• The social aspect of pubnixes

    From lkh@lkh@cosmic.voyage to tilde.meta on Sat Aug 6 17:42:09 2022
    Here's another question about everyones tilde usage, and maybe
    some food for thought:

    I can see how a shared unix/linux system is a social construct.
    People are sharing resources like saving their stuff in a common
    file tree. Also, right from the beginning UNIX has supported
    direct communication like motd, write and the very basic news
    system described in *The UNIX Programming Environment* by
    Kernighan and Pike. Using these systems users are literally
    meeting and communicating on their one system.

    However, while most of these basic services are still there,
    they've been superseeded decades ago; by Email, Netnews and IRC.
    Even talk and finger transcend the one system, making the net the
    place people are actally meeting.

    And so the question here is: in the tildeverse it seems to be
    common lingo to meet on this or that tilde. Which seems to make
    sense especially for the special interest tildes like breadpunk
    or cosmic. But don't those users actually rather meet in the
    #cosmic or #breadpunk irc channels? or the respective newsgroups?
    That would be quite different from meeting on a particular single
    system, wouln't it?

    To ask the other way around, how much do folks actually use the
    traditional communications systems like write, when interacting
    on a single system?

    Cheers,

    lkh
    --- Synchronet 3.19a-Linux NewsLink 1.113
  • From anthk@anthk@texto-plano.xyz to tilde.meta on Sun Aug 7 07:54:40 2022
    On 2022-08-06, <lkh@cosmic.voyage> <lkh@cosmic.voyage> wrote:
    Here's another question about everyones tilde usage, and maybe
    some food for thought:

    I can see how a shared unix/linux system is a social construct.
    People are sharing resources like saving their stuff in a common
    file tree. Also, right from the beginning UNIX has supported
    direct communication like motd, write and the very basic news
    system described in *The UNIX Programming Environment* by
    Kernighan and Pike. Using these systems users are literally
    meeting and communicating on their one system.

    However, while most of these basic services are still there,
    they've been superseeded decades ago; by Email, Netnews and IRC.
    Even talk and finger transcend the one system, making the net the
    place people are actally meeting.

    And so the question here is: in the tildeverse it seems to be
    common lingo to meet on this or that tilde. Which seems to make
    sense especially for the special interest tildes like breadpunk
    or cosmic. But don't those users actually rather meet in the
    #cosmic or #breadpunk irc channels? or the respective newsgroups?
    That would be quite different from meeting on a particular single
    system, wouln't it?

    To ask the other way around, how much do folks actually use the
    traditional communications systems like write, when interacting
    on a single system?

    Cheers,

    lkh

    In such case, I'd say ITS was best suited, because it has no perms and everyone
    shared its projects everywhere, a true hackers' dream, and basically what
    RMS used.
    I still use daily some stuff from that OS, such as Maxima under OpenBSD (Macsyma?).
    --- Synchronet 3.19a-Linux NewsLink 1.113
  • From xwindows@xwindows@tilde.club to tilde.meta on Sun Aug 7 18:04:51 2022
    On Sat, 6 Aug 2022, lkh wrote:

    However, while most of these basic services are still there,
    they've been superseeded decades ago; by Email

    Actually, Unix-local email (i.e. `mail` command) originated within
    Unix itself, has been available from the *first* edition of AT&T UNIX
    circa 1971; thus qualifies as one of the Unix social interaction facilities.

    What this program originally did was simply appending the message to the recepient user's inbox (MBOX) file; and recepient would run `mail` command without parameter to page through (and manipulate) this file.

    When ARPANET mail (i.e. "email" in modern sense) came along, it just bolted standardized headers and addressing scheme into the message
    that this program send; making it possible for `mail` program to recognize foreign address and post the message through appropriate handler program
    to send tthat mail over to different host, through whatever mechanism
    that could be utilized. (FTP, UUCP, etc. to SMTP like it's used today)

    One can still use this `mail` command to send message to other local user
    today (use `mail SOMEONE` where "SOMEONE" is Unix username, without `@`).
    On modern implementation of `mail`, as far as I understand,
    it will pass the mail to system's MTA (on pubnix/Tilde system,
    that would probably be a fully-featured Internet-routable MTA/MDA)
    to do the actual delivery. This means the message would end up in
    target user's familiar Internet mailbox, whether local MBOX-based,
    local maildir-based, IMAP-based, or ~/.forward -ed.

    Nice side effect of this is you can use this old-school `mail` program
    to send an Internet email on such system; but a less-nice side effect
    is you probably cannot use this old-school `mail` program to read email
    sent to you unless your system's MDA is configured to deliver emails
    to you in traditional MBOX-in-homedir (or MBOX in spooldir) fashion;
    and you would have to use whatever Internet email program you're
    already using instead.

    finger transcend the one system

    Fingering user without `@` is a strictly local operation.
    You'll notice the difference immediately on systems with EfingerD daemon
    (like Tilde.club, Tilde.team, and Tilde.instute)
    that allows user to provide custom Internet Finger status text.

    For example, if someone run `finger xwindows` on Tilde.club,
    the information would be gathered via system calls and files;
    and that person would see my (actual) login stats
    as well as my ~/.project and ~/.plan content. No daemon is involved.

    However, if someone rather run `finger xwindows@tilde.club` there
    or from somewhere out there on the Internet, the Internet Finger daemon (EfingerD in this case) would handle the request,
    my Internet Finger banner would be printed instead
    (which in my case, contain ASCII art and ~/.plan content,
    but devoid of actual login stats).

    To ask the other way around, how much do folks actually use the
    traditional communications systems like write, when interacting
    on a single system?

    I'm reluctant to use `write` (or `wall`); mainly because people today spend large chunk of terminal time in full screen TUI applications
    rather than traditional line-oriented programs operating directly
    on top of TTY. [1] This means apart from the message polluting the
    receipient's workspace, the message text tend to get mangled
    (or disappeared) as soon as the foreground program redraw the screen;
    which means the message (apart from being annoying) is also hard to read.

    This does not even count users who use multiple TTYs/VTYs at the same time (very common these days due to terminal multiplexers and graphical
    terminal emulators); which one should the sender `write` to, that it would catch the attention in timely fashion? I'm doubt `write`'s built-in
    heuristics would help that much in this case.

    For `talk`; I do wish that Tilde.club has a functional `talk` implementation. (Currently, it got a YTalk without the actual daemon runing, which would just hang when one run the client) Personally, I would prefer a `talk` implementation that works in peer-to-peer fashion over Unix stream socket residing in recepient's home directory without any central daemon;
    but I hadn't tried to seek out and see if there is any actual `talk` implementation that works this way.

    But don't those users actually rather meet in the
    [...]
    respective newsgroups?

    I actually treat Newsgroups very much as a local facility,
    just with wormholes that make things from "the other side(s)"
    magically appear. (Considering that I actually read your post
    in-spool on-Tilde with `less` and posted this reply via `inews`)

    Regards,
    ~xwindows


    [1] When was the last time you edited text file in `ed`, `ex`, or `sed`
    rather than the likes of `nano`, `vim` or `emacs`?
    Or when was the last time you read/sent mail using `mail`/`sendmail`
    rather than the likes of `alpine` or `mutt`?
    --
    xwindows' gallery of freely-licensed artworks
    https://tilde.club/~xwindows/ http://tilde.club/~xwindows/ gopher://tilde.club/1/~xwindows/
    --- Synchronet 3.19a-Linux NewsLink 1.113
  • From lkh@lkh@cosmic.voyage to tilde.meta on Mon Aug 8 13:38:02 2022
    xwindows <xwindows@tilde.club> wrote:
    On Sat, 6 Aug 2022, lkh wrote:

    However, while most of these basic services are still there,
    they've been superseeded decades ago; by Email

    Actually, Unix-local email (i.e. `mail` command) originated within
    Unix itself, has been available from the *first* edition of AT&T UNIX
    circa 1971; thus qualifies as one of the Unix social interaction facilities.

    What this program originally did was simply appending the message to the recepient user's inbox (MBOX) file; and recepient would run `mail` command without parameter to page through (and manipulate) this file.

    I'm always fascinated by the simplicity of early UNIX solutions. The UNIX philosophy at it's best.

    Actually just your free form description would suffice to build a local
    mail system on short notice, just with the basic tools.

    Nice side effect of this is you can use this old-school `mail` program
    to send an Internet email on such system; but a less-nice side effect
    is you probably cannot use this old-school `mail` program to read email
    sent to you unless your system's MDA is configured to deliver emails
    to you in traditional MBOX-in-homedir (or MBOX in spooldir) fashion;
    and you would have to use whatever Internet email program you're
    already using instead.

    On most *nix Systems, mail still get's appended to /var/mail/<username>, doesn't it? Just like in the early days ...

    finger transcend the one system

    Fingering user without `@` is a strictly local operation.
    You'll notice the difference immediately on systems with EfingerD daemon (like Tilde.club, Tilde.team, and Tilde.instute)
    that allows user to provide custom Internet Finger status text.

    Intersting, I didn't know that basic finger works locally. Again, just
    as with basic mail, I could totally see how to implement a basic local
    finger command just with the standard UNIX tools.

    For example, if someone run `finger xwindows` on Tilde.club,
    the information would be gathered via system calls and files;
    and that person would see my (actual) login stats
    as well as my ~/.project and ~/.plan content. No daemon is involved.

    However, if someone rather run `finger xwindows@tilde.club` there
    or from somewhere out there on the Internet, the Internet Finger daemon (EfingerD in this case) would handle the request,
    my Internet Finger banner would be printed instead
    (which in my case, contain ASCII art and ~/.plan content,
    but devoid of actual login stats).

    Interesting, implementations obviously vary. While Linux finger deamons
    don't send any login stats to remote requests, NetBSD does ...
    not sure it that's a good idea.

    To ask the other way around, how much do folks actually use the
    traditional communications systems like write, when interacting
    on a single system?

    I'm reluctant to use `write` (or `wall`); mainly because people today spend large chunk of terminal time in full screen TUI applications
    rather than traditional line-oriented programs operating directly
    on top of TTY. [1] This means apart from the message polluting the receipient's workspace, the message text tend to get mangled
    (or disappeared) as soon as the foreground program redraw the screen;
    which means the message (apart from being annoying) is also hard to read.

    Right, using write is probably not a good idea. And translating the
    application to terminal multiplexers and desktop environments would
    probably seem too unwieldy.

    (Currently, it got a YTalk without the actual daemon runing, which would just hang when one run the client) Personally, I would prefer a `talk` implementation that works in peer-to-peer fashion over Unix stream socket residing in recepient's home directory without any central daemon;
    but I hadn't tried to seek out and see if there is any actual `talk` implementation that works this way.

    I like the idea, but looking at talk(1), doesn't talk have the same
    issues as write(1), with users logged in on multiple terminals?

    But don't those users actually rather meet in the
    [...]
    respective newsgroups?

    I actually treat Newsgroups very much as a local facility,
    just with wormholes that make things from "the other side(s)"
    magically appear. (Considering that I actually read your post
    in-spool on-Tilde with `less` and posted this reply via `inews`)

    I like that, so we meet on certain systems, while having mystical
    glimpses of users on other systems, too.

    Regards,
    ~xwindows

    Hey, thanks for your interesting reply!

    Best,

    ~lkh
    --- Synchronet 3.19a-Linux NewsLink 1.113