Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > alt.folklore.computers > #163292 > unrolled thread
| Started by | Neo Fortune <serenity3043@gmail.com> |
|---|---|
| First post | 2016-05-06 02:04 -0700 |
| Last post | 2016-05-17 20:05 -0400 |
| Articles | 20 on this page of 268 — 43 participants |
Back to article view | Back to alt.folklore.computers
What Would Be Your Ultimate Computer? Neo Fortune <serenity3043@gmail.com> - 2016-05-06 02:04 -0700
Re: What Would Be Your Ultimate Computer? philo <philo@privacy.net> - 2016-05-06 11:00 -0500
Re: What Would Be Your Ultimate Computer? Stan Barr <plan.b@bluesomatic.org> - 2016-05-06 16:34 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-06 18:20 +0100
Re: What Would Be Your Ultimate Computer? Stan Barr <plan.b@bluesomatic.org> - 2016-05-07 07:12 +0000
Re: What Would Be Your Ultimate Computer? Bud Frede <frede@mouse-potato.com> - 2016-05-11 14:55 -0400
Re: What Would Be Your Ultimate Computer? cb@elaine.df.lth.se (Christian Brunschen) - 2016-05-11 19:20 +0000
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-11 21:16 +0200
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-07 13:14 +0000
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-07 20:18 -0700
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-08 04:22 +0000
Re: What Would Be Your Ultimate Computer? Anne & Lynn Wheeler <lynn@garlic.com> - 2016-05-07 21:46 -0700
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-08 06:58 +0100
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-08 07:31 -0400
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-06 12:32 -0400
Re: What Would Be Your Ultimate Computer? "Rod Speed" <rod.speed.aaa@gmail.com> - 2016-05-07 02:30 +1000
Re: What Would Be Your Ultimate Computer? Jorgen Grahn <grahn+nntp@snipabacken.se> - 2016-05-06 17:19 +0000
Re: What Would Be Your Ultimate Computer? Melzzzzz <mel@zzzzz.com> - 2016-05-06 19:41 +0200
Re: What Would Be Your Ultimate Computer? Neo Fortune <serenity3043@gmail.com> - 2016-05-09 02:43 -0700
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-09 06:58 -0700
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-09 17:05 +0000
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-10 12:34 +0000
Re: What Would Be Your Ultimate Computer? Bud Frede <frede@mouse-potato.com> - 2016-05-11 14:38 -0400
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-11 21:13 +0200
Re: What Would Be Your Ultimate Computer? simon@twoplaces.co.uk (Simon Turner) - 2016-05-12 12:50 +0100
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-12 17:49 +0000
Re: What Would Be Your Ultimate Computer? simon@twoplaces.co.uk (Simon Turner) - 2016-05-13 12:10 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-12 13:04 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-13 11:18 -0700
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-15 14:09 +0000
Re: What Would Be Your Ultimate Computer? Ibmekon <Ibmekon> - 2016-05-15 21:53 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-16 12:12 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-16 06:16 -0700
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-17 12:32 +0000
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-17 10:22 -0500
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-18 07:32 -0400
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-18 12:28 +0000
Re: What Would Be Your Ultimate Computer? mausg@mail.com - 2016-05-18 19:23 +0000
Re: What Would Be Your Ultimate Computer? sidd@situ.com (sidd) - 2016-05-18 20:15 -0400
Re: What Would Be Your Ultimate Computer? Andrew Swallow <am.swallow@btopenworld.com> - 2016-05-19 08:06 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-19 12:27 +0000
Re: What Would Be Your Ultimate Computer? sidd@situ.com (sidd) - 2016-05-19 14:41 -0400
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-19 12:27 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-21 04:59 -0700
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-22 13:15 +0000
Re: What Would Be Your Ultimate Computer? "Charles Richmond" <numerist@aquaporin4.com> - 2016-05-23 20:14 -0500
Re: What Would Be Your Ultimate Computer? Alfred Falk <falk@arc.ab.ca> - 2016-05-24 04:45 +0000
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-24 12:11 +0000
Re: What Would Be Your Ultimate Computer? Andrew Swallow <am.swallow@btopenworld.com> - 2016-05-24 14:19 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-25 13:11 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-25 09:22 -0700
Re: What Would Be Your Ultimate Computer? Andrew Swallow <am.swallow@btopenworld.com> - 2016-05-25 17:31 +0100
Re: What Would Be Your Ultimate Computer? Alfred Falk <falk@arc.ab.ca> - 2016-05-26 21:41 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-18 23:19 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-19 12:27 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-16 06:17 -0700
Re: What Would Be Your Ultimate Computer? "Charles Richmond" <numerist@aquaporin4.com> - 2016-05-23 20:16 -0500
Re: What Would Be Your Ultimate Computer? Ibmekon <Ibmekon> - 2016-05-16 20:08 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-17 12:32 +0000
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-09 16:29 -0300
Re: What Would Be Your Ultimate Computer? Neo Fortune <serenity3043@gmail.com> - 2016-05-10 02:10 -0700
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-10 06:42 -0400
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-10 07:43 -0500
Re: What Would Be Your Ultimate Computer? Neo Fortune <serenity3043@gmail.com> - 2016-05-11 01:49 -0700
Re: What Would Be Your Ultimate Computer? Walter Banks <walter@bytecraft.com> - 2016-05-11 09:06 -0400
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-11 09:47 -0400
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-11 09:33 -0500
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-11 10:26 -0500
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-11 11:11 -0500
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-11 17:48 +0100
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-16 20:02 -0400
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-11 12:05 -0500
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-11 14:59 -0500
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-11 21:32 -0500
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-16 19:59 -0400
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-11 21:34 -0500
Re: What Would Be Your Ultimate Computer? Andrew Swallow <am.swallow@btopenworld.com> - 2016-05-12 03:39 +0100
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-12 06:13 +0100
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-12 14:46 -0400
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-11 14:31 -0400
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-11 14:21 -0700
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-11 17:50 -0500
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-11 16:01 -0700
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-11 21:35 -0500
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-12 02:50 +0000
Re: What Would Be Your Ultimate Computer? Bob Eager <news0006@eager.cx> - 2016-05-12 10:07 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-12 11:39 +0100
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-12 17:49 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-12 21:55 +0100
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-12 13:12 -0700
Re: What Would Be Your Ultimate Computer? Andy Burns <feb2017-usenet@adslpipe.co.uk> - 2016-05-12 00:20 +0100
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-11 21:40 -0400
Re: What Would Be Your Ultimate Computer? Joy Beeson <jbeeson@invalid.net.invalid> - 2016-05-14 23:47 -0300
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-15 14:09 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-11 11:37 -0700
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-11 11:40 -0700
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-11 15:15 -0500
Re: What Would Be Your Ultimate Computer? cb@elaine.df.lth.se (Christian Brunschen) - 2016-05-11 21:50 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-11 15:49 -0700
Re: What Would Be Your Ultimate Computer? Walter Banks <walter@bytecraft.com> - 2016-05-11 23:36 -0400
Re: What Would Be Your Ultimate Computer? cb@elaine.df.lth.se (Christian Brunschen) - 2016-05-12 10:32 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-13 11:16 -0700
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-13 15:11 -0700
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-15 17:42 -0400
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-15 17:34 -0500
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-15 20:32 -0400
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-15 22:21 -0700
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-16 06:57 -0400
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-16 13:12 -0500
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-16 23:25 +0000
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-16 19:53 -0400
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-17 08:23 -0500
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-17 09:16 -0500
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-17 10:40 -0500
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-17 16:59 +0000
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-17 13:42 -0500
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-17 11:56 -0700
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-17 12:10 -0700
Re: What Would Be Your Ultimate Computer? mausg@mail.com - 2016-05-17 19:58 +0000
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-18 02:50 -0300
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-17 23:51 -0700
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-18 07:32 -0400
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-18 14:28 -0700
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-18 23:18 +0100
Re: What Would Be Your Ultimate Computer? RJH <patchmoney@gmx.com> - 2016-05-19 11:07 +0100
Re: What Would Be Your Ultimate Computer? Dan Espen <despen@verizon.net> - 2016-05-19 10:00 -0400
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-19 12:53 -0400
Re: What Would Be Your Ultimate Computer? Bud Frede <frede@mouse-potato.com> - 2016-05-23 10:56 -0400
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-23 12:59 -0500
Re: What Would Be Your Ultimate Computer? Joy Beeson <jbeeson@invalid.net.invalid> - 2016-05-24 00:01 -0300
Re: What Would Be Your Ultimate Computer? Bob Eager <news0006@eager.cx> - 2016-05-23 18:48 +0000
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-19 02:18 -0300
Re: What Would Be Your Ultimate Computer? Stan Barr <plan.b@bluesomatic.org> - 2016-05-19 07:03 +0000
Re: What Would Be Your Ultimate Computer? mausg@mail.com - 2016-05-19 13:11 +0000
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-18 12:28 +0000
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-19 02:38 -0300
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-21 21:26 -0400
Re: What Would Be Your Ultimate Computer? Lawrence Statton NK1G <lawrence@senguio.mx> - 2016-05-18 10:39 -0500
Re: What Would Be Your Ultimate Computer? mausg@mail.com - 2016-05-18 19:21 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-18 12:35 -0700
Re: What Would Be Your Ultimate Computer? mausg@mail.com - 2016-05-18 21:06 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-18 12:34 -0700
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-17 14:35 -0500
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-17 15:43 -0500
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-17 14:32 -0500
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-17 14:53 -0500
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-18 07:32 -0400
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-19 02:43 -0300
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-17 19:44 +0000
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-17 22:40 +0000
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-17 14:42 +0000
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-17 17:01 +0000
Re: What Would Be Your Ultimate Computer? Lawrence Statton NK1G <lawrence@senguio.mx> - 2016-05-17 12:09 -0500
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-17 14:31 -0300
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-18 12:28 +0000
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-17 12:33 -0500
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-17 10:31 -0500
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-16 20:37 -0400
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-17 11:12 +0200
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-17 05:49 -0700
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-17 08:21 -0500
Re: What Would Be Your Ultimate Computer? Bud Frede <frede@mouse-potato.com> - 2016-05-23 12:23 -0400
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-16 22:03 -0700
Re: What Would Be Your Ultimate Computer? "Rod Speed" <rod.speed.aaa@gmail.com> - 2016-05-12 05:09 +1000
Re: What Would Be Your Ultimate Computer? Bud Frede <frede@mouse-potato.com> - 2016-05-23 09:39 -0400
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-23 09:48 -0500
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-23 13:51 -0400
Re: What Would Be Your Ultimate Computer? Dave Garland <dave.garland@wizinfo.com> - 2016-05-23 13:49 -0500
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-11 11:34 -0700
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-11 14:06 -0500
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-11 17:49 +0200
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-11 16:53 +0000
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-12 03:11 -0300
Re: What Would Be Your Ultimate Computer? Bernd Felsche <berfel@innovative.iinet.net.au> - 2016-05-14 16:59 +0800
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-14 10:48 +0100
Re: What Would Be Your Ultimate Computer? Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-05-14 15:53 +0000
Re: What Would Be Your Ultimate Computer? Anne & Lynn Wheeler <lynn@garlic.com> - 2016-05-14 09:37 -0700
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-14 12:09 -0500
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-15 14:09 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-11 17:22 +0100
Re: What Would Be Your Ultimate Computer? Walter Banks <walter@bytecraft.com> - 2016-05-11 12:59 -0400
Re: What Would Be Your Ultimate Computer? usenet@only.tnx (Questor) - 2016-05-21 00:56 +0000
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-20 22:17 -0400
Re: What Would Be Your Ultimate Computer? usenet@only.tnx (Questor) - 2016-05-22 05:53 +0000
Re: What Would Be Your Ultimate Computer? "gareth G4SDW GQRP #3339" <no.spam@thank.you.invalid> - 2016-05-22 11:26 +0100
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-22 07:13 -0400
Re: What Would Be Your Ultimate Computer? usenet@only.tnx (Questor) - 2016-05-21 00:56 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-22 06:43 +0100
Re: What Would Be Your Ultimate Computer? "gareth G4SDW GQRP #3339" <no.spam@thank.you.invalid> - 2016-05-22 11:22 +0100
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-22 07:14 -0400
Re: What Would Be Your Ultimate Computer? "gareth G4SDW GQRP #3339" <no.spam@thank.you.invalid> - 2016-05-22 13:35 +0100
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-22 07:29 -0700
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-22 08:35 -0500
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-23 06:41 +0100
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-23 16:11 +0200
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-23 16:43 -0300
Re: What Would Be Your Ultimate Computer? Dan Espen <despen@verizon.net> - 2016-05-23 16:50 -0400
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-24 01:50 -0300
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-24 12:11 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-24 13:40 +0100
Re: What Would Be Your Ultimate Computer? mausg@mail.com - 2016-05-24 19:33 +0000
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-23 21:33 -0400
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-24 08:54 -0500
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-24 16:38 -0300
Re: What Would Be Your Ultimate Computer? Bob Eager <news0006@eager.cx> - 2016-05-24 20:04 +0000
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-24 22:24 -0400
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-25 12:45 -0700
Re: What Would Be Your Ultimate Computer? JimP <solosam90@gmail.com> - 2016-05-24 17:15 -0500
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-24 11:55 -0400
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-24 18:18 +0100
Re: What Would Be Your Ultimate Computer? "Osmium" <r124c4u102@comcast.net> - 2016-05-24 13:26 -0500
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-24 11:59 -0700
Re: What Would Be Your Ultimate Computer? Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-05-24 17:03 -0300
Re: What Would Be Your Ultimate Computer? usenet@only.tnx (Questor) - 2016-05-26 17:55 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-24 11:53 -0700
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-10 17:47 +0100
Re: What Would Be Your Ultimate Computer? Bud Frede <frede@mouse-potato.com> - 2016-05-11 14:41 -0400
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-10 12:34 +0000
Re: What Would Be Your Ultimate Computer? Neo Fortune <serenity3043@gmail.com> - 2016-05-11 01:44 -0700
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-11 11:32 -0700
Re: What Would Be Your Ultimate Computer? Lawrence Statton NK1G <lawrence@senguio.mx> - 2016-05-06 13:55 -0500
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-06 19:29 -0400
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-07 13:14 +0000
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-07 05:39 -0400
Re: What Would Be Your Ultimate Computer? Bob Eager <news0006@eager.cx> - 2016-05-07 11:35 +0000
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-07 17:38 -0400
Re: What Would Be Your Ultimate Computer? Bob Eager <news0006@eager.cx> - 2016-05-07 22:30 +0000
Re: What Would Be Your Ultimate Computer? Anne & Lynn Wheeler <lynn@garlic.com> - 2016-05-07 13:25 -0700
Re: What Would Be Your Ultimate Computer? Anne & Lynn Wheeler <lynn@garlic.com> - 2016-05-07 17:16 -0700
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-07 20:39 -0400
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-07 21:57 -0400
Re: What Would Be Your Ultimate Computer? "J. Clarke" <j.clarke.873638@gmail.com> - 2016-05-07 22:49 -0400
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-08 07:08 -0700
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-08 20:58 -0400
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-07 13:14 +0000
Re: What Would Be Your Ultimate Computer? Quadibloc <jsavard@ecn.ab.ca> - 2016-05-06 12:09 -0700
Re: What Would Be Your Ultimate Computer? "gareth G4SDW GQRP #3339" <no.spam@thank.you.invalid> - 2016-05-07 10:39 +0100
Re: What Would Be Your Ultimate Computer? Gene Wirchenko <genew@telus.net> - 2016-05-07 20:16 -0700
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-08 13:08 +0000
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-08 16:20 +0200
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-09 13:08 +0000
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name,invalid> - 2016-05-09 15:34 +0200
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-10 12:34 +0000
Re: What Would Be Your Ultimate Computer? cb@elaine.df.lth.se (Christian Brunschen) - 2016-05-10 13:42 +0000
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-11 12:18 +0000
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-11 13:19 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-11 16:57 +0100
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-11 16:20 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-11 17:38 +0100
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-10 15:49 +0200
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-11 12:18 +0000
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-10 11:07 -0400
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-11 12:18 +0000
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-11 13:16 +0000
Re: What Would Be Your Ultimate Computer? Michael Black <et472@ncf.ca> - 2016-05-11 09:45 -0400
Re: What Would Be Your Ultimate Computer? Andrew Swallow <am.swallow@btopenworld.com> - 2016-05-11 16:22 +0100
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-11 15:47 +0000
Re: What Would Be Your Ultimate Computer? Peter Flass <peter_flass@yahoo.com> - 2016-05-11 10:52 -0400
Re: What Would Be Your Ultimate Computer? scott@slp53.sl.home (Scott Lurndal) - 2016-05-11 15:49 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-10 17:44 +0100
Re: What Would Be Your Ultimate Computer? jmfbahciv <See.above@aol.com> - 2016-05-11 12:18 +0000
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-11 13:44 +0100
Re: What Would Be Your Ultimate Computer? Morten Reistad <first@last.name.invalid> - 2016-05-11 17:46 +0200
Re: What Would Be Your Ultimate Computer? Thomas Lake <tomlake07@gmail.com> - 2016-05-10 10:47 -0700
Re: What Would Be Your Ultimate Computer? Bob Eager <news0006@eager.cx> - 2016-05-10 18:36 +0000
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-16 19:45 -0400
Re: What Would Be Your Ultimate Computer? Ahem A Rivet's Shot <steveo@eircom.net> - 2016-05-17 06:35 +0100
Re: What Would Be Your Ultimate Computer? Walter Bushell <proto@panix.com> - 2016-05-17 20:05 -0400
Page 13 of 14 — ← Prev page 1 … 11 12 [13] 14 Next page →
| From | jmfbahciv <See.above@aol.com> |
|---|---|
| Date | 2016-05-09 13:08 +0000 |
| Message-ID | <PM0005326827D1037E@aca40e74.ipt.aol.com> |
| In reply to | #163417 |
Morten Reistad wrote: > In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, > jmfbahciv <See.above@aol.com> wrote: >>Gene Wirchenko wrote: >>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>> <no.spam@thank.you.invalid> wrote: >>> >>> [snip] >>> >>>>There would be no OS, unless limited to the 10 or so calls that CP/M had, >>>>so allowing me to proceed with complete freedom in whatever direction >>>>I chose. >>> >>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>> more. >> >>If you need a network, there would be triple the calls. > > No, assign/bind, connect, listen, accept, plus send and receive; and some extra > code in open,read,write and close would do the trick. You would want a > select/poll call, but that should already be there. > > That makes half a dozen new calls, and extra code in half a dozen more. > > Networking is not that hard. We learned how to do this right by ca 1980. > > The rest is implementation detail; roughly done in the *n*xes by 1985. > > Legacy, not-done-quite-right protocols would need more calls. Hence my too high estimate. And, for efficiency, it might be better to have a new call rather than trying to use an old call with new arguments. /BAH
[toc] | [prev] | [next] | [standalone]
| From | Morten Reistad <first@last.name,invalid> |
|---|---|
| Date | 2016-05-09 15:34 +0200 |
| Message-ID | <a8g60d-akf.ln1@sambook.reistad.name> |
| In reply to | #163461 |
In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, jmfbahciv <See.above@aol.com> wrote: >Morten Reistad wrote: >> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >> jmfbahciv <See.above@aol.com> wrote: >>>Gene Wirchenko wrote: >>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>> <no.spam@thank.you.invalid> wrote: >>>> >>>> [snip] >>>> >>>>>There would be no OS, unless limited to the 10 or so calls that CP/M had, >>>>>so allowing me to proceed with complete freedom in whatever direction >>>>>I chose. >>>> >>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>> more. >>> >>>If you need a network, there would be triple the calls. >> >> No, assign/bind, connect, listen, accept, plus send and receive; and some >extra >> code in open,read,write and close would do the trick. You would want a >> select/poll call, but that should already be there. >> >> That makes half a dozen new calls, and extra code in half a dozen more. >> >> Networking is not that hard. We learned how to do this right by ca 1980. >> >> The rest is implementation detail; roughly done in the *n*xes by 1985. >> >> Legacy, not-done-quite-right protocols would need more calls. > >Hence my too high estimate. And, for efficiency, it might be better to have >a new call rather than trying to use an old call with new arguments. > >/BAH No; have general calls for open,read,write,close and a select/poll (waits until something happens, can be many streams providing input or output buffers accepting data, or exceptions). For the servers you need bind/listen/accept, and for more elaborate network stuff you need connect and send/receive. That is it. Plus some statistics and management calls for root/admin. But they can be in sysctl etc. Half a dozen new calls and expansion of half a dozen more. KISS. Also, nfs uses mount/umount just like any other file system. So does afs. And zfs. The magic is in the file system driver in the kernel. -- mrr
[toc] | [prev] | [next] | [standalone]
| From | jmfbahciv <See.above@aol.com> |
|---|---|
| Date | 2016-05-10 12:34 +0000 |
| Message-ID | <PM0005327C2668A564@aca4069b.ipt.aol.com> |
| In reply to | #163463 |
Morten Reistad wrote: > In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, > jmfbahciv <See.above@aol.com> wrote: >>Morten Reistad wrote: >>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>> jmfbahciv <See.above@aol.com> wrote: >>>>Gene Wirchenko wrote: >>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>> <no.spam@thank.you.invalid> wrote: >>>>> >>>>> [snip] >>>>> >>>>>>There would be no OS, unless limited to the 10 or so calls that CP/M had, >>>>>>so allowing me to proceed with complete freedom in whatever direction >>>>>>I chose. >>>>> >>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>> more. >>>> >>>>If you need a network, there would be triple the calls. >>> >>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>extra >>> code in open,read,write and close would do the trick. You would want a >>> select/poll call, but that should already be there. >>> >>> That makes half a dozen new calls, and extra code in half a dozen more. >>> >>> Networking is not that hard. We learned how to do this right by ca 1980. >>> >>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>> >>> Legacy, not-done-quite-right protocols would need more calls. >> >>Hence my too high estimate. And, for efficiency, it might be better to have >>a new call rather than trying to use an old call with new arguments. >> >>/BAH > > No; have general calls for open,read,write,close and a select/poll (waits > until something happens, can be many streams providing input or output buffers > accepting data, or exceptions). For the servers you need bind/listen/accept, > and for more elaborate network stuff you need connect and send/receive. > That is it. Plus some statistics and management calls for root/admin. But they > can be in sysctl etc. Half a dozen new calls and expansion of half a dozen more. > KISS. > > Also, nfs uses mount/umount just like any other file system. So does afs. > And zfs. The magic is in the file system driver in the kernel. If a call has to be extended, wouldn't it be better to create a new call instead of trying to retrofit the new code into old design/code? It would be KISS for backwards compatibility; nothing old breaks. Note that this is a philosophical discussion. /BAH
[toc] | [prev] | [next] | [standalone]
| From | cb@elaine.df.lth.se (Christian Brunschen) |
|---|---|
| Date | 2016-05-10 13:42 +0000 |
| Message-ID | <ngsofh$dv7$1@dont-email.me> |
| In reply to | #163509 |
In article <PM0005327C2668A564@aca4069b.ipt.aol.com>, jmfbahciv <See.above@aol.com> wrote: >Morten Reistad wrote: >> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >> jmfbahciv <See.above@aol.com> wrote: >>>Morten Reistad wrote: >>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>> jmfbahciv <See.above@aol.com> wrote: >>>>>Gene Wirchenko wrote: >>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>> <no.spam@thank.you.invalid> wrote: >>>>>> >>>>>> [snip] >>>>>> >>>>>>>There would be no OS, unless limited to the 10 or so calls that CP/M >had, >>>>>>>so allowing me to proceed with complete freedom in whatever direction >>>>>>>I chose. >>>>>> >>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>> more. >>>>> >>>>>If you need a network, there would be triple the calls. >>>> >>>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>>extra >>>> code in open,read,write and close would do the trick. You would want a >>>> select/poll call, but that should already be there. >>>> >>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>> >>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>> >>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>> >>>> Legacy, not-done-quite-right protocols would need more calls. >>> >>>Hence my too high estimate. And, for efficiency, it might be better to have >>>a new call rather than trying to use an old call with new arguments. >>> >>>/BAH >> >> No; have general calls for open,read,write,close and a select/poll (waits >> until something happens, can be many streams providing input or output >buffers >> accepting data, or exceptions). For the servers you need bind/listen/accept, >> and for more elaborate network stuff you need connect and send/receive. >> That is it. Plus some statistics and management calls for root/admin. But >they >> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen >more. >> KISS. >> >> Also, nfs uses mount/umount just like any other file system. So does afs. >> And zfs. The magic is in the file system driver in the kernel. > >If a call has to be extended, wouldn't it be better to create a new call >instead of trying to retrofit the new code into old design/code? It would >be KISS for backwards compatibility; nothing old breaks. No; the better thing is to design the abstractions to be flexible enough to be possible to adapt to different kinds of things. That can be easier in hindsight than in advance, though there are examples of it working well. Unix's read() and write() calls were originally designed for files, but were flexible enough to be used for network connections as well; thus, all code that could read sequentially from or write to an open file, can now also read from (or write to) an open network connection. The magic that tells the different kinds of data streams apart (file vs inter-process pipe vs across-the-network socket) lives inside the kernel. > >Note that this is a philosophical discussion. > >/BAH // Christian
[toc] | [prev] | [next] | [standalone]
| From | jmfbahciv <See.above@aol.com> |
|---|---|
| Date | 2016-05-11 12:18 +0000 |
| Message-ID | <PM0005328FF6F3B9E3@aca46c75.ipt.aol.com> |
| In reply to | #163517 |
Christian Brunschen wrote: > In article <PM0005327C2668A564@aca4069b.ipt.aol.com>, > jmfbahciv <See.above@aol.com> wrote: >>Morten Reistad wrote: >>> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >>> jmfbahciv <See.above@aol.com> wrote: >>>>Morten Reistad wrote: >>>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>>> jmfbahciv <See.above@aol.com> wrote: >>>>>>Gene Wirchenko wrote: >>>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>>> <no.spam@thank.you.invalid> wrote: >>>>>>> >>>>>>> [snip] >>>>>>> >>>>>>>>There would be no OS, unless limited to the 10 or so calls that CP/M >>had, >>>>>>>>so allowing me to proceed with complete freedom in whatever direction >>>>>>>>I chose. >>>>>>> >>>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>>> more. >>>>>> >>>>>>If you need a network, there would be triple the calls. >>>>> >>>>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>>>extra >>>>> code in open,read,write and close would do the trick. You would want a >>>>> select/poll call, but that should already be there. >>>>> >>>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>>> >>>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>>> >>>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>>> >>>>> Legacy, not-done-quite-right protocols would need more calls. >>>> >>>>Hence my too high estimate. And, for efficiency, it might be better to have >>>>a new call rather than trying to use an old call with new arguments. >>>> >>>>/BAH >>> >>> No; have general calls for open,read,write,close and a select/poll (waits >>> until something happens, can be many streams providing input or output >>buffers >>> accepting data, or exceptions). For the servers you need bind/listen/accept, >>> and for more elaborate network stuff you need connect and send/receive. >>> That is it. Plus some statistics and management calls for root/admin. But >>they >>> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen >>more. >>> KISS. >>> >>> Also, nfs uses mount/umount just like any other file system. So does afs. >>> And zfs. The magic is in the file system driver in the kernel. >> >>If a call has to be extended, wouldn't it be better to create a new call >>instead of trying to retrofit the new code into old design/code? It would >>be KISS for backwards compatibility; nothing old breaks. > > No; the better thing is to design the abstractions to be flexible enough > to be possible to adapt to different kinds of things. That can be easier > in hindsight than in advance, though there are examples of it working well. > > Unix's read() and write() calls were originally designed for files, but > were flexible enough to be used for network connections as well; thus, > all code that could read sequentially from or write to an open file, > can now also read from (or write to) an open network connection. > > The magic that tells the different kinds of data streams apart (file vs > inter-process pipe vs across-the-network socket) lives inside the kernel. By the time, Unix was begun, disk files, directories and multiple users existed. There were enough implementations of these aspects of computing that Unix' designers could skip the hard part about learning something completely new. /BAH
[toc] | [prev] | [next] | [standalone]
| From | scott@slp53.sl.home (Scott Lurndal) |
|---|---|
| Date | 2016-05-11 13:19 +0000 |
| Message-ID | <cjGYy.30038$H_5.3409@fx18.iad> |
| In reply to | #163555 |
jmfbahciv <See.above@aol.com> writes: >> The magic that tells the different kinds of data streams apart (file vs >> inter-process pipe vs across-the-network socket) lives inside the kernel. > >By the time, Unix was begun, disk files, directories and multiple users >existed. There were enough implementations of these aspects of computing >that Unix' designers could skip the hard part about learning something >completely new. All those things existed prior to TOPS-10, as well. So your point is moot (and mute). I note that the Burroughs systems had no problems with proliferation of interfaces due to new hardware capabilities. The same 'write' BCT (system call) was used for all file types, and internally in the MCP, we had a VFS-like layer that handled specific types of hardware (magtape, disk, datacomm, timesharing stations, cards, inter-system communications (ISC)).
[toc] | [prev] | [next] | [standalone]
| From | Ahem A Rivet's Shot <steveo@eircom.net> |
|---|---|
| Date | 2016-05-11 16:57 +0100 |
| Message-ID | <20160511165700.28f465cb02b53a840649c6ef@eircom.net> |
| In reply to | #163563 |
On Wed, 11 May 2016 13:19:04 GMT scott@slp53.sl.home (Scott Lurndal) wrote: > jmfbahciv <See.above@aol.com> writes: > > >> The magic that tells the different kinds of data streams apart (file vs > >> inter-process pipe vs across-the-network socket) lives inside the > >> kernel. > > > >By the time, Unix was begun, disk files, directories and multiple users > >existed. There were enough implementations of these aspects of computing > >that Unix' designers could skip the hard part about learning something > >completely new. > > All those things existed prior to TOPS-10, as well. So your point is Multics introduced the concept of a filesystem and that to the filesystem all files were just byte streams. I thought TOPS-10 was earlier than that. It's having good abstractions that makes things clean. -- Steve O'Hara-Smith | Directable Mirror Arrays C:>WIN | A better way to focus the sun The computer obeys and wins. | licences available see You lose and Bill collects. | http://www.sohara.org/
[toc] | [prev] | [next] | [standalone]
| From | scott@slp53.sl.home (Scott Lurndal) |
|---|---|
| Date | 2016-05-11 16:20 +0000 |
| Message-ID | <xZIYy.99117$6X3.91407@fx34.iad> |
| In reply to | #163578 |
Ahem A Rivet's Shot <steveo@eircom.net> writes: >On Wed, 11 May 2016 13:19:04 GMT >scott@slp53.sl.home (Scott Lurndal) wrote: > >> jmfbahciv <See.above@aol.com> writes: >> >> >> The magic that tells the different kinds of data streams apart (file vs >> >> inter-process pipe vs across-the-network socket) lives inside the >> >> kernel. >> > >> >By the time, Unix was begun, disk files, directories and multiple users >> >existed. There were enough implementations of these aspects of computing >> >that Unix' designers could skip the hard part about learning something >> >completely new. >> >> All those things existed prior to TOPS-10, as well. So your point is > > Multics introduced the concept of a filesystem and that to the BAH noted disk files, directories and multiple users. All of these existed on the Burroughs systems in the early 60's (albeit under different names e.g. catalog instead of directory, pack family instead of filesystem).
[toc] | [prev] | [next] | [standalone]
| From | Ahem A Rivet's Shot <steveo@eircom.net> |
|---|---|
| Date | 2016-05-11 17:38 +0100 |
| Message-ID | <20160511173810.88ba46186c570dd3533f31d9@eircom.net> |
| In reply to | #163580 |
On Wed, 11 May 2016 16:20:45 GMT scott@slp53.sl.home (Scott Lurndal) wrote: > Ahem A Rivet's Shot <steveo@eircom.net> writes: > >On Wed, 11 May 2016 13:19:04 GMT > >scott@slp53.sl.home (Scott Lurndal) wrote: > > > >> jmfbahciv <See.above@aol.com> writes: > >> > >> >> The magic that tells the different kinds of data streams apart > >> >> (file vs inter-process pipe vs across-the-network socket) lives > >> >> inside the kernel. > >> > > >> >By the time, Unix was begun, disk files, directories and multiple > >> >users existed. There were enough implementations of these aspects of > >> >computing that Unix' designers could skip the hard part about > >> >learning something completely new. > >> > >> All those things existed prior to TOPS-10, as well. So your point is > > > > Multics introduced the concept of a filesystem and that to the > > BAH noted disk files, directories and multiple users. All of these > existed on the Burroughs systems in the early 60's (albeit under > different names e.g. catalog instead of directory, pack family instead > of filesystem). Yes they did and anyone could have made the simplifying observations that led to the powerful abstractions introduced in Multics, but they didn't. That was the key thing that the Unix developers had that nobody before had. -- Steve O'Hara-Smith | Directable Mirror Arrays C:>WIN | A better way to focus the sun The computer obeys and wins. | licences available see You lose and Bill collects. | http://www.sohara.org/
[toc] | [prev] | [next] | [standalone]
| From | Morten Reistad <first@last.name.invalid> |
|---|---|
| Date | 2016-05-10 15:49 +0200 |
| Message-ID | <6h590d-9li.ln1@sambook.reistad.name> |
| In reply to | #163509 |
In article <PM0005327C2668A564@aca4069b.ipt.aol.com>, jmfbahciv <See.above@aol.com> wrote: >Morten Reistad wrote: >> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >> jmfbahciv <See.above@aol.com> wrote: >>>Morten Reistad wrote: >>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>> jmfbahciv <See.above@aol.com> wrote: >>>>>Gene Wirchenko wrote: >>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>> <no.spam@thank.you.invalid> wrote: >>>>>> >>>>>> [snip] >>>>>> >>>>>>>There would be no OS, unless limited to the 10 or so calls that CP/M >had, >>>>>>>so allowing me to proceed with complete freedom in whatever direction >>>>>>>I chose. >>>>>> >>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>> more. >>>>> >>>>>If you need a network, there would be triple the calls. >>>> >>>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>>extra >>>> code in open,read,write and close would do the trick. You would want a >>>> select/poll call, but that should already be there. >>>> >>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>> >>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>> >>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>> >>>> Legacy, not-done-quite-right protocols would need more calls. >>> >>>Hence my too high estimate. And, for efficiency, it might be better to have >>>a new call rather than trying to use an old call with new arguments. >>> >>>/BAH >> >> No; have general calls for open,read,write,close and a select/poll (waits >> until something happens, can be many streams providing input or output >buffers >> accepting data, or exceptions). For the servers you need bind/listen/accept, >> and for more elaborate network stuff you need connect and send/receive. >> That is it. Plus some statistics and management calls for root/admin. But >they >> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen >more. >> KISS. >> >> Also, nfs uses mount/umount just like any other file system. So does afs. >> And zfs. The magic is in the file system driver in the kernel. > >If a call has to be extended, wouldn't it be better to create a new call >instead of trying to retrofit the new code into old design/code? It would >be KISS for backwards compatibility; nothing old breaks. > >Note that this is a philosophical discussion. Showing the side from one of the most important distinctions between tops10 and tops20. Device independence. A write is a write is a write(). It works on 30 different file systems, nearly a dozen different network stacks, and to another two dozen or so different IO devices, from serial and parallell ports, via raw devices to virtual stuff like raids, etc. open,close,read,write and select are the main vehicles for i/o in all of posix. Windows and VMS included, at least when you run posix code. -- mrr
[toc] | [prev] | [next] | [standalone]
| From | jmfbahciv <See.above@aol.com> |
|---|---|
| Date | 2016-05-11 12:18 +0000 |
| Message-ID | <PM0005328FEC5B6EF4@aca46c75.ipt.aol.com> |
| In reply to | #163518 |
Morten Reistad wrote: > In article <PM0005327C2668A564@aca4069b.ipt.aol.com>, > jmfbahciv <See.above@aol.com> wrote: >>Morten Reistad wrote: >>> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >>> jmfbahciv <See.above@aol.com> wrote: >>>>Morten Reistad wrote: >>>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>>> jmfbahciv <See.above@aol.com> wrote: >>>>>>Gene Wirchenko wrote: >>>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>>> <no.spam@thank.you.invalid> wrote: >>>>>>> >>>>>>> [snip] >>>>>>> >>>>>>>>There would be no OS, unless limited to the 10 or so calls that CP/M >>had, >>>>>>>>so allowing me to proceed with complete freedom in whatever direction >>>>>>>>I chose. >>>>>>> >>>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>>> more. >>>>>> >>>>>>If you need a network, there would be triple the calls. >>>>> >>>>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>>>extra >>>>> code in open,read,write and close would do the trick. You would want a >>>>> select/poll call, but that should already be there. >>>>> >>>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>>> >>>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>>> >>>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>>> >>>>> Legacy, not-done-quite-right protocols would need more calls. >>>> >>>>Hence my too high estimate. And, for efficiency, it might be better to have >>>>a new call rather than trying to use an old call with new arguments. >>>> >>>>/BAH >>> >>> No; have general calls for open,read,write,close and a select/poll (waits >>> until something happens, can be many streams providing input or output >>buffers >>> accepting data, or exceptions). For the servers you need bind/listen/accept, >>> and for more elaborate network stuff you need connect and send/receive. >>> That is it. Plus some statistics and management calls for root/admin. But >>they >>> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen >>more. >>> KISS. >>> >>> Also, nfs uses mount/umount just like any other file system. So does afs. >>> And zfs. The magic is in the file system driver in the kernel. >> >>If a call has to be extended, wouldn't it be better to create a new call >>instead of trying to retrofit the new code into old design/code? It would >>be KISS for backwards compatibility; nothing old breaks. >> >>Note that this is a philosophical discussion. > > Showing the side from one of the most important distinctions between tops10 > and tops20. <grin> Not quite. -20 had problems, too. > > Device independence. > > A write is a write is a write(). > > It works on 30 different file systems, nearly a dozen different network > stacks, and to another two dozen or so different IO devices, from serial > and parallell ports, via raw devices to virtual stuff like raids, etc. > > open,close,read,write and select are the main vehicles for i/o in all > of posix. Windows and VMS included, at least when you run posix code. Logical names, structures, files, directories and sub-directories were not in TOPS-10 before it was TOPS-10. Even the IBM card-based systems dealt with direct I/O and took a long time to acclimate to the "features" of a disk. /BAH
[toc] | [prev] | [next] | [standalone]
| From | Peter Flass <peter_flass@yahoo.com> |
|---|---|
| Date | 2016-05-10 11:07 -0400 |
| Message-ID | <293972997.484585386.453086.peter_flass-yahoo.com@news.eternal-september.org> |
| In reply to | #163509 |
jmfbahciv <See.above@aol.com> wrote: > Morten Reistad wrote: >> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >> jmfbahciv <See.above@aol.com> wrote: >>> Morten Reistad wrote: >>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>> jmfbahciv <See.above@aol.com> wrote: >>>>> Gene Wirchenko wrote: >>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>> <no.spam@thank.you.invalid> wrote: >>>>>> >>>>>> [snip] >>>>>> >>>>>>> There would be no OS, unless limited to the 10 or so calls that CP/M > had, >>>>>>> so allowing me to proceed with complete freedom in whatever direction >>>>>>> I chose. >>>>>> >>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>> more. >>>>> >>>>> If you need a network, there would be triple the calls. >>>> >>>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>> extra >>>> code in open,read,write and close would do the trick. You would want a >>>> select/poll call, but that should already be there. >>>> >>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>> >>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>> >>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>> >>>> Legacy, not-done-quite-right protocols would need more calls. >>> >>> Hence my too high estimate. And, for efficiency, it might be better to have >>> a new call rather than trying to use an old call with new arguments. >>> >>> /BAH >> >> No; have general calls for open,read,write,close and a select/poll (waits >> until something happens, can be many streams providing input or output > buffers >> accepting data, or exceptions). For the servers you need bind/listen/accept, >> and for more elaborate network stuff you need connect and send/receive. >> That is it. Plus some statistics and management calls for root/admin. But > they >> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen > more. >> KISS. >> >> Also, nfs uses mount/umount just like any other file system. So does afs. >> And zfs. The magic is in the file system driver in the kernel. > > If a call has to be extended, wouldn't it be better to create a new call > instead of trying to retrofit the new code into old design/code? It would > be KISS for backwards compatibility; nothing old breaks. > > Note that this is a philosophical discussion. > If the interface had been designed right originally this would not be a problem. -- Pete
[toc] | [prev] | [next] | [standalone]
| From | jmfbahciv <See.above@aol.com> |
|---|---|
| Date | 2016-05-11 12:18 +0000 |
| Message-ID | <PM0005328FDCB4E2D4@aca46c75.ipt.aol.com> |
| In reply to | #163521 |
Peter Flass wrote: > jmfbahciv <See.above@aol.com> wrote: >> Morten Reistad wrote: >>> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >>> jmfbahciv <See.above@aol.com> wrote: >>>> Morten Reistad wrote: >>>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>>> jmfbahciv <See.above@aol.com> wrote: >>>>>> Gene Wirchenko wrote: >>>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>>> <no.spam@thank.you.invalid> wrote: >>>>>>> >>>>>>> [snip] >>>>>>> >>>>>>>> There would be no OS, unless limited to the 10 or so calls that CP/M >> had, >>>>>>>> so allowing me to proceed with complete freedom in whatever direction >>>>>>>> I chose. >>>>>>> >>>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>>> more. >>>>>> >>>>>> If you need a network, there would be triple the calls. >>>>> >>>>> No, assign/bind, connect, listen, accept, plus send and receive; and some >>>> extra >>>>> code in open,read,write and close would do the trick. You would want a >>>>> select/poll call, but that should already be there. >>>>> >>>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>>> >>>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>>> >>>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>>> >>>>> Legacy, not-done-quite-right protocols would need more calls. >>>> >>>> Hence my too high estimate. And, for efficiency, it might be better to have >>>> a new call rather than trying to use an old call with new arguments. >>>> >>>> /BAH >>> >>> No; have general calls for open,read,write,close and a select/poll (waits >>> until something happens, can be many streams providing input or output >> buffers >>> accepting data, or exceptions). For the servers you need bind/listen/accept, >>> and for more elaborate network stuff you need connect and send/receive. >>> That is it. Plus some statistics and management calls for root/admin. But >> they >>> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen >> more. >>> KISS. >>> >>> Also, nfs uses mount/umount just like any other file system. So does afs. >>> And zfs. The magic is in the file system driver in the kernel. >> >> If a call has to be extended, wouldn't it be better to create a new call >> instead of trying to retrofit the new code into old design/code? It would >> be KISS for backwards compatibility; nothing old breaks. >> >> Note that this is a philosophical discussion. >> > > > If the interface had been designed right originally this would not be a > problem. > <grin> The interface had been designed "right" at the time. Noone was prescient enough to predict all the extensions required by new hardware designs and software demands. /BAH
[toc] | [prev] | [next] | [standalone]
| From | scott@slp53.sl.home (Scott Lurndal) |
|---|---|
| Date | 2016-05-11 13:16 +0000 |
| Message-ID | <ngGYy.30037$H_5.16559@fx18.iad> |
| In reply to | #163554 |
jmfbahciv <See.above@aol.com> writes: >Peter Flass wrote: >> >> If the interface had been designed right originally this would not be a >> problem. >> > ><grin> The interface had been designed "right" at the time. Noone >was prescient enough to predict all the extensions required by new >hardware designs and software demands. Ken and Dennis had no problems designing interfaces that "predicted" future needs, quite successfully, as it turned out, as their interfaces are still in widespread use forty years later.
[toc] | [prev] | [next] | [standalone]
| From | Michael Black <et472@ncf.ca> |
|---|---|
| Date | 2016-05-11 09:45 -0400 |
| Message-ID | <alpine.LNX.2.02.1605110944020.30685@darkstar.example.org> |
| In reply to | #163562 |
On Wed, 11 May 2016, Scott Lurndal wrote: > jmfbahciv <See.above@aol.com> writes: >> Peter Flass wrote: > >>> >>> If the interface had been designed right originally this would not be a >>> problem. >>> >> >> <grin> The interface had been designed "right" at the time. Noone >> was prescient enough to predict all the extensions required by new >> hardware designs and software demands. > > Ken and Dennis had no problems designing interfaces that "predicted" > future needs, quite successfully, as it turned out, as their interfaces > are still in widespread use forty years later. > I think it's one of those things people look at from different angles. Unix might seem to some like an almost fifty year old operating system that is stale. And others see it as something that came in early, but still covers the territory well. I'm in the second camp. Besides, the effort has been put into making it better, while some commercial operating systems have kept reinventing themselves, mostly for the sake of selling a new version to the customers. Michael
[toc] | [prev] | [next] | [standalone]
| From | Andrew Swallow <am.swallow@btopenworld.com> |
|---|---|
| Date | 2016-05-11 16:22 +0100 |
| Message-ID | <B_KdnQZaUJp90K7KnZ2dnUU78UOdnZ2d@giganews.com> |
| In reply to | #163562 |
On 11/05/2016 14:16, Scott Lurndal wrote: > jmfbahciv <See.above@aol.com> writes: >> Peter Flass wrote: > >>> >>> If the interface had been designed right originally this would not be a >>> problem. >>> >> >> <grin> The interface had been designed "right" at the time. Noone >> was prescient enough to predict all the extensions required by new >> hardware designs and software demands. > > Ken and Dennis had no problems designing interfaces that "predicted" > future needs, quite successfully, as it turned out, as their interfaces > are still in widespread use forty years later. > The operating system had to support COBOL's read, write, open, close, rewind and seek. Although going from record I/O to byte stream was a significant change.
[toc] | [prev] | [next] | [standalone]
| From | scott@slp53.sl.home (Scott Lurndal) |
|---|---|
| Date | 2016-05-11 15:47 +0000 |
| Message-ID | <UtIYy.99113$6X3.10969@fx34.iad> |
| In reply to | #163571 |
Andrew Swallow <am.swallow@btopenworld.com> writes: >On 11/05/2016 14:16, Scott Lurndal wrote: >> jmfbahciv <See.above@aol.com> writes: >>> Peter Flass wrote: >> >>>> >>>> If the interface had been designed right originally this would not be a >>>> problem. >>>> >>> >>> <grin> The interface had been designed "right" at the time. Noone >>> was prescient enough to predict all the extensions required by new >>> hardware designs and software demands. >> >> Ken and Dennis had no problems designing interfaces that "predicted" >> future needs, quite successfully, as it turned out, as their interfaces >> are still in widespread use forty years later. >> > >The operating system had to support COBOL's read, write, open, close, >rewind and seek. Although going from record I/O to byte stream was a >significant change. There are several versions of COBOL that work just fine on unix and unix-like systems. Microfocus comes immediately to mind. In any case, record oriented I/O can _always_ be built on top of byte-stream I/O in run-time libraries or even in the application itself. No need to include it in the OS. Even in VMS, RMS wasn't part of the kernel, it executed in a different privilege ring.
[toc] | [prev] | [next] | [standalone]
| From | Peter Flass <peter_flass@yahoo.com> |
|---|---|
| Date | 2016-05-11 10:52 -0400 |
| Message-ID | <1729324411.484670756.076816.peter_flass-yahoo.com@news.eternal-september.org> |
| In reply to | #163554 |
jmfbahciv <See.above@aol.com> wrote: > Peter Flass wrote: >> jmfbahciv <See.above@aol.com> wrote: >>> Morten Reistad wrote: >>>> In article <PM0005326827D1037E@aca40e74.ipt.aol.com>, >>>> jmfbahciv <See.above@aol.com> wrote: >>>>> Morten Reistad wrote: >>>>>> In article <PM00053253F3BBD6C0@aca41104.ipt.aol.com>, >>>>>> jmfbahciv <See.above@aol.com> wrote: >>>>>>> Gene Wirchenko wrote: >>>>>>>> On Sat, 7 May 2016 10:39:56 +0100, "gareth G4SDW GQRP #3339" >>>>>>>> <no.spam@thank.you.invalid> wrote: >>>>>>>> >>>>>>>> [snip] >>>>>>>> >>>>>>>>> There would be no OS, unless limited to the 10 or so calls that CP/M >>> had, >>>>>>>>> so allowing me to proceed with complete freedom in whatever direction >>>>>>>>> I chose. >>>>>>>> >>>>>>>> Even CP/M 2.2 had a few dozen calls. Later versions had even >>>>>>>> more. >>>>>>> >>>>>>> If you need a network, there would be triple the calls. >>>>>> >>>>>> No, assign/bind, connect, listen, accept, plus send and receive; and > some >>>>> extra >>>>>> code in open,read,write and close would do the trick. You would want a >>>>>> select/poll call, but that should already be there. >>>>>> >>>>>> That makes half a dozen new calls, and extra code in half a dozen more. >>>>>> >>>>>> Networking is not that hard. We learned how to do this right by ca 1980. >>>>>> >>>>>> The rest is implementation detail; roughly done in the *n*xes by 1985. >>>>>> >>>>>> Legacy, not-done-quite-right protocols would need more calls. >>>>> >>>>> Hence my too high estimate. And, for efficiency, it might be better to > have >>>>> a new call rather than trying to use an old call with new arguments. >>>>> >>>>> /BAH >>>> >>>> No; have general calls for open,read,write,close and a select/poll (waits >>>> until something happens, can be many streams providing input or output >>> buffers >>>> accepting data, or exceptions). For the servers you need > bind/listen/accept, >>>> and for more elaborate network stuff you need connect and send/receive. >>>> That is it. Plus some statistics and management calls for root/admin. But >>> they >>>> can be in sysctl etc. Half a dozen new calls and expansion of half a dozen >>> more. >>>> KISS. >>>> >>>> Also, nfs uses mount/umount just like any other file system. So does afs. >>>> And zfs. The magic is in the file system driver in the kernel. >>> >>> If a call has to be extended, wouldn't it be better to create a new call >>> instead of trying to retrofit the new code into old design/code? It would >>> be KISS for backwards compatibility; nothing old breaks. >>> >>> Note that this is a philosophical discussion. >>> >> >> >> If the interface had been designed right originally this would not be a >> problem. >> > > <grin> The interface had been designed "right" at the time. Noone > was prescient enough to predict all the extensions required by new > hardware designs and software demands. > > /BAH > The idea is not to lock yourself in. If you pass arguments to the OS pass a version, so you can go to version 2 transparently. If you pass the address of an area to receive a response put either a version or an area length in that so the OS knows what to pass back to you. -- Pete
[toc] | [prev] | [next] | [standalone]
| From | scott@slp53.sl.home (Scott Lurndal) |
|---|---|
| Date | 2016-05-11 15:49 +0000 |
| Message-ID | <MvIYy.99114$6X3.37557@fx34.iad> |
| In reply to | #163569 |
Peter Flass <peter_flass@yahoo.com> writes: >jmfbahciv <See.above@aol.com> wrote: >> >> <grin> The interface had been designed "right" at the time. Noone >> was prescient enough to predict all the extensions required by new >> hardware designs and software demands. >> >> /BAH >> > >The idea is not to lock yourself in. If you pass arguments to the OS pass >a version, so you can go to version 2 transparently. If you pass the >address of an area to receive a response put either a version or an area >length in that so the OS knows what to pass back to you. VMS did this well. Particuarly with $GETJPI, $GETDVI & friends.
[toc] | [prev] | [next] | [standalone]
| From | Ahem A Rivet's Shot <steveo@eircom.net> |
|---|---|
| Date | 2016-05-10 17:44 +0100 |
| Message-ID | <20160510174447.d79a8516c7747b9ea1a5ce42@eircom.net> |
| In reply to | #163509 |
On 10 May 2016 12:34:19 GMT jmfbahciv <See.above@aol.com> wrote: > If a call has to be extended, wouldn't it be better to create a new call > instead of trying to retrofit the new code into old design/code? It would > be KISS for backwards compatibility; nothing old breaks. The trick is to design the layer beneath the call so that it can be implemented on a wide variety of substrates. That way the call can be used with anything that implements the APIs it uses to communicate to the layer below without touching the code of the call. Now recurse and you get a system which is pluggable at many levels depending only on correct implementation of well defined APIs at each level. -- Steve O'Hara-Smith | Directable Mirror Arrays C:>WIN | A better way to focus the sun The computer obeys and wins. | licences available see You lose and Bill collects. | http://www.sohara.org/
[toc] | [prev] | [next] | [standalone]
Page 13 of 14 — ← Prev page 1 … 11 12 [13] 14 Next page →
Back to top | Article view | alt.folklore.computers
csiph-web