SDF inspired me to start setting up my own retro computing lab for fun; to experiment with actual working history is very entertaining! Since SDF mostly has older systems (https://wiki.sdf.org/doku.php?id=vintage_systems:start) I tried to get the later ones - Sparc M10, M7 & Ultrasparc, HP-C9000, etc.
I remember seeing the TOAD systems when I visited in 2016 long before they closed, it’s very sad that people no longer get to experience computer history in person the same way.
I run a public unix(openbsd) shell for fun, I call it my social network platform it sort of sucks, no users, ip6 only, a bunch of vm's on an old underpowered router running in my closet. But feel free to stop by and set up a .plan if you have nothing better to do.
It is pretty pointless, nobody needs or wants a unix shell account in this day and age. But I had fun setting it up, it started as an exercise to see what a shared multiuser postgres install would look like and got a little out of control. My current project is getting a rack of raspberry pi's(6 of them in a cute little case) hooked in as physical application nodes.
> nobody needs or wants a unix shell account in this day and age
I do. But I do not need just any Unix shell account, I need old and weird ones! I develop and maintain a portable utility (rlwrap) that is aimed at users of older software, who are often also using older or even obsolete systems.
For years, I used Polarhome (http://www.polarhome.com/) as a "dinosaur zoo" of obsolete systems (thans, Zoltan!) For every new release, building it on a creaky Solaris or HP-UX machine would expose a few bugs.
Because older systems are being replaced by (much more uniform) newer ones, there is a diminishing need for such extreme portability. This is also the reason that Polarhome closed in 2022.
In spite of this, testing on many different systems improves general code quality, even for users of mainstream systems like linux, BSD or OSX.
Of course, I could setup a couple of virtual machines, but that is a lot of hassle, especially for machines with uncommon processor architectures.
I have had many of these machines at various points in time, some even running on the public internet. They are a giant PITA to keep running and alive. This is why they don't usually last very long if they ever even get to public access(most don't).
Unless it's a super fun hobby for you, I wouldn't plan on this being very fun after the first dozen random crashes.
I'm curious, do you know which virtual machines (i.e. what emulator and what OS) you would want? Does the software exist and it's just a matter of the time to set it up? Or is it harder to get ahold of all the necessary old software (even if you have the emulator)?
Maybe in the modern age someone could make a "polarhome in a box" that offers a similar gamut of systems, but via preconfigured emulators that you can simply download and run.
Wow! I hope you know you are having a real impact in the world. Rlwrap has made my life easier so many times, it's in my top 3 most useful CLI tools. Thank you :)
> nobody needs or wants a unix shell account in this day and age
> I do. But I do not need just any Unix shell account, I need old and weird ones! I develop and maintain a portable utility (rlwrap) that is aimed at users of older software
Thank you, personally. I've used it in several contexts not just old systems, for example rlwrap is recommended with Clojure (okay, perhaps that's a comparatively small audience).
+1, same here, I've used line editors a fair bit (and enjoying line-oriented interface in general), so rlwrap has been an essential tool for me. Many thanks for your work!
Named after the Super Dimension Fortress from the Macross anime series. If you like mecha i recommend checking out the original series (it might look dated in some regards but still worth a watch. And the Do You Remember Love is a must watch after you finished the series, a grandiose animated spectacle, one of the most impressive animated films I've seen)
If you are not feeling like watching a long series, I recommend checking out Macross Plus, from the author of Cowboy Bebop and Samurai Champloo
The series is known as Robotech in the USA. The original series is not available legally in the USA to my knowledge but should be available on Japanese blu rays with english subtitles or on your favorite Linux ISO sharing website. The rest of the entries are on Disney+ or the aforementioned websites.
I first saw it when I was 10 in 1985 and it changed my life. I never knew that "cartoons" could be so serious. Lisa Hayes was such an incredible character. I didn't really know what anime was until I went to college and then I learned what Robotech really was and how it was created.
Thank you for mentioning those! I saw both Robotech and Macross Plus (in French) when I was a teenager. I still think about the SDF more often I would admit and the soundtrack is a regular occurrence in my work playlist :)
the VMS shell had so many good ideas. If i ever write a shell, I'm including VMS style abbreviations. If there is any modern POSIX shell that implements such a feature, let me know, because if there isn't I have to write one
I had access to a VMS system in my BBS days, and I had no idea it wasn't just some hard to use BBS software. When it clicked that it was a real operating system on a giant machine (I believe 11/380) it changed everything for me!
Apparently, the only place the VAX 380 exists is in a writing sample by Pearson Education. Otherwise, there is no evidence of DEC ever producing something called "VAX 380".
It looks entirely made up because the procedure described is also entirely alien to me, and I had professional experience with both VMS and Ultrix when they were still supported by DEC. (And it's certainly not BSD...)
I've only ever read about VMS in an historic context, like Wikipedia articles and blog posts. DEC and VMS are not well known. That's a shame, considering how much influence they had, especially on WinNT.
I don't know about VMS specifically (more people will just know it as the thing the VAX runs), but DEC is very well known to anyone in the computer space.
The PDP series brought us Unix and GNU, and the VAX was the only mainframe capable of competing with IBM. DEC was the largest terminal manufacturer (they made the vt100 and vt220. if you've ever run a terminal emulator, chances are it's emulating one of those or a machine that did). They created CP/M (and by extension DOS). DEC is very well known
SDF is cool, I commend their efforts of keeping a pub unix going! To me it feels like a stronghold of the "old school" web, similar to certain builtin board systems.
I regularly visit and enjoy reading the phlogs of their members as well.
I found a way to escape their shell (so you can run whatever you want), if you're not verified, it involves multiple steps to archive this. I mailed them 2x to their membership address, but since today no reaction. I asked also in their IRC.
Just a question to HN: should I wait more, try again? Or should I simply publish the vulnerabilities somewhere? If yes, where? It's my first time that I found a vulnerability at my own, not sure how to deal with that.
You shall wait. It's a volunteer powered system and while the ops are silent and terse in their mails, they're nice people.
Their plate is already quite full and they operate a whole universe of services, so cut them some slack.
It's not an ordinary service which is exposed to internet trying to turn a profit. They run SDF, two Mastodon instances, a mail server, a Git server, trying to salvage/keep alive living computer museum (SDF Vintage Systems), etc. etc.
I tried signing up to their mastodon three times and just never received the email accepting me. It's a shame because I wanted to be part of their community
Unless things have changed recently (last year or so), the SDF Mastodon servers are really slow and terrible at federating. They even had an incident where the servers failed, everyone lost their posts and had to start over again. Downtime was terrible.
SDF welcomed everyone openly during the initial Mastodon waves, so it was all very Eternal September.
If you're joining to make a spare account to participate with SDF people, awesome! But if you want it as your identity for all of Fedi, I think that would be a bad experience. I ended up getting my own MastoHost account for a while and it was a vastly better experience, until I burned out on Fedi.
SDF is a super fun place to experiment with Gopher though. I absolutely recommend getting your own Gopherhole on SDF. It's like the old Geocities days but in ASCII. (And make sure you grab Lagrange as your GUI Gopher / Gemini client. I liked Phetch as my terminal Gopher client.)
The performance hit was due to database work they were doing on the instance. Now it's a lot faster. The latest announcement reads as follows:
We've completed our first phase of database clean up, thank you for your patience. The impact on performance was heavy, but it was a necessary step. All active users and their posts, profile, connections and media will be migrated to the new servers. Once that has been completed, any remaining data will stay online for further migration and clean up. Our instance is nearly 10 years old of constant daily operation, but we ran into a migration wall which held us back on 4.1.x. Now that it is deprecated, we will do our best to jump to the latest version rather than migrate through. Your support and patience has been greatly appreciated.
I get that it's a volunteer system, but having donated for 2 years to help support their Lemmy instance, it's frustrating it's been down for 2 weeks without much of an update, just a hint "there's a good chance" it will come back. To me that seems lacking of transparency, not terse. How much disk space is it using? Maybe others in the community could help? How can they if they don't respond to emails? It was a nice thing while it lasted, but for federated social media, that kind of downtime hurts communities the most.
Don't publish. You already notified them, your shell escape isn't a big deal, publishing it will only be a pain for the volunteers running the service.
You can't have it both ways: if it's not a big deal, then he can publish it.
If you say "Don't publish", then you acknowledge that it's a big deal.
I say to GP: "Congrats for finding a shell escape, it's always a big deal. But don't publish it... Yet".
Give them a chance to fix it. But it they don't even answer to the emails, even just saying: "thx we're busy we can't fix right now but will do", then at some point you just publish.
It doesn't take long to answer an email saying "thanks, we'll fix it eventually".
"We'll fix it eventually" is not good enough. If a human can find a flaw, then a bot can find the same flaw, and the bots are always watching and always testing. If someone can't commit to immediate security response when running a public-facing internet service then they should not be running that service, because the rest of the internet will not forgive them when their machine gets popped and becomes everyone else's problem.
If they can't commit to a hard timeline of less than a few days, then publish. What happens next is not your fault - it was inevitable anyway.
Edit for clarity: This is just in general, not specifically SDF or small orgs or large orgs. The internet does not care about the difference. The internet just does not care period. Nobody is going to give anyone else any breaks, and especially not a botnet.
I think you should create some visible but harmless nuisance using this shell escape, so that it's likely to get noticed, but doesn't damage anyone's valuable data.
Perhaps just run "bash -c 'stress --cpu 64 ; echo fix your shell escape'"l " or something like that.
Creating a nuisance is not a good way to go about it.
Some security practices sometimes feels like someone stabbing you just to prove you could be stabbed.
Then they point at the wound and say: "See? You should be more careful."
Yes, the risk is real, but creating harm to demonstrate it isnt the same as protecting people.
Well, ruining everyone's day on that particular host is not a nice way to "bring this to attention".
If I ever experienced something like that, I'd be banning the person (or limiting their resources drastically) for 60 to 90 days to bring the impact of this matter to their attention.
Anything affecting users on a system is not harmless.
I did it too but TBH as I used small tools such as tcc, jimsh, eforth+muxleq, sacc, smu, catpoint+pointtools, compilers from https://t3x.org... I didn't care a lot on the rest, I'm pretty happy with my current account.
You can do a lot with S9 Scheme and the Unix API/syscalls it supports.
just got my stickers from there yesterday! :-) i wish my less cs-oriented friends could see how cool i think the sdf is, lol; and, that some kind of "small-web" system, complete with the self-expression the sdf offers via web-hosting, a radio station(!), etc., was accessible to more people (not at the fault of anyone; just that there's a lot to the internet that most people will never see). :>
> While we did initially start out on a single computer in 1987, the
> SDF is now a network of 8 64bit enterprise class servers running
> NetBSD realising a combined processing power of over 21.1 GFLOPS!
Which piqued my interest about how that compares to today's computers. nVidia's venerable 1080Ti from 2017 measures about 11300 GFLOPS, or 11.3 teraFLOPS. About a fifty times increase.
I had an account there years ago but never really saw the point. I was already SSHing in from a shell, just to end up at another, different one. Kind of whimsical I guess, but ultimately of scant practical use.
In the 90s and even into the 2000s getting an email account that had shell access was really nice; there were things you could do that would have been hard to do locally.
Very cool how they tried to move and preserve many of the living computer museum’s computers before Paul Allens sister could sell them all off. https://wiki.sdf.org/doku.php?id=vintage_systems:lcml_collec...
I remember seeing the TOAD systems when I visited in 2016 long before they closed, it’s very sad that people no longer get to experience computer history in person the same way.
ssh to applicant@register.public.outband.net
instructions at https://www.public.outband.net note that it's ip6 only.
It is pretty pointless, nobody needs or wants a unix shell account in this day and age. But I had fun setting it up, it started as an exercise to see what a shared multiuser postgres install would look like and got a little out of control. My current project is getting a rack of raspberry pi's(6 of them in a cute little case) hooked in as physical application nodes.
I do. But I do not need just any Unix shell account, I need old and weird ones! I develop and maintain a portable utility (rlwrap) that is aimed at users of older software, who are often also using older or even obsolete systems.
For years, I used Polarhome (http://www.polarhome.com/) as a "dinosaur zoo" of obsolete systems (thans, Zoltan!) For every new release, building it on a creaky Solaris or HP-UX machine would expose a few bugs.
Because older systems are being replaced by (much more uniform) newer ones, there is a diminishing need for such extreme portability. This is also the reason that Polarhome closed in 2022.
In spite of this, testing on many different systems improves general code quality, even for users of mainstream systems like linux, BSD or OSX.
Of course, I could setup a couple of virtual machines, but that is a lot of hassle, especially for machines with uncommon processor architectures.
a powerpc xserve (running OSX server)
a sparc box (on solaris)
an alpha box (on either VMS or Digital Unix)
a pa-risc box (hp-ux)
a modern power box (Rocky or AIX)
an itanium box (running either VMS or NT depending on what the alpha is running)
a pi cluster (plan 9)
and a commodity x86 server (running OpenBSD, FreeBSD, Debian, Hurd, Redox, Serenity, reactos, and AROS).
and make a MOAP (mother of all pubnixes). if anyone has any hardware they'd like to donate, get in contact :)
Unless it's a super fun hobby for you, I wouldn't plan on this being very fun after the first dozen random crashes.
Maybe in the modern age someone could make a "polarhome in a box" that offers a similar gamut of systems, but via preconfigured emulators that you can simply download and run.
> I do. But I do not need just any Unix shell account, I need old and weird ones! I develop and maintain a portable utility (rlwrap) that is aimed at users of older software
Thank you, personally. I've used it in several contexts not just old systems, for example rlwrap is recommended with Clojure (okay, perhaps that's a comparatively small audience).
If you are not feeling like watching a long series, I recommend checking out Macross Plus, from the author of Cowboy Bebop and Samurai Champloo
The series is known as Robotech in the USA. The original series is not available legally in the USA to my knowledge but should be available on Japanese blu rays with english subtitles or on your favorite Linux ISO sharing website. The rest of the entries are on Disney+ or the aforementioned websites.
Somehow I still remembered most of the shell syntax in a book I read about it probably in 2001. Don't ask me ... I don't know how either.
Got bored in about 10 minutes but still, another box checked off!
https://www.pearsonhighered.com/assets/samplechapter/0/2/0/5...
The PDP series brought us Unix and GNU, and the VAX was the only mainframe capable of competing with IBM. DEC was the largest terminal manufacturer (they made the vt100 and vt220. if you've ever run a terminal emulator, chances are it's emulating one of those or a machine that did). They created CP/M (and by extension DOS). DEC is very well known
He's an absolutely kind soul who is deeply interested in all kinds of retro projects. I wish there were more folks like him in tech generally
https://sdf.org/plan9/
Side note: here's my workflow for running Plan 9 on Windows:
https://youtu.be/IzEa2L_Pgw0?si=unM5l2-_i_g-NYKP
I regularly visit and enjoy reading the phlogs of their members as well.
Just a question to HN: should I wait more, try again? Or should I simply publish the vulnerabilities somewhere? If yes, where? It's my first time that I found a vulnerability at my own, not sure how to deal with that.
Their plate is already quite full and they operate a whole universe of services, so cut them some slack.
It's not an ordinary service which is exposed to internet trying to turn a profit. They run SDF, two Mastodon instances, a mail server, a Git server, trying to salvage/keep alive living computer museum (SDF Vintage Systems), etc. etc.
SDF welcomed everyone openly during the initial Mastodon waves, so it was all very Eternal September.
If you're joining to make a spare account to participate with SDF people, awesome! But if you want it as your identity for all of Fedi, I think that would be a bad experience. I ended up getting my own MastoHost account for a while and it was a vastly better experience, until I burned out on Fedi.
SDF is a super fun place to experiment with Gopher though. I absolutely recommend getting your own Gopherhole on SDF. It's like the old Geocities days but in ASCII. (And make sure you grab Lagrange as your GUI Gopher / Gemini client. I liked Phetch as my terminal Gopher client.)
You can't have it both ways: if it's not a big deal, then he can publish it.
If you say "Don't publish", then you acknowledge that it's a big deal.
I say to GP: "Congrats for finding a shell escape, it's always a big deal. But don't publish it... Yet".
Give them a chance to fix it. But it they don't even answer to the emails, even just saying: "thx we're busy we can't fix right now but will do", then at some point you just publish.
It doesn't take long to answer an email saying "thanks, we'll fix it eventually".
If they can't commit to a hard timeline of less than a few days, then publish. What happens next is not your fault - it was inevitable anyway.
Edit for clarity: This is just in general, not specifically SDF or small orgs or large orgs. The internet does not care about the difference. The internet just does not care period. Nobody is going to give anyone else any breaks, and especially not a botnet.
Perhaps just run "bash -c 'stress --cpu 64 ; echo fix your shell escape'"l " or something like that.
Some security practices sometimes feels like someone stabbing you just to prove you could be stabbed. Then they point at the wound and say: "See? You should be more careful."
Yes, the risk is real, but creating harm to demonstrate it isnt the same as protecting people.
If I ever experienced something like that, I'd be banning the person (or limiting their resources drastically) for 60 to 90 days to bring the impact of this matter to their attention.
Anything affecting users on a system is not harmless.
You can do a lot with S9 Scheme and the Unix API/syscalls it supports.
SDF Public Access Unix System - https://news.ycombinator.com/item?id=32340635 - Aug 2022 (29 comments)
SDF Public Access Unix System - https://news.ycombinator.com/item?id=31076886 - April 2022 (46 comments)
SDF Public Access Unix System - https://news.ycombinator.com/item?id=14940790 - Aug 2017 (29 comments)
SDF – Public Access Unix System - https://news.ycombinator.com/item?id=14134798 - April 2017 (51 comments)
> While we did initially start out on a single computer in 1987, the
> SDF is now a network of 8 64bit enterprise class servers running
> NetBSD realising a combined processing power of over 21.1 GFLOPS!
Which piqued my interest about how that compares to today's computers. nVidia's venerable 1080Ti from 2017 measures about 11300 GFLOPS, or 11.3 teraFLOPS. About a fifty times increase.
"this page was generated using ksh, sed and awk"
It’s much less needed now.