Uptime numbers (was Re: [MUD-Dev] Orthogonality and invariants)

[Jay Carlson]

From: "Miroslav Silovic" <silovic at zesoi.fer.hr> writes at April 05, 2000
9:43 AM:

> First, in my experience, the main time-sink in writing MUDs is not
> coding per se, it's debugging. In most 1,000,000 line applications,
> you won't even notice a crash every 12 hours, or a memory leak that
> fills the swap after a couple of days - this is because most
> applications don't run for hours, let alone days. A MUD that crashes
> daily is (in my experience) unplayable. This goes especially for
> persistent MUDs that don't even see regular reboots more often than
> twice per year.
>
> So, MUD servers (and more generally, application servers) have
> EXTREMELY severe stability constraint: they have to be able to run
> stably for months, without visible bloat, fd leak, heap corruption,
> thread forking, thread blocking, and all other types of weirdness that
> plague applications.
>
> Several MUD codebases already reached this plateaou (MOO, various
> types of MUSH, Cold are those I know about for sure). I don't think
> that a less stable codebase can compete at all (because crashes and
> losses are just too annoying to bother with - especially with all
> those stable MUDs laying around).

OK, I gotta gloat.  From one of the semi-public MOO servers on a machine I
own:

|  The server has been up for 1 year, 2 months, 19 days, 1 hour, 25 minutes,
and 45 seconds.

Anybody else wanna contribute cool uptime numbers?

Jay




_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev