[sf-lug] BBS Software for Linux (Ubuntu)

[Rick Moen]

Quoting Blake Haggerty (Blake.Haggerty at Sapphire.com):

> Does anyone have experience running a BBS on Ubuntu? What software did
> you use? Was maintainence a nightmare? Where are some good resources
> for running a BBS on Linux?
> 
> I was a sysop of a BBS back in the early 90s it was DOS based WWIV and
> everything was all dial up and only one user at a time, etc.. etc... I
> am sure things have changed considerably but it was a fun hobby I
> think I want to get back into. Any suggestions of where to start would
> be greatly appreciated.

(Me:  Six years building, owning, maintaining, and funding The Skeptic's
Board, 1:125/27 at fidonet, RBBS-PC behind BinkleyTerm and UFgate under
MS-DOS/DESQview with QWK and Silver Xpress offline mail doors.)

I don't have the information to answer your question, but I can answer
(IMVAO) a related one you _didn't_ exactly ask:



 From rick Wed Dec 12 14:14:50 2007
 Date: Wed, 12 Dec 2007 14:14:50 -0800
 To: luv-main at luv.asn.au
 Subject: Re: Life wo M$

Quoting Tony Langdon (vk3jed at gmail.com):

> I always remember DESQview as a text mode environment.

It was actually able to do both -- thus preventing incurring the
overhead of running graphical services when you didn't need them.  The
DESQview/X variant also did X11 (though I never used it much).

> The only time I had performance issues was when a poorly written DOS
> program would be stuck in a tight loop without calling any of the idle
> interrupts.  Under all other conditions, you hardly noticed DESQview
> was there.

Yes, and it was actually the most reasonable way to corral MS-Windows
3.x and prevent it from clobbering the rest of the machine.

The dial-up BBS that I ran for six years (custom setup using
BinkleyTerm, UFgate, and RBBS-PC) ran under DESQview.  I spent some
months trying to prototype an updated-technology replacement for it
using either Linux or OS/2-HPFS when I gradually realised that the whole
Fidotech/BBS was so riddled with technological underdesign that it just
was hopeless except as a means of doing low-tech connectivity in
countries with no Internet infrastructure.  (Anyway, if someone wants to
resurrect my BBS, let me know and I'll send you a stack of floppies.
;->)


 Date: Wed, 12 Dec 2007 15:02:34 -0800
 From: Rick Moen <rick at linuxmafia.com>
 To: Tony Langdon <vk3jed at gmail.com>
 Subject: Re: Life wo M$

Hi, Tony.

> LOL, well Fido is still around, though the Internet has taken a huge
> bite out of it. :)  I did manage to assist in the transition of a BBS
> from DOS/DESQview to OS/2 Warp Connect 3.  The BBS was more stable
> and performed better under OS/2 as it turned out (still running the
> same DOS software).  It was a Binkley/RA setup.

My prototype rebuild was some sort of Binkley/Maximus thing running
native OS/2 code entirely on HPFS.  However, the more I played with all
the software required, the more dumb limitations kept coming up, e.g.,
needless and avoidable dependency on 8.3 filenames, file-typing based
on three-letter extensions, kludged network protocols, etc.  At that
point, I started to remember the _other_ limitations of the Fidotech
world, e.g., stupid squabbling over the division of piddly telephone
bills for echomail, Policy4 netgod disputes, inability to coordinate
network namespace among multiple Fidotech networks, lack of availabilty
of most needed software under open source licensing terms[1], etc.

I remembered that echomail was a really bad, really slow, really fragile
imitation of Usenet, and that netmail was a really bad, really slow,
really fragile imitation of SMTP (well, of UUCP, really).  And I
remembered that my entire building in San Francisco was on static IPs on
a T1 line, that I could do real protocols using Linux, and didn't _have_
to deal with that other bullshit anymore.

So, at that point, I shut the thing down after a one-month grace period
for the users, and converted the machine over to a real Linux box on the
open Internet.

[1] This was the rationale for continued use (except in native OS/2
implementation) of BinkleyTerm, and previously for RBBS-PC, practically
the _only_ such software that was open source at the time.  Maximus
source code was freely available but proprietary:  It didn't go open
source until 2002.


 Date: Wed, 12 Dec 2007 17:19:47 -0800
 From: Rick Moen <rick at linuxmafia.com>
 To: Tony Langdon <vk3jed at gmail.com>
 Subject: Re: Life wo M$
                                                                                
Quoting Tony Langdon (vk3jed at gmail.com):                                        

> Yep, there were all sorts of DOS imposed limitations.  Running UUCP
> on DOS (for email-netmail and Usenet - echo gating) was interesting,
> the filenames had to be munged in all sorts of ways. :)

Yes, the munging routines in my friend Tim Pozar's UFgate package, for
example, were truly hideous.

> Yep, all sorts of personality issues.

It actually went way beyond personalities:  It amounted to pervasive
small-time thinking and underdesign.  One of the things I found
preferable about the Linux and open source world was far greater clue
quotient in exactly those areas.  Here's a trivial example:  You might
have noticed that X11 really doesn't work properly at less than
1024x768.  That is, your X server will image, but many dialogues will
extend down past the bottom of the screen and impose other, similar
irksome artifacts.

Even back in 1994, if you complained about that to, basically, anyone
who knew about X11, the answer you got was "Dude, sorry, the expected
practical minimum is 1024x768.  Get a real monitor and/or video card."

_That_ is the sort of no-we-won't-make-stupid-design-compromises mindset
I found refreshingly prevalent in Linux and open source, and entirely
lacking in the BBS world:  the insistence on doing things right, making
tools that scale and are reliable, making protocols that are robust and
efficient, insisting on modularity and network-transparency from top to
bottom.

(More recently, GNOME and KDE people have attempted to damage modularity
and network-transparency, but I consider them eminently ignorable.)

You might actually enjoy the two satirical articles I published in
_FidoNews_ writing at "The Fido Phool":

http://wps.com/FidoNet/FidoNews/1994/FIDO1129.NWS
http://wps.com/FidoNet/FidoNews/1993/FIDO1048.NWS

It's possible that 75% of the humour is in the You Had to Be There
category on account of topical references that are now lost to time's
bitbucket, but the pieces were deemed very amusing at the time.  Most of
my barbs were directed at people being notably petty and stupid.
Which, again, was the core reason for my dissatisfaction.

And the "petty and stupid" part seems to adequately account for extreme
slowness in converting essential BBS infrastructure to open source
licensing.  All of the copyright owners had their minds permanently
stuck in 1985, and were desparately afraid of losing their chance at
earning around US $200 total in shareware registration fees.  Which
explained why all of that software remained (because it was proprietary,
mostly binary-only, and poorly documented) badly debugged, inflexible,
and prone to mysterious failures.