was leaking between sessions before. Oops. Fixed now.
See https://meta.wikimedia.org/wiki/Telnet_gateway for technical details. The technology behind this is used for Wikipedia's Visual Editor, PDF export, and plain text export, so it is very likely to be supported if enough folks find it useful.
Oh -- and report bugs/feature requests at https://github.com/cscott/wikipedia-telnet/issues please. Patches are even better.
It's a joke, but it really works :)
I almost skipped it, but I tried anyway. Wonderful! I like the wiktionary:
use en.wiktionary.org
Somehow I got disconnected frequently. Is it intentional? Or is it just my company's network hate telnet?
It's quite possibly because they're getting a lot of traffic. As people are realizing that it's actually functional I'm sure they've got more telnet connections than all the MUDs in the world combined.
Maybe an aggressive timeout when there hasn't been any new input recently?
Much faster than the web interface. Even over Tor! The fast disconnections are a bit lame, would be a bit better if it used curses so that it would start scrolled up.
Out of interest, is it possible to set up an anonymous service with SSH? Or would it rely on configuring an SSH server to accept any user-provided credentials?
The public NetHack server at alt.org has an SSH interface alongside the telnet interface. Try
ssh nethack@alt.org
It works properly with no password. It would be cool if Wikimedia would set up a similar thing for this service.
Patches welcome!
I thought it was a funny joke - I was curious how far they took it. Wholy crap, `telnet telnet.wmflabs.org` works and the searches do too. Nice job :)
This is really nice for those of us who run vintage computers (e.g. Apple II) with ethernet adapters. I hope they keep this around.
No money but lots of time and ideas I see.
It's actually a 269-line test case for more serious projects at the foundation: the Offline Content Generator and Parsoid. We're allowed to have fun in the service of the greater goal. More technical details at https://meta.wikimedia.org/wiki/Telnet_gateway
I was meanly poking at Wales mails and hyper disrupting popup for money (you guess how much I like all this).
That said, I friggin love the telnet access point. I am a bit fed up with the ever more weighty web so simple text link + repl gets my vote.
I would really like it if HN could mark things that are jokes as such, or constrain them to one thread.
This is not just a joke; the interface is working (apart from some bugs you might expect at this stage) and while the announcement is clearly a joke, the interface is IMHO pretty cool
hrmf, for some reason it disconnects me seconds after giving the prompt.
Edit: Never mind, was using an overly "clever" terminal emulator...
I was immediately kicked off when using real telnet, but it worked fine with netcat (nc). I think this service gets confused if you send it the telnet control stuff.
Could be. but using a "dumber" terminal seems to allow the connection to last longer...
The comments in this thread make the joke even better.