Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

swap and page files

161 views
Skip to first unread message

Phillip Helbig---undress to reply

unread,
Dec 27, 2012, 5:33:33 AM12/27/12
to
I have the regular swap and page files on the system disk. In addition,
I have some additional (larger) ones on another disk. These are
installed in SYPAGSWPFILES.COM.

In MODPARAMS.DAT (example from one node), I have:

PAGEFILE2_NAME="DISK$SWAPPAGE_110:[SYS110.SYSEXE]PAGEFILE2.SYS"
MIN_PAGEFILE2_SIZE=2101200
SWAPFILE2_NAME="DISK$SWAPPAGE_110:[SYS110.SYSEXE]SWAPFILE2.SYS"
MIN_SWAPFILE2_SIZE=36500

and in a file included by MODPARAMS.DAT by all nodes, I have:

PAGEFILE1_NAME="SYS$SYSTEM:PAGEFILE.SYS"
PAGEFILE1_SIZE=0 ! tell AUTOGEN not to resize primary page file
SWAPFILE1_NAME="SYS$SYSTEM:SWAPFILE.SYS"
SWAPFILE1_SIZE=0 ! tell AUTOGEN not to resize primary swap file

SHOW MEMORY tells me:

Swap File Usage (8KB pages): Index Free Size
DISK$ALPHASYS_110:[SYS110.SYSEXE]SWAPFILE.SYS
1 1272 1272
DISK$SWAPPAGE_110:[SYS110.SYSEXE]SWAPFILE2.SYS;1
2 2280 2280

Total size of all swap files: 3552

Paging File Usage (8KB pages): Index Free Size
DISK$SWAPPAGE_110:[SYS110.SYSEXE]PAGEFILE2.SYS;1
253 131319 131320
DISK$ALPHASYS_110:[SYS110.SYSEXE]PAGEFILE.SYS
254 109664 109664

Total size of all paging files: 240984
Total committed paging file usage: 12394

The names are correct. Apparently the Index counts swap files starting
at 1 ascending and page files starting at 254 descending.

What is confusing is that the system has reversed the primary and
secondary page files. Thus, AUTOGEN will size the primary files, even
though I tell it not to. This is also shown in AGEN$FEEDBACK.DAT:

SWAPFILE1_NAME = "SYS$SYSTEM:SWAPFILE.SYS"
SWAPFILE1_PEAK = 0
SWAPFILE2_NAME = "DISK$SWAPPAGE_110:[SYS110.SYSEXE]SWAPFILE2.SYS;1"
SWAPFILE2_PEAK = 0
PAGEFILE1_NAME = "DISK$SWAPPAGE_110:[SYS110.SYSEXE]PAGEFILE2.SYS;1"
PAGEFILE1_PEAK = 7
PAGEFILE2_NAME = "SYS$SYSTEM:PAGEFILE.SYS"
PAGEFILE2_PEAK = 7

Yes, they aren't needed much; yes, I can manually size them and avoid
the sizing step in AUTOGEN. However, I want to understand what is going
wrong. Since they are set up according to the same scheme, any mistake
on my part should reverse the swap files as well, not just the page
files.

Apparently PAGEFILE1_NAME refers to the primary page file. I gather
that the actual names of secondary page files don't matter at all, since
they are always given explicitly. Thus, the fact that I have a "2" in
the name is neither here nor there and can't be the source of the
confusion.

SYPAGSWPFILES.COM says

$ ! Note that the name SWAPFILE1.SYS is reserved for use on
$ ! tailored VMS systems.

This refers to the file name, not the parameter SWAPFILE1_NAME, and I
don't have any file called SWAPFILE1.SYS, so that can't be the source of
confusion.

This shows up in AGEN$PARAMS.REPORT:

PAGEFILE1_SIZE information (for DISK$SWAPPAGE_110:[SYS110.SYSEXE]PAGEFILE2.SYS;1):
PAGEFILE2_SIZE information (for SYS$SYSTEM:PAGEFILE.SYS):
SWAPFILE1_SIZE information (for SYS$SYSTEM:SWAPFILE.SYS):
SWAPFILE2_SIZE information (for DISK$SWAPPAGE_110:[SYS110.SYSEXE]SWAPFILE2.SYS;1):

Where the reversal for the page files is seen again.

Any ideas what can cause this reversal? Any ideas how to fix it?

Stephen Hoffman

unread,
Dec 27, 2012, 8:39:18 AM12/27/12
to
On 2012-12-27 10:33:33 +0000, Phillip Helbig---undress to reply said:

> Any ideas what can cause this reversal? Any ideas how to fix it?

As the available memory is increased and as the workload is better
partitioned to avoid memory pile-ups, the need to configure larger and
multiple page and swap is reduced. Sure, there are cases where a
couple of files are definitely necessary and that primarily for
performance, but the process pile-ups haven't been nearly as severly in
recent years as the old VAX boxes were once overloaded, and
environments that require higher performance have added physical memory
or off-loaded the pile-ups during critical processing...

Preferably, there's no (hard) paging and swapping on an OpenVMS
system... But if there is...

1: "Buy more memory." Even virtual memory needs physical memory.

2: I'd get rid of the dinky system disk files[1], and go with fewer and
larger and preferably local pagefiles.

3: If the system is paging heavily or particularly if the system is
swapping, then there are probably other issues with the configuration
that should be addressed before paging and swapping files.

4: I'd rip out all of those settings in MODPARAMS here, set the sizes
of the page, swap and dump files to 0 to disable all adjustments by
AUTOGEN, and manage the file sizes manually. This configuration is
comparatively complex, and AUTOGEN likely won't do what would probably
be desired here if/when it decides to adjust anything. Less time on
getting the swap files and HBVS and HBMM and policies all "just so",
and more time on getting "larger-capacity" disks on-line, too.

5: I'd seriously consider getting rid of all system disk page, swap and
dump files and all but one of the off-system-disk pagefiles, and
(maybe) all but one (small) off-disk swapfile, and would set up DOSD if
the disks here are as small as they appear. This given the system disk
is apparently small-capacity, or full, or the dreaded both
small-capacity and full that's so common with older VMS disks these
days.


==

[1] or to one file, with a "different" name, that I could connect and
use if I really needed to boot with just the system disk around, and if
I didn't have enough memory to do what I needed without OpenVMS
encountering the need to page or (less desirably) swap.

--
Pure Personal Opinion | HoffmanLabs LLC

Phillip Helbig---undress to reply

unread,
Dec 27, 2012, 8:56:29 AM12/27/12
to
In article <kbhj26$707$1...@dont-email.me>, Stephen Hoffman
<seao...@hoffmanlabs.invalid> writes:

> As the available memory is increased and as the workload is better
> partitioned to avoid memory pile-ups, the need to configure larger and
> multiple page and swap is reduced. Sure, there are cases where a
> couple of files are definitely necessary and that primarily for
> performance, but the process pile-ups haven't been nearly as severly in
> recent years as the old VAX boxes were once overloaded, and
> environments that require higher performance have added physical memory
> or off-loaded the pile-ups during critical processing...

True. Actually, I rarely need them. In this case, the motivation is to
understand what went wrong.

> 2: I'd get rid of the dinky system disk files[1], and go with fewer and
> larger and preferably local pagefiles.

I like to keep them so that I can boot a node with just its system disk
for test purposes.

> 4: I'd rip out all of those settings in MODPARAMS here, set the sizes
> of the page, swap and dump files to 0 to disable all adjustments by
> AUTOGEN, and manage the file sizes manually. This configuration is
> comparatively complex, and AUTOGEN likely won't do what would probably
> be desired here if/when it decides to adjust anything. Less time on
> getting the swap files and HBVS and HBMM and policies all "just so",
> and more time on getting "larger-capacity" disks on-line, too.

Good suggestion, and actually what I've done, but I would still like to
understand this strange reversal.

> 5: I'd seriously consider getting rid of all system disk page, swap and
> dump files and all but one of the off-system-disk pagefiles, and
> (maybe) all but one (small) off-disk swapfile, and would set up DOSD if
> the disks here are as small as they appear. This given the system disk
> is apparently small-capacity, or full, or the dreaded both
> small-capacity and full that's so common with older VMS disks these
> days.

DOSD has to dump to a single disk, not a shadow set or a member thereof,
right? So, in the absence of RAID etc, one has a single point of
failure. :-(

Stephen Hoffman

unread,
Dec 27, 2012, 12:20:43 PM12/27/12
to
On 2012-12-27 13:56:29 +0000, Phillip Helbig---undress to reply said:

> In article <kbhj26$707$1...@dont-email.me>, Stephen Hoffman
> <seao...@hoffmanlabs.invalid> writes:
>
>> As the available memory is increased and as the workload is better
>> partitioned to avoid memory pile-ups, the need to configure larger and
>> multiple page and swap is reduced. Sure, there are cases where a
>> couple of files are definitely necessary and that primarily for
>> performance, but the process pile-ups haven't been nearly as severly in
>> recent years as the old VAX boxes were once overloaded, and
>> environments that require higher performance have added physical memory
>> or off-loaded the pile-ups during critical processing...
>
> True. Actually, I rarely need them. In this case, the motivation is to
> understand what went wrong.

Sometimes you're just so cute...

>> 2: I'd get rid of the dinky system disk files[1], and go with fewer and
>> larger and preferably local pagefiles.
>
> I like to keep them so that I can boot a node with just its system disk
> for test purposes.

VMS doesn't need files to do that. Not if you have a sane amount of memory.

>
>> 4: I'd rip out all of those settings in MODPARAMS here, set the sizes
>> of the page, swap and dump files to 0 to disable all adjustments by
>> AUTOGEN, and manage the file sizes manually. This configuration is
>> comparatively complex, and AUTOGEN likely won't do what would probably
>> be desired here if/when it decides to adjust anything. Less time on
>> getting the swap files and HBVS and HBMM and policies all "just so",
>> and more time on getting "larger-capacity" disks on-line, too.
>
> Good suggestion, and actually what I've done, but I would still like to
> understand this strange reversal.

How about "who cares, that's just the way it works?" Would that help?

>
>> 5: I'd seriously consider getting rid of all system disk page, swap and
>> dump files and all but one of the off-system-disk pagefiles, and
>> (maybe) all but one (small) off-disk swapfile, and would set up DOSD if
>> the disks here are as small as they appear. This given the system disk
>> is apparently small-capacity, or full, or the dreaded both
>> small-capacity and full that's so common with older VMS disks these
>> days.
>
> DOSD has to dump to a single disk, not a shadow set or a member thereof,
> right? So, in the absence of RAID etc, one has a single point of
> failure. :-(

I thought VMS never crashed?

Phillip Helbig---undress to reply

unread,
Dec 28, 2012, 1:59:00 AM12/28/12
to
In article <kbi01b$oiv$1...@dont-email.me>, Stephen Hoffman
<seao...@hoffmanlabs.invalid> writes:

> > Good suggestion, and actually what I've done, but I would still like to
> > understand this strange reversal.
>
> How about "who cares, that's just the way it works?" Would that help?

I specify the names explicitly in MODPARAMS.DAT and they are reversed in
AGEN$FEEDBACK.DAT, but only for page, not swap files. I wouldn't call
that "works". Perhaps there is a bug in the version of VMS I have, thus
my query here.

Yes, I probably don't need them. I don't "need" VMS at home. It's just
a hobby. There are worse hobbies.

MG

unread,
Dec 28, 2012, 8:30:08 AM12/28/12
to
On 28-dec-2012 7:59, Phillip Helbig---undress to reply wrote:
> Yes, I probably don't need them. I don't "need" VMS at home. It's
> just a hobby. There are worse hobbies.

You didn't specify it, you ought to know by now that you need to
specify such information each time and perhaps even repeat it a few
times before it sinks in. Don't you know that by now, after posting
here for all this time?

Besides, how productive or worthwhile is it to be using VMS anyway?
Isn't gazing at the latest Samsung or Apple 'appliance' a much more
sound and socially acceptable pass-time nowadays? We're obviously
stuck in the past, obsessed with unnecessary functionality. Dare I
call it, fundamentalism?

The worrisome lack of Google Ads throughout VMS is needless to say
also an immense tragedy.

- MG

Paul Sture

unread,
Dec 28, 2012, 9:02:03 AM12/28/12
to
In article <kbhk2d$8va$2...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

> DOSD has to dump to a single disk, not a shadow set or a member thereof,
> right? So, in the absence of RAID etc, one has a single point of
> failure. :-(

I used DOSD for a cluster of workstations which booted off a pair of
Alpha 4100 systems. We also configured those workstations to have their
page and swap files on their internal disks.

The reason for implementing DOSD on that cluster was to get rid of the
shadow merges on the system disk every time a workstation started.[1]

This was particularly problematic when all the satellites in the cluster
rebooted, for example as the result of a network outage. The load on
the system with 20+ workstations booting at once was such that
"something" in DECnet Phase V startup would time out and therefore
DECnet would not come up properly. The manual workaround before DOSD
was implemented was to drop one of the system disk shadow members until
the whole cluster was up.

In several years managing that cluster we only had a couple of
workstations which suffered local disk failures and they were both back
up and running within minutes of a new disk being fitted.

With mission critical uses, having complete workstations ready to
replace failed ones might be more appropriate than fitting RAID
controllers to each workstation. You are going to need those anyway in
case things other than disks fail.

[1] http://h71000.www7.hp.com/doc/731final/5423/5423pro_014.html

"virtual-unit: shadow master has changed. Dump file will be written if
system crashes. Volume Processing in progress.

Explanation: The shadowing software has determined a new master disk for
the system disk shadow set. You can write a dump file for this system
only if the master is the same disk as the one the system booted from.
This is because the boot drivers are not connected with the shadow
driver, and different boot drivers from the ones that interact with the
booted system disk might be needed to interact with the new master disk.
For example, a system disk could be served and also locally connected,
causing the served path to use different drivers from the local path."

--
Paul Sture

Paul Sture

unread,
Dec 28, 2012, 9:12:58 AM12/28/12
to
In article <50dd9ee7$0$3162$e4fe...@dreader36.news.xs4all.nl>,
MG <marc...@SPAMxs4all.nl> wrote:

> Besides, how productive or worthwhile is it to be using VMS anyway?
> Isn't gazing at the latest Samsung or Apple 'appliance' a much more
> sound and socially acceptable pass-time nowadays? We're obviously
> stuck in the past, obsessed with unnecessary functionality. Dare I
> call it, fundamentalism?

Whatever floats your boat, and Phillip clearly enjoys VMS. I am taking
note of Hoff's warnings about career opportunities passing you by if you
don't move on.

While Phillip was posting about free disk space I was researching to see
if I could grab more than one audio stream to disk simultaneously (with
mixed success). Oh, and testing deduplicated backups.

P.S. It's "pastime" not "pass-time".

--
Paul Sture

MG

unread,
Dec 28, 2012, 9:28:27 AM12/28/12
to
On 28-dec-2012 15:12, Paul Sture wrote:
> Whatever floats your boat, and Phillip clearly enjoys VMS. I am taking
> note of Hoff's warnings about career opportunities passing you by if you
> don't move on.

Ever heard of sarcasm? Also, what career opportunities?! Now you're
simply being facetious.


> While Phillip was posting about free disk space I was researching to see
> if I could grab more than one audio stream to disk simultaneously (with
> mixed success). Oh, and testing deduplicated backups.

Yes, okay...?


> P.S. It's "pastime" not "pass-time".

Mea culpa, I only ever heard it pronounced. English isn't exactly a
phonetic language.

- MG

VAXman-

unread,
Dec 28, 2012, 9:35:56 AM12/28/12
to
In article <50ddac92$0$3517$e4fe...@dreader37.news.xs4all.nl>, MG <marc...@SPAMxs4all.nl> writes:
>{...snip...}
>> P.S. It's "pastime" not "pass-time".
>
>Mea culpa, I only ever heard it pronounced. English isn't exactly a
>phonetic language.

Neither is Phrench! ;)

--
VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)ORG

Well I speak to machines with the voice of humanity.

Bill Gunshannon

unread,
Dec 28, 2012, 9:42:32 AM12/28/12
to
In article <00ACC867...@sendspamhere.org>,
VAXman- @SendSpamHere.ORG writes:
> In article <50ddac92$0$3517$e4fe...@dreader37.news.xs4all.nl>, MG <marc...@SPAMxs4all.nl> writes:
>>{...snip...}
>>> P.S. It's "pastime" not "pass-time".
>>
>>Mea culpa, I only ever heard it pronounced. English isn't exactly a
>>phonetic language.

Based on my experience with english :-) both wold be pronounced the same.
>
> Neither is Phrench! ;)

German, on the other hand.........

bill

--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>

David Froble

unread,
Dec 28, 2012, 1:00:33 PM12/28/12
to
Paul Sture wrote:

> With mission critical uses, having complete workstations ready to
> replace failed ones might be more appropriate than fitting RAID
> controllers to each workstation. You are going to need those anyway in
> case things other than disks fail.

As usual, simpler is better ....

Stephen Hoffman

unread,
Dec 28, 2012, 2:30:30 PM12/28/12
to
On 2012-12-28 06:59:00 +0000, Phillip Helbig---undress to reply said:

> In article <kbi01b$oiv$1...@dont-email.me>, Stephen Hoffman
> <seao...@hoffmanlabs.invalid> writes:
>
>>> Good suggestion, and actually what I've done, but I would still like to
>>> understand this strange reversal.
>>
>> How about "who cares, that's just the way it works?" Would that help?
>
> I specify the names explicitly in MODPARAMS.DAT and they are reversed in
> AGEN$FEEDBACK.DAT, but only for page, not swap files. I wouldn't call
> that "works". Perhaps there is a bug in the version of VMS I have, thus
> my query here.

Chainsaw out all that stuff and replace it with one file. Or a few
same-sized files on separate (non-system-disk) spindles. Or more
memory. Or a combination.

And FWIW, it's quite possible to run OpenVMS with no page file.

That's what the CD and DVD do, after all, and — unless the system is
very memory-constrained — should work just fine.

I've had a few (and rare) cases where I've created a "runt" page file
under a different and non-default name, and connected it at boot.

If you're interested in seeing what AUTOGEN is up to, the DCL is available.

> Yes, I probably don't need them. I don't "need" VMS at home. It's just
> a hobby. There are worse hobbies.

You've been at this retrocomputing stuff long enough that it's arguably
progressed past a "hobby" and become an "affliction". :-)

Stephen Hoffman

unread,
Dec 28, 2012, 2:34:18 PM12/28/12
to
On 2012-12-28 13:30:08 +0000, MG said:

> On 28-dec-2012 7:59, Phillip Helbig---undress to reply wrote:
>> Yes, I probably don't need them. I don't "need" VMS at home. It's
>> just a hobby. There are worse hobbies.
>
> You didn't specify it, you ought to know by now that you need to
> specify such information each time and perhaps even repeat it a few
> times before it sinks in. Don't you know that by now, after posting
> here for all this time?

Good idea.

> Besides, how productive or worthwhile is it to be using VMS anyway?

Yep; it gets tougher each year, too.

> Isn't gazing at the latest Samsung or Apple 'appliance' a much more
> sound and socially acceptable pass-time nowadays?

Far easier, certainly. Or to a more modern server-oriented system, if
you're hosting services for your clients.

> We're obviously
> stuck in the past, obsessed with unnecessary functionality.

Disagree. The functionality of the old gear is much lower.

> Dare I call it, fundamentalism?

Fundamentals, certainly. Retrocomputing lacks features.

> The worrisome lack of Google Ads throughout VMS is needless to say
> also an immense tragedy.

That omission is lost revenue for somebody, certainly.

Single Stage to Orbit

unread,
Dec 28, 2012, 2:15:58 PM12/28/12
to
On Fri, 2012-12-28 at 14:35 +0000, VAX...@SendSpamHere.ORG wrote:
> >Mea culpa, I only ever heard it pronounced. English isn't exactly a
> >phonetic language.
>
> Neither is Phrench! ;)

Oh gods, 1415 all over again :-)
--
Tactical Nuclear Kittens

Stephen Hoffman

unread,
Dec 28, 2012, 3:02:26 PM12/28/12
to
On 2012-12-28 14:12:58 +0000, Paul Sture said:

> In article <50dd9ee7$0$3162$e4fe...@dreader36.news.xs4all.nl>,
> MG <marc...@SPAMxs4all.nl> wrote:
>
>> Besides, how productive or worthwhile is it to be using VMS anyway?
>> Isn't gazing at the latest Samsung or Apple 'appliance' a much more
>> sound and socially acceptable pass-time nowadays? We're obviously
>> stuck in the past, obsessed with unnecessary functionality. Dare I
>> call it, fundamentalism?
>
> Whatever floats your boat, and Phillip clearly enjoys VMS. I am taking
> note of Hoff's warnings about career opportunities passing you by if you
> don't move on.

Ayup. HP BCS is prominently featuring their ProLiant DL980 box, as
well as their upcoming DragonHawk SD blades and the HydraLynx blades,
and pushing x86 servers with RAS features, and targetting Windows
Server and Linux as their platforms for their future of
mission-critical computing offerings, and has stated that
mission-critical x86 is where they're spending most of their R&D and
what they're full-speed working on.

> While Phillip was posting about free disk space I was researching to see
> if I could grab more than one audio stream to disk simultaneously (with
> mixed success).

FWIW (and don't know what you're up to with those audio streams), one
of the older Mac Mini boxes I deal with fairly regularly streams four
parallel OTA HDTV streams from GbE to disk. And it's not very busy.
That benchmark was also before the most recent upgrade, so that
customer's box can undoubtedly go rather faster.

Stephen Hoffman

unread,
Dec 28, 2012, 3:05:43 PM12/28/12
to
On 2012-12-28 14:28:27 +0000, MG said:

> On 28-dec-2012 15:12, Paul Sture wrote:
>> Whatever floats your boat, and Phillip clearly enjoys VMS. I am taking
>> note of Hoff's warnings about career opportunities passing you by if you
>> don't move on.
>
> Ever heard of sarcasm?

This is why we ask, and why we don't assume. (And why I inevitably get
in trouble when I do.)

>
>> P.S. It's "pastime" not "pass-time".
>
> Mea culpa, I only ever heard it pronounced. English isn't exactly a
> phonetic language.

It's also not the first language of various folks around here, and
humor and sarcasm often get missed or misinterpreted, and folks have
various degrees of experience with stuff and good and bad days, meaning
that details can and do get omitted, and "what's obvious to {somebody}"
may not be obvious to somebody else, nor is it certain what's obvious
is what's really going on.

Phillip Helbig---undress to reply

unread,
Dec 28, 2012, 4:02:24 PM12/28/12
to
In article <50dd9ee7$0$3162$e4fe...@dreader36.news.xs4all.nl>, MG
<marc...@SPAMxs4all.nl> writes:

> On 28-dec-2012 7:59, Phillip Helbig---undress to reply wrote:
> > Yes, I probably don't need them. I don't "need" VMS at home. It's
> > just a hobby. There are worse hobbies.
>
> You didn't specify it, you ought to know by now that you need to
> specify such information each time and perhaps even repeat it a few
> times before it sinks in. Don't you know that by now, after posting
> here for all this time?

Well, Hoff chides me for repeating stuff too much. I'm trying to strike
a balance, and be cute at the same time. :-|

Phillip Helbig---undress to reply

unread,
Dec 28, 2012, 4:34:51 PM12/28/12
to
In article <kbks0l$96j$1...@dont-email.me>, Stephen Hoffman
<seao...@hoffmanlabs.invalid> writes:

> >>> Good suggestion, and actually what I've done, but I would still like to
> >>> understand this strange reversal.
> >>
> >> How about "who cares, that's just the way it works?" Would that help?
> >
> > I specify the names explicitly in MODPARAMS.DAT and they are reversed in
> > AGEN$FEEDBACK.DAT, but only for page, not swap files. I wouldn't call
> > that "works". Perhaps there is a bug in the version of VMS I have, thus
> > my query here.
>
> Chainsaw out all that stuff and replace it with one file. Or a few
> same-sized files on separate (non-system-disk) spindles. Or more
> memory. Or a combination.
>
> And FWIW, it's quite possible to run OpenVMS with no page file.

Again, the swap and page files are there for historical reasons. I
probably don't need them now. I might get rid of them. I am
reorganizing my hardware now anyway. However, I like to understand
things.

> If you're interested in seeing what AUTOGEN is up to, the DCL is available.

Yes, I can try to find out where this is coming from. OK.

Let me phrase this another way: Does anyone else see hard-coded names
for page files reversed in AGEN$FEEDBACK.DAT? Does anyone else think
that this is the way it should work and is not some sort of bug? Has it
always been that way? Certainly the docs don't mention this "feature".

Retrocomputing? No. I even switched off my VAXen a while back. Do I
have the latest, greatest(?) stuff? No. But then again, I listen to
60s and 70s rock music and not Justin Bieber. And Bach instead of John
Cage. I don't mind being retro when I consider some of the alternatives
though, again, in computing that is not my goal. VMS is best for what I
use it for at home, and I enjoy working with it, though most of the time
it is a means to an end for other stuff. With reasonably fast (for my
needs) hardware available for free, why spend money on the latest stuff
from southeast Asia? What are the alternatives? Windows and worry
about viruses? Apple or Google and find files replaced with censored
versions which don't offend the Daughters of the American Revolution?
Gnu/Linux and use the stuff promoted by the guy who said it is a crime
against humanity (his words, really---check them out; some of this Gnu
philosophy would be funny if it weren't meant seriously) to sell
software without source code? Some other niche product which I would
have to spend time learning when VMS meets my needs? New skills? I
have a VMS day job, which should keep me going for a few years. Would I
be competitive in some other aspect of computing compared to all the
young dudes? No. Why waste time trying? I can hopefully run VMS at
home for as long as I need to. If it's no longer possible, and even if
the day job caves in, I don't have to work in computing. I've worked in
a slaughterhouse, as a lorry driver, as a pizza deliverer, in a garden
shop, on an assembly line, as a scientist, as a translator, as a
teacher. There are other things I would enjoy doing and be able to do
if I can't earn my living doing computing. Maybe at some point I can
retire and won't even have to earn my living. In the meantime, I enjoy
VMS and DEC hardware. This also has something to do with respect for
people who created a good product. When I switched off my VAXes, some
had been running continuously for 20 years, and were running the newest
VMS version; an executable from 20 years before would still run though.
Do that with the latest Windows/Intel box. Quality. Like in the VAX at
20 book where "quality is not a goal; it is a given". There is nothing
wrong remembering the good old days if they were actually better than
the present. Yes, other things have changed but there is no need to
avoid eclecticism; one can have the best of the past and the best of the
present. One doesn't have to bow down to trends. One can actually try
to understand things. I've been in comp.os.vms for 20 years or so and
have learned a lot and have actually helped a few people, be it
answering questions which I had learned the answers to from other people
and/or by doing or reminding people who had been unjustly flamed by Carl
Lydick that life is still worth living. I've met a few people from here
"offline" and they were nice folks. I like it here. :-)

MG

unread,
Dec 28, 2012, 5:51:57 PM12/28/12
to
On 28-dec-2012 22:34, Phillip Helbig---undress to reply wrote:
> Again, the swap and page files are there for historical reasons.
> I probably don't need them now. I might get rid of them. I am
> reorganizing my hardware now anyway. However, I like to understand
> things.

Forbidden!


> Retrocomputing? No. I even switched off my VAXen a while back. Do I
> have the latest, greatest(?) stuff? No. But then again, I listen to
> 60s and 70s rock music and not Justin Bieber. And Bach instead of John
> Cage. I don't mind being retro when I consider some of the alternatives
> though, again, in computing that is not my goal. VMS is best for what I
> use it for at home, and I enjoy working with it, though most of the time
> it is a means to an end for other stuff. With reasonably fast (for my
> needs) hardware available for free, why spend money on the latest stuff
> from southeast Asia? What are the alternatives? Windows and worry
> about viruses? Apple or Google and find files replaced with censored
> versions which don't offend the Daughters of the American Revolution?

Seriously, you don't need to apologize or motivate anything.

I myself also love classical music and I find the metaphor very
fitting! VMS is indeed like classical music to me, amongst computers
and in the context of computing, but in terrible (and quite possibly,
intentional) 'decline'.


> Gnu/Linux and use the stuff promoted by the guy who said it is a crime
> against humanity (his words, really---check them out; some of this Gnu
> philosophy would be funny if it weren't meant seriously) to sell
> software without source code? Some other niche product which I would
> have to spend time learning when VMS meets my needs? New skills? I
> have a VMS day job, which should keep me going for a few years. Would I
> be competitive in some other aspect of computing compared to all the
> young dudes? No. Why waste time trying? I can hopefully run VMS at
> home for as long as I need to. If it's no longer possible, and even if
> the day job caves in, I don't have to work in computing. I've worked in
> a slaughterhouse, as a lorry driver, as a pizza deliverer, in a garden
> shop, on an assembly line, as a scientist, as a translator, as a
> teacher.

I take my hat off to you sir, that's impressive. You sound like a very
stand-up person, who was never afraid to do any honest work.


> There are other things I would enjoy doing and be able to do if I can't
> earn my living doing computing. Maybe at some point I can retire and
> won't even have to earn my living. In the meantime, I enjoy VMS and
> DEC hardware. This also has something to do with respect for people
> who created a good product.

We, or I, sometimes/often forget that it's about cash flow... Ohne
Ehre, like everything in this fake "global economy".


> When I switched off my VAXes, some had been running continuously for 20
> years, and were running the newest VMS version; an executable from 20
> years before would still run though. Do that with the latest Windows/Intel
> box. Quality.

You know what I find interesting, now that you mention it, that all
these big corporations like HP claim that they're "green".

How many can say they ran their system, happily, for more than half
a decade? (Let alone productively.)


> Like in the VAX at 20 book where "quality is not a goal; it is a given".
> There is nothing wrong remembering the good old days if they were actually
> better than the present.

I'm sure it's slandered as being "20th century thinking" by the "a
China-made PC every two/three years" crowd, who're intensely concerned
about 'progress'.

What is also amazing, that PCs aren't exactly cheap either. Nothing
is really cheap, especially not for the volume and actual production
costs.

Do you remember what SATA disks used to cost a few years (in my own
estimation, about half a decade ago) costed and how they've seemingly
permanently spiked since the "Thailand floods"?

Prices, here anyway, still haven't recovered. Then there were also
the illegal LCD monitor price agreements!


> Yes, other things have changed but there is no need to avoid eclecticism;
> one can have the best of the past and the best of the present. One does
> not have to bow down to trends.

Heretic!!


> I've been in comp.os.vms for 20 years or so and have learned a lot and
> have actually helped a few people, be it answering questions which I had
> learned the answers to from other people and/or by doing or reminding
> people who had been unjustly flamed by Carl Lydick that life is still
> worth living. I've met a few people from here "offline" and they were
> nice folks. I like it here. :-)

Dear lord, do I want to know? Probably not. Parts of the historical
comp.os.vms archive should be merged with something along the lines of
*.psychiatry...


Happy new year to you! Let's hope people like yourself will remain
here in 2013.

- MG

John E. Malmberg

unread,
Dec 28, 2012, 8:16:29 PM12/28/12
to
On 12/28/2012 3:34 PM, Phillip Helbig---undress to reply wrote:

> Again, the swap and page files are there for historical reasons. I
> probably don't need them now. I might get rid of them. I am
> reorganizing my hardware now anyway. However, I like to understand
> things.

Quite simply, for virtual memory to useful your disk DMA throughput
needs to be able to move data to and from memory at about the same speed
that your processor can execute through memory.

If the DISK DMA can not keep up, then you need more memory to compensate.

IBM had some formulas in their VM/370 documentation set that could be
used to determine how much paging bandwidth you needed for a given CPU
speed. It had other recommendations on how to calculate how much memory
to keep on the free list and such. I do not have that manual any more
though, but it helped me a lot with VAX 11/750s, and MicroVAX machines.

Those formulas would generally predict what you could get away with for
a given hardware setup. I never saw anything equivalent in Digital
literature.

The tricks to get the disk speed up in the past were to use more
independent spindles on independent controllers. With more modern
storage, the different interconnects to storage as well has having CPUs
that operate in the GHz range has made it harder to get the memory
transfer speed up high enough.

So now we are basically back to buy more memory, as it is hard to get
the disk transfer speeds up to do any major paging.

Pagefiles were originally for small transfers and swapfiles where for
big transfers. Over time the way that VMS managed virtual memory
basically made the swapfiles less important for most users.

> Let me phrase this another way: Does anyone else see hard-coded names
> for page files reversed in AGEN$FEEDBACK.DAT? Does anyone else think
> that this is the way it should work and is not some sort of bug? Has it
> always been that way? Certainly the docs don't mention this "feature".

I learned long ago that while Autogen was your friend, it was a friend
that was wiling to betray you if given the slightest opportunity.

The first rule that I learned is never let Autogen adjust your
pagefiles. Use its advice, but never let it change them. Autogen at
its heart seems to know that you have only one pagefile and one
swapfile. It never seemed to understand that VMS wanted all pagefiles
to be the same size.

So based on my first rule, I would never have seen the condition that
you are seeing. I never looked in AGEN$FEEDBACK.DAT, I only looked at
the autogen report to make sure that it had values that I thought were same.

The second rule that I learned is to have Modparams allocate enough
resources so that they did not need to expand into their virtual region
under a normal production load. That expansion seems to slow things down.

The third rule that I learned only applied to VAX/VMS before about 5.4,
where Autogen would always calculate freelim and freegoal too small for
systems that had a Q-bus or smaller.

The fourth rule that I learned is never let feedback control what your
autogen values are set to. Use the MIN_ or MAX_ settings. You can use
the feedback as a guideline.

An upgrade of VAX/VMS usually results in at least one automatic autogen
being run with feedback that is only a few minutes old, which means that
any custom settings that you were relying on feedback were loss.

And at one site, a DDOS attack from a PING storm threw the feedback for
the pool expansion so high that it was useless.

> Retrocomputing? No. I even switched off my VAXen a while back.

I still have three, that I have not had the time / location to set them
up. One can be either a MicroVAX-I or Microvax-II depending on what
board set I install. I am getting there on getting the location set up,
but I will probably unload the smaller ones once I get them working.
For me SimH or other free emulator is the future for any VAX use. I
still think my VAX 4000-500 is faster than what I can get SimH or
another free emulator to run on affordable hardware. But even then, I
will probably leave it off much of the time because of the power
consumption.

-John
wb8...@qsl.network
Personal Opinion Only

Phillip Helbig---undress to reply

unread,
Dec 29, 2012, 3:54:55 AM12/29/12
to
In article <wYadnWlN4sRX2UPN...@mchsi.com>, "John E.
Malmberg" <wb8...@qsl.network> writes:

> The first rule that I learned is never let Autogen adjust your
> pagefiles. Use its advice, but never let it change them. Autogen at
> its heart seems to know that you have only one pagefile and one
> swapfile. It never seemed to understand that VMS wanted all pagefiles
> to be the same size.

Right, although the documentation claims it does. Also, even if it
things there is only one, it chooses differently for page and swap
files.

> So based on my first rule, I would never have seen the condition that
> you are seeing. I never looked in AGEN$FEEDBACK.DAT, I only looked at
> the autogen report to make sure that it had values that I thought were same.

I looked at AGEN$FEEDBACK.DAT only because I wanted to see where the
swapped (no pun intended) values were coming from.

David Froble

unread,
Dec 29, 2012, 5:50:38 AM12/29/12
to
My actions with respect to tuning are close to John's. Being a VMS
system manager is more than knowing how to run AUTOGEN. I had my own
favorite parameters for MODPARAMS in the early days, however, it appears
AUTOGEN has caught up a bit and the number of specified parameters in
MODPARAMS has decreased. (More physical memory in later hardware was a
big help.)

I always decide the size of the page and swap files. I might accept
some recommendations. But I retain the decision making.

I've always felt that the MicroVAX 3100 Model 98 was the best VAX ever
produced. Not the fastest, and a sucky SCSI interface, but cheap,
small, and if one didn't do the job, just add another.

Stephen Hoffman

unread,
Dec 29, 2012, 9:48:05 AM12/29/12
to
On 2012-12-29 10:50:38 +0000, David Froble said:

> My actions with respect to tuning are close to John's. Being a VMS
> system manager is more than knowing how to run AUTOGEN. I had my own
> favorite parameters for MODPARAMS in the early days, however, it
> appears AUTOGEN has caught up a bit and the number of specified
> parameters in MODPARAMS has decreased. (More physical memory in later
> hardware was a big help.)

Working with AUTOGEN has always been a rearguard action, as the AUTOGEN
tool has been improved, and as OpenVMS systems and requirements changed.

Stuffing three or five programmers into the five megabytes on a
VAX-11/750 or the eight megabytes of a typical VAX-11/780 was "fun".
These days, various smartphone have more RAM and more storage and more
graphics power and more cores than most VAX systems ever had. Times
change. Older Alpha boxes tune differently than newer Alpha boxes, and
tossing a gigabyte or two at an Alpha or an Integrity box means you
might not hit the page file for any but the largest of your tasks. And
no page file use means... fast. MODPARAMS looks sparse in comparison
to the old days, and the boxes run better.

Times change, constraints change and costs change.
<http://www.eecs.berkeley.edu/~rcs/research/interactive_latency.html>

> I always decide the size of the page and swap files. I might accept
> some recommendations. But I retain the decision making.

More physical memory usually wins.
Prices on that have cratered.
Fewer page files often wins.
Same-sized page files usually wins.
Disabling the AUTOGEN adjustments of page, swap and dump usually wins,
because that just doesn't work as well as manual override.

> I've always felt that the MicroVAX 3100 Model 98 was the best VAX ever
> produced. Not the fastest, and a sucky SCSI interface, but cheap,
> small, and if one didn't do the job, just add another.

Or the roughly equivalent Model 100 series boxes, or the VAXstation
4000 model 9x boxes...

Or the XP1000, DS10, DS15, VS10 or XP900 Alpha boxes...

Or my recent favorite, the zx2000.

Or the quad-core Xeon boxes or the quad-core Core i7 boxes or...

But then there's a decision around whether the goal is retrocomputing
for fun or profit — that's Alpha and older, these days — or whether the
newer gear is best suited for the task. These dinky Core i7 boxes run
rings around the VAX boxes and even the older Alpha boxes, and various
of the new boxes can be dead quiet, and they can get far more done with
far less fuss. Or the ARM tablet gear for portability; getting ten
hours of use on batteries is very hard to beat. The ease of some of
these new systems is... still surprising.

The complexity still lurks, but it's moved from tussling with one box
to tussling with a dozen, a hundred or thousands of boxes, and memory
and storage capacity has just completely exploded.

Old computers are like old cars. Many folks will have favorites, too.
Fun for some stuff, but not so good for others.

Phillip Helbig---undress to reply

unread,
Dec 29, 2012, 11:00:21 AM12/29/12
to
In article <kbmvr5$ltc$1...@dont-email.me>, Stephen Hoffman
<seao...@hoffmanlabs.invalid> writes:

> > I've always felt that the MicroVAX 3100 Model 98 was the best VAX ever
> > produced. Not the fastest, and a sucky SCSI interface, but cheap,
> > small, and if one didn't do the job, just add another.
>
> Or the roughly equivalent Model 100 series boxes, or the VAXstation
> 4000 model 9x boxes...

I like the VAXstation 4000 models. They are all nice, but the 90A is
quite fast (for a VAX).

> Or the XP1000, DS10, DS15, VS10 or XP900 Alpha boxes...

I have XP1000 and DS10. IIRC, XP900 is roughly the same as the DS10.
DS15 is DS10 on steroids. VS10? When did that come out?

The VAXstation 4000 are smaller, quieter, lighter and use less power,
though.

MG

unread,
Dec 29, 2012, 12:37:04 PM12/29/12
to
On 29-dec-2012 17:00, Phillip Helbig---undress to reply wrote:
> I have XP1000 and DS10. IIRC, XP900 is roughly the same as the DS10.
> DS15 is DS10 on steroids. VS10? When did that come out?

The VS10 was exclusively for VMS, I believe; my guess it it has
limitations imposed through SRM and no AlphaBIOS. The TS10/DS10CO
(for 'telco') and DS10L (1U) are the only truly, or physically,
differing ones, amongst the DS10 systems.

A bit outdated, in terms of not listing later upgrades, but
concise and informative nonetheless:
<http://din-tech.e717.net/product/ds10.htm>

- MG

Stephen Hoffman

unread,
Dec 29, 2012, 3:15:41 PM12/29/12
to
On 2012-12-29 17:37:04 +0000, MG said:

> On 29-dec-2012 17:00, Phillip Helbig---undress to reply wrote:
>> I have XP1000 and DS10. IIRC, XP900 is roughly the same as the DS10.
>> DS15 is DS10 on steroids. VS10? When did that come out?
>
> The VS10 was exclusively for VMS, I believe; my guess it it has
> limitations imposed through SRM and no AlphaBIOS.

OpenVMS sees no difference among what's variously been called
AlphaStation VS10, AlphaStation XP900, AlphaServer DS10 and
AlphaStation DS10 is the same box, modulo a graphics controller. The
rationale was a boring series of boring with some boring packaging and
boring details, with none of the boring being particularly relevent or
less than boring here. Treat these boxes as identical to an
AlphaServer DS10 for all discussions, and you'll be fine.

Paul Sture

unread,
Dec 30, 2012, 2:37:13 AM12/30/12
to
In article <kbktsi$jva$1...@dont-email.me>,
Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:

> On 2012-12-28 14:12:58 +0000, Paul Sture said:
>
> > In article <50dd9ee7$0$3162$e4fe...@dreader36.news.xs4all.nl>,
> > MG <marc...@SPAMxs4all.nl> wrote:
> >
> >> Besides, how productive or worthwhile is it to be using VMS anyway?
> >> Isn't gazing at the latest Samsung or Apple 'appliance' a much more
> >> sound and socially acceptable pass-time nowadays? We're obviously
> >> stuck in the past, obsessed with unnecessary functionality. Dare I
> >> call it, fundamentalism?
> >
> > Whatever floats your boat, and Phillip clearly enjoys VMS. I am taking
> > note of Hoff's warnings about career opportunities passing you by if you
> > don't move on.
>
> Ayup. HP BCS is prominently featuring their ProLiant DL980 box, as
> well as their upcoming DragonHawk SD blades and the HydraLynx blades,
> and pushing x86 servers with RAS features, and targetting Windows
> Server and Linux as their platforms for their future of
> mission-critical computing offerings, and has stated that
> mission-critical x86 is where they're spending most of their R&D and
> what they're full-speed working on.

Noted.

> > While Phillip was posting about free disk space I was researching to see
> > if I could grab more than one audio stream to disk simultaneously (with
> > mixed success).
>
> FWIW (and don't know what you're up to with those audio streams), one
> of the older Mac Mini boxes I deal with fairly regularly streams four
> parallel OTA HDTV streams from GbE to disk. And it's not very busy.
> That benchmark was also before the most recent upgrade, so that
> customer's box can undoubtedly go rather faster.

That's worth knowing thanks, because during that process I ran into a
disk I/O bottleneck - the audio stuff, several VMs, duplicity off to
another machine, and when Time Machine kicked in everything slowed to a
crawl. This is on the base 2011 model Mac mini. A quick test shows
that copying a file of several gigabytes from one place to another on
the system disk runs at somewhere between 14 and 17 MB/s, but when using
scp to another system I get 35 to 40+ MB/s. Offloading some files to a
file server via SMB, NFS or AFP will probably bring some benefits here,
(as would getting another disk).

On the audio front, my digital TV box will record TV but not radio, and
I wanted to record some programmes which clashed. I have used
SoundflowerBed on OS X to divert incoming audio to a recording app
before but had to look up the instructions because it's been a while
since I last used it. As you can see from this link there are about 10
steps even with the apps already installed.

<http://www.wikihow.com/Record-Application-Audio-With-Soundflower>

In stark contrast on my openSUSE box it was three steps:

1. sudo zypper install audio-recorder
2. fire up Firefox and point it at the desired radio station
3. run audio-recorder and start recording

To get back onto the thread topic, albeit with an OS X flavour, I have
found that OS X slows to a crawl when its swapfiles grow above a certain
size, much akin to the way VMS slows down when heavy swapping occurs.
With a G3 iBook that was anything above 1GB, with a G4 PowerBook it was
somewhat higher, and the Mac mini in its original 2GB configuration it
was about 4GB. The VMS adage of "add memory" works for Macs too.

FWIW although Apple and Kingston quote a maximum of 8 GB for the 2011
Mac mini, Crucial and various other sources recommend 16GB and that does
work.

--
Paul Sture

Paul Sture

unread,
Dec 30, 2012, 3:10:27 AM12/30/12
to
In article <50de2295$0$3166$e4fe...@dreader36.news.xs4all.nl>,
MG <marc...@SPAMxs4all.nl> wrote:

> Do you remember what SATA disks used to cost a few years (in my own
> estimation, about half a decade ago) costed and how they've seemingly
> permanently spiked since the "Thailand floods"?
>
> Prices, here anyway, still haven't recovered. Then there were also
> the illegal LCD monitor price agreements!

I bought a 2 TB disk shortly before the floods and a few weeks later the
price had more than doubled. The prices have come down, but are still
above the pre-flood prices.

For USB disks they have the excuse of USB 3 and I would guess that the
average consumer sees them as a performance upgrade so are happy with
that.

--
Paul Sture

Paul Sture

unread,
Dec 30, 2012, 6:07:33 AM12/30/12
to
In article <wYadnWlN4sRX2UPN...@mchsi.com>,
"John E. Malmberg" <wb8...@qsl.network> wrote:


> Pagefiles were originally for small transfers and swapfiles where for
> big transfers. Over time the way that VMS managed virtual memory
> basically made the swapfiles less important for most users.

I typically found that swapping was more expensive than paging so tried
to avoid swapping except for things like the print symbiont processes.

> > Let me phrase this another way: Does anyone else see hard-coded names
> > for page files reversed in AGEN$FEEDBACK.DAT? Does anyone else think
> > that this is the way it should work and is not some sort of bug? Has it
> > always been that way? Certainly the docs don't mention this "feature".
>
> I learned long ago that while Autogen was your friend, it was a friend
> that was wiling to betray you if given the slightest opportunity.
>
> The first rule that I learned is never let Autogen adjust your
> pagefiles. Use its advice, but never let it change them. Autogen at
> its heart seems to know that you have only one pagefile and one
> swapfile. It never seemed to understand that VMS wanted all pagefiles
> to be the same size.

I learned that rule the first time Autogen suggested smaller files and
went and created new versions.

> So based on my first rule, I would never have seen the condition that
> you are seeing. I never looked in AGEN$FEEDBACK.DAT, I only looked at
> the autogen report to make sure that it had values that I thought were same.
>
> The second rule that I learned is to have Modparams allocate enough
> resources so that they did not need to expand into their virtual region
> under a normal production load. That expansion seems to slow things down.

At one point I monitored pool usage for several weeks (we had a weekly
reboot to do standalone backups), and found that Autogen would try to
increase the values one week and if implemented would suggest decreasing
them the next week. There was a difference in the percentage change
calculated by VAX and Alpha - I forget which way around but one was 5%
and the other 10 or 15%. IIRC I nailed these down using MIN_ values.

> The third rule that I learned only applied to VAX/VMS before about 5.4,
> where Autogen would always calculate freelim and freegoal too small for
> systems that had a Q-bus or smaller.
>
> The fourth rule that I learned is never let feedback control what your
> autogen values are set to. Use the MIN_ or MAX_ settings. You can use
> the feedback as a guideline.
>
> An upgrade of VAX/VMS usually results in at least one automatic autogen
> being run with feedback that is only a few minutes old, which means that
> any custom settings that you were relying on feedback were loss.

This is very relevant to the pagefile size problem.

After any upgrade I always review MODPARAMS.DAT because of the way older
settings are brought forward. SCSSYSTEMID will be brought forward as
the raw number and I prefer to have the formula

SCSSYSTEMID = ((DECnet area number) * 1024) + (DECnet node number)

in there instead. Plus comments of course, particularly those relating
to changes required for layered software or recommended by release notes
or support recommendations.

Also waiting to trap you with the way parameters are brought forward
during an upgrade is that you will end up with duplicate entries in
MODPARAMS.DAT. It is way too easy to edit what you think is a unique
entry, but in fact you only hit the first occurrence of it - whoops the
old value contained in a later entry overrides that.

> And at one site, a DDOS attack from a PING storm threw the feedback for
> the pool expansion so high that it was useless.

Likewise running Autogen too soon after a reboot.

--
Paul Sture

Paul Sture

unread,
Dec 30, 2012, 6:10:39 AM12/30/12
to
In article <50ddac92$0$3517$e4fe...@dreader37.news.xs4all.nl>,
MG <marc...@SPAMxs4all.nl> wrote:

> > P.S. It's "pastime" not "pass-time".
>
> Mea culpa, I only ever heard it pronounced. English isn't exactly a
> phonetic language.

Indeed it isn't. French is arguably worse, and as someone else pointed
out German is easy compared to either English or French.

--
Paul Sture

Phillip Helbig---undress to reply

unread,
Dec 30, 2012, 6:30:49 AM12/30/12
to
In article <nospam-334AC9....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> Indeed it isn't. French is arguably worse, and as someone else pointed
> out German is easy compared to either English or French.

German is easy in one direction: when you know how something is spelled,
it is almost always 100% clear how it is pronounced. On the other hand,
knowing how something is pronounced is sometimes not enough since there
are sometimes alternate spellings. (In some cases, these might be
distinguished in pronunciation in one dialect but not in another.) For
example, ai, ay, ei, and ey are all pronounced the same (like the
English long i), so the names Reiner and Rainer, Kaiser and Kayser, and
Maier and Meier and Mayer and Meier are all pronounced the same (and, in
the last case, usually Meir and Mair and Meyr and Mayr as well).

French is even worse here, with more alternate spellings for a given
pronunciation, though the pronunciation can almost always be deduced
from the spelling.

English is difficult in both directions.

Some languages are very close to being easy in both directions:
Norwegian, Macedonian, Spanish, Italian (and others as well).
Macedonian even has one sound per letter and one letter per sound.

Bill Gunshannon

unread,
Dec 30, 2012, 9:42:44 AM12/30/12
to
In article <kbp8l9$iov$2...@online.de>,
One of the best examples of the ease of pronunciation for German is
Roger Whitaker. He sings a number of his songs in German as well as
English and yet, according to the usual sources, does not understand
or "speak" a word of German. But, given the lyrics he can pronounce
them in time to the music perfectly. :-)

Stephen Hoffman

unread,
Dec 30, 2012, 11:16:11 AM12/30/12
to
On 2012-12-30 07:37:13 +0000, Paul Sture said:

> In article <kbktsi$jva$1...@dont-email.me>,
> Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:
>>
>> FWIW (and don't know what you're up to with those audio streams), one
>> of the older Mac Mini boxes I deal with fairly regularly streams four
>> parallel OTA HDTV streams from GbE to disk. And it's not very busy.
>> That benchmark was also before the most recent upgrade, so that
>> customer's box can undoubtedly go rather faster.
>
> That's worth knowing thanks, because during that process I ran into a
> disk I/O bottleneck - the audio stuff, several VMs, duplicity off to
> another machine, and when Time Machine kicked in everything slowed to a
> crawl. This is on the base 2011 model Mac mini. A quick test shows
> that copying a file of several gigabytes from one place to another on
> the system disk runs at somewhere between 14 and 17 MB/s, but when using
> scp to another system I get 35 to 40+ MB/s. Offloading some files to a
> file server via SMB, NFS or AFP will probably bring some benefits here,
> (as would getting another disk).

The Mac Mini (Non-Unibody; 2009?) was running I/O to a FireWire 800
array configured JBOD.

FireWire 800 was (and is) a substantial improvement over USB 2.0.

WiFi on that box was getting ~200 to ~220 Mbps, per the displays and
not measured. That wasn't used for the tests.

There was no TM operating on that box. As happens with some VMS
configurations, you don't need to back up some disks and some system
disks; there's not much point in spending time processing backups
periodically for what can be restored from distro or from a single
backup.

The replacement box is running Thunderbolt to RAID, which is (no
benchmarks for it yet) substantially faster.

For comparison with VMS, Ultra-4 / SCSI 320 — which is comparatively
rarely found on mid-vintage and older VMS boxes — maxes out at ~2.5
Gbps, and Thunderbolt maxes out at ~10 Gbps. YMMV. If you can get the
disks going that fast, and which usually means cache. Which means
bringing the data into the host in bigger and bigger hunks — classic
RMS record-oriented processing stinks at this, too — and using memory
speeds where you can.

Again, times change, constraints change, costs change, which means
application designs and system designs change:
<http://www.eecs.berkeley.edu/~rcs/research/interactive_latency.html>
The VAX architecture was optimized for gonzo-expensive memory, and
various OpenVMS data structures have maintained that same model — for
compatibility, and to avoid changes. We're no longer in that world;
memory prices are far lower, and memory configurations far larger.

> ...To get back onto the thread topic, albeit with an OS X flavour, I have
> found that OS X slows to a crawl when its swapfiles grow above a certain
> size, much akin to the way VMS slows down when heavy swapping occurs.
> With a G3 iBook that was anything above 1GB, with a G4 PowerBook it was
> somewhat higher, and the Mac mini in its original 2GB configuration it
> was about 4GB. The VMS adage of "add memory" works for Macs too.

More physical memory, and — for when you do hit the swap disks — faster
swap disks, more managed memory loads mean better performance. The
same for OS X and VMS. The difference between the 5400 and 7200 RPM is
noticable with some of the loads on various boxes, too. Options
include disks with faster rotation, faster interconnects, faster
storage arrays, Fusion (hybrid) drives, and SSDs.

> FWIW although Apple and Kingston quote a maximum of 8 GB for the 2011
> Mac mini, Crucial and various other sources recommend 16GB and that does
> work.

Interesting. Useful. Most of the Mac boxes I deal with are near max
(supported) configurations. Many of the Alpha and Itanium boxes are
not. Watching a VMS box page heavily is... sad.

Paul Sture

unread,
Dec 30, 2012, 1:41:56 PM12/30/12
to
In article <akb274...@mid.individual.net>,
bill...@cs.uofs.edu (Bill Gunshannon) wrote:

> One of the best examples of the ease of pronunciation for German is
> Roger Whitaker. He sings a number of his songs in German as well as
> English and yet, according to the usual sources, does not understand
> or "speak" a word of German. But, given the lyrics he can pronounce
> them in time to the music perfectly. :-)

I never knew that he couldn't speak German, and his official website
confirms it.

<http://www.rogerwhittaker.com/bio.htm>

(I could have sworn I had seen or heard him being interviewed in German.
Obviously not.)

--
Paul Sture

Paul Sture

unread,
Dec 30, 2012, 2:00:54 PM12/30/12
to
In article <kbp8l9$iov$2...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

> In article <nospam-334AC9....@news.chingola.ch>, Paul Sture
> <nos...@sture.ch> writes:
>
> > Indeed it isn't. French is arguably worse, and as someone else pointed
> > out German is easy compared to either English or French.
>
> German is easy in one direction: when you know how something is spelled,
> it is almost always 100% clear how it is pronounced.

True although in Switzerland a lot of French words have crept into
everyday use. English too: a Spanish speaking girl in my German class
pronounced "homepage" as "homma pagga" when she first came across it (in
the context how was she to know that it was English?).

I find it pretty easy to find a German word in a dictionary when
pronounced clearly. Mostly anyway.

> On the other hand, knowing how something is pronounced is sometimes
> not enough since there are sometimes alternate spellings. (In some
> cases, these might be distinguished in pronunciation in one dialect
> but not in another.)

Switzerland is full of dialects and the formal written language is High
German. Swiss German writing is pretty much restricted to poetry,
songs, adverts and cartoons, and everyone has their own flavour of
spelling.

"Learn High German first" is the advice you receive when you first come
here, but that doesn't help with the Swiss German you will come across.
Fortunately most folks can switch to High German if you ask and that
gets you started.

> For example, ai, ay, ei, and ey are all pronounced the same (like the
> English long i), so the names Reiner and Rainer, Kaiser and Kayser, and
> Maier and Meier and Mayer and Meier are all pronounced the same (and, in
> the last case, usually Meir and Mair and Meyr and Mayr as well).

The Swiss telephone book is full of variations on Meyer and Maya is a
girl's name.

> French is even worse here, with more alternate spellings for a given
> pronunciation, though the pronunciation can almost always be deduced
> from the spelling.

Which is why it is normal when taking someone's name or address on the
phone in French to ask how it is spelled.

> English is difficult in both directions.
>
> Some languages are very close to being easy in both directions:
> Norwegian, Macedonian, Spanish, Italian (and others as well).
> Macedonian even has one sound per letter and one letter per sound.

IIRC Dutch (oops I mean Netherlands) was quite easy in that respect,
again with the occasional French word cropping up.

--
Paul Sture

Paul Sture

unread,
Dec 30, 2012, 2:57:05 PM12/30/12
to
In article <kbppca$bj6$1...@dont-email.me>,
Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:

> On 2012-12-30 07:37:13 +0000, Paul Sture said:
>
> > In article <kbktsi$jva$1...@dont-email.me>,
> > Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:
> >>
> >> FWIW (and don't know what you're up to with those audio streams), one
> >> of the older Mac Mini boxes I deal with fairly regularly streams four
> >> parallel OTA HDTV streams from GbE to disk. And it's not very busy.
> >> That benchmark was also before the most recent upgrade, so that
> >> customer's box can undoubtedly go rather faster.
> >
> > That's worth knowing thanks, because during that process I ran into a
> > disk I/O bottleneck - the audio stuff, several VMs, duplicity off to
> > another machine, and when Time Machine kicked in everything slowed to a
> > crawl. This is on the base 2011 model Mac mini. A quick test shows
> > that copying a file of several gigabytes from one place to another on
> > the system disk runs at somewhere between 14 and 17 MB/s, but when using
> > scp to another system I get 35 to 40+ MB/s. Offloading some files to a
> > file server via SMB, NFS or AFP will probably bring some benefits here,
> > (as would getting another disk).
>
> The Mac Mini (Non-Unibody; 2009?) was running I/O to a FireWire 800
> array configured JBOD.
>
> FireWire 800 was (and is) a substantial improvement over USB 2.0.

Unfortunately in my neck of the woods Firewire disks are rather
expensive. I have one Firewire 800 disk which I use for TM, and
periodically take copies onto other systems as well.

> WiFi on that box was getting ~200 to ~220 Mbps, per the displays and
> not measured. That wasn't used for the tests.
>
> There was no TM operating on that box. As happens with some VMS
> configurations, you don't need to back up some disks and some system
> disks; there's not much point in spending time processing backups
> periodically for what can be restored from distro or from a single
> backup.

Yep. I just realised the other day that we never backed up our cluster
Alpha PWS systems individually, since all the important stuff was on the
host systems.

> The replacement box is running Thunderbolt to RAID, which is (no
> benchmarks for it yet) substantially faster.
>
> For comparison with VMS, Ultra-4 / SCSI 320 — which is comparatively
> rarely found on mid-vintage and older VMS boxes — maxes out at ~2.5
> Gbps, and Thunderbolt maxes out at ~10 Gbps. YMMV. If you can get the
> disks going that fast, and which usually means cache. Which means
> bringing the data into the host in bigger and bigger hunks — classic
> RMS record-oriented processing stinks at this, too — and using memory
> speeds where you can.
>
> Again, times change, constraints change, costs change, which means
> application designs and system designs change:
> <http://www.eecs.berkeley.edu/~rcs/research/interactive_latency.html>

That's a useful table. Thanks.

>
> More physical memory, and — for when you do hit the swap disks — faster
> swap disks, more managed memory loads mean better performance. The
> same for OS X and VMS. The difference between the 5400 and 7200 RPM is
> noticable with some of the loads on various boxes, too. Options
> include disks with faster rotation, faster interconnects, faster
> storage arrays, Fusion (hybrid) drives, and SSDs.
>
> > FWIW although Apple and Kingston quote a maximum of 8 GB for the 2011
> > Mac mini, Crucial and various other sources recommend 16GB and that does
> > work.
>
> Interesting. Useful. Most of the Mac boxes I deal with are near max
> (supported) configurations.

Apparently Apple only ever quote the maximum amounts of RAM that a given
model was/is shipped with by Apple, and don't update those details when
greater capacity chips become available/affordable.

Thus far I have maxed out the RAM on all my Mac systems to date and this
is a rewarding approach. Unfortunately both the 2002 vintage iBook and
the later PowerBook had rather small maximum values here.

The general advice was/is to buy the basic Apple configuration and add
third party RAM on cost grounds. However the base models of the latest
iMacs don't provide you the ability to do RAM upgrades yourself (short
of the kind of surgery the iFixit folks indulge in), so you are better
off with a "Build To Order" (BTO) system with the amount of memory you
anticipate using over the lifetime of the machine. Or going for the
more expensive model where you can upgrade RAM yourself, of course.

> Many of the Alpha and Itanium boxes are not. Watching a VMS box page
> heavily is... sad.

After a reboot 24 hours ago my Mac mini is happily sitting at zero bytes
swap usage and zero page outs.

--
Paul Sture

Johnny Billquist

unread,
Dec 30, 2012, 4:50:45 PM12/30/12
to
Just want to point one thing out, in all this bickering.
Run an OS X box for a month without rebooting, and it will be paging
heavily, and the results are much worse than an old VMS box.
Todays machines (well OSes) are designed for regular reboots.
The memory is used rather aggressively for disk caching, and eventually
you'll be out of memory, and your processes will be paging a lot, since
disk caching is regarded as a more useful usage of memory than for
processes.

I hate having to reboot machines twice a month. (Or course, mandatory
updates means you often have to reboot that often anyway...)

Johnny

Stephen Hoffman

unread,
Dec 30, 2012, 5:36:08 PM12/30/12
to
On 2012-12-30 21:50:45 +0000, Johnny Billquist said:

> On 2012-12-30 17:16, Stephen Hoffman wrote:
>> On 2012-12-30 07:37:13 +0000, Paul Sture said:
>>
>>> FWIW although Apple and Kingston quote a maximum of 8 GB for the 2011
>>> Mac mini, Crucial and various other sources recommend 16GB and that does
>>> work.
>>
>> Interesting. Useful. Most of the Mac boxes I deal with are near max
>> (supported) configurations. Many of the Alpha and Itanium boxes are
>> not. Watching a VMS box page heavily is... sad.
>
> Just want to point one thing out, in all this bickering.
> Run an OS X box for a month without rebooting, and it will be paging
> heavily, and the results are much worse than an old VMS box.
> Todays machines (well OSes) are designed for regular reboots.

For some cases, uptime is great. Uptime isn't always a panacea, and
it can sometimes be a siren. Everything's working... Or appears to
be...

> The memory is used rather aggressively for disk caching, and eventually
> you'll be out of memory, and your processes will be paging a lot, since
> disk caching is regarded as a more useful usage of memory than for
> processes.

Interesting. Was that with 10.8? Or 10.7? Or earlier?

> I hate having to reboot machines twice a month. (Or course, mandatory
> updates means you often have to reboot that often anyway...)

Boxes that are up for months or years are a production risk, even when
the operating system itself is working perfectly. If it's production
servers, then Chaos Monkey
<http://techblog.netflix.com/2012/07/chaos-monkey-released-into-wild.html>
and similar can help spot production errors, avoiding single-point
problems, and can help keep your servers online when Bad Things happen.
Better to force the reboots and restarts and software updates — on a
schedule, or automated, or entirely randomly — when staff is available
to deal with the fallout.

OpenVMS has some useful features here with cluster rolling upgrades,
and it's commonly why many applications increasingly use multiple
hosts, multiple DCs and related. And why tools such as Chaos Monkey
can be used to trigger failures.

When you're doing DT, it helps to be triggering your own disasters, and
recovering from them.

There was a senior software engineer an eon or two ago, and one that
was quite fond of poking off-line and power buttons in server rooms.
He's been posited as an important reason why various parts of the
software he was working on functions as well as it does, too. His
initials are mysteriously embedded in one key subsystem of that
software, too, though I've heard denials of that...

Johnny Billquist

unread,
Dec 30, 2012, 6:03:55 PM12/30/12
to
On 2012-12-30 23:36, Stephen Hoffman wrote:
> On 2012-12-30 21:50:45 +0000, Johnny Billquist said:
>
>> On 2012-12-30 17:16, Stephen Hoffman wrote:
>>> On 2012-12-30 07:37:13 +0000, Paul Sture said:
>>>
>>>> FWIW although Apple and Kingston quote a maximum of 8 GB for the 2011
>>>> Mac mini, Crucial and various other sources recommend 16GB and that
>>>> does
>>>> work.
>>>
>>> Interesting. Useful. Most of the Mac boxes I deal with are near max
>>> (supported) configurations. Many of the Alpha and Itanium boxes are
>>> not. Watching a VMS box page heavily is... sad.
>>
>> Just want to point one thing out, in all this bickering.
>> Run an OS X box for a month without rebooting, and it will be paging
>> heavily, and the results are much worse than an old VMS box.
>> Todays machines (well OSes) are designed for regular reboots.
>
> For some cases, uptime is great. Uptime isn't always a panacea, and it
> can sometimes be a siren. Everything's working... Or appears to be...

True. It's just that I like it.

>> The memory is used rather aggressively for disk caching, and
>> eventually you'll be out of memory, and your processes will be paging
>> a lot, since disk caching is regarded as a more useful usage of memory
>> than for processes.
>
> Interesting. Was that with 10.8? Or 10.7? Or earlier?

Running 10.8.2 now, but it's been this way for a few years actually. I
have the same issues in Linux, as well as NetBSD. All Unixes seems to
have taken the same path here...

>> I hate having to reboot machines twice a month. (Or course, mandatory
>> updates means you often have to reboot that often anyway...)
>
> Boxes that are up for months or years are a production risk, even when
> the operating system itself is working perfectly. If it's production
> servers, then Chaos Monkey
> <http://techblog.netflix.com/2012/07/chaos-monkey-released-into-wild.html>
> and similar can help spot production errors, avoiding single-point
> problems, and can help keep your servers online when Bad Things happen.
> Better to force the reboots and restarts and software updates — on a
> schedule, or automated, or entirely randomly — when staff is available
> to deal with the fallout.

While I agree in a way in the concept of having software services
designed to be resilient to failure of individual machines, and so on,
when it comes to my personal workstation it does not impress me.
I don't care that my neighbors machine is up and running if I need to
reboot.

And having things crash often as a solution to having insecure and badly
designed software do seem more like a "we can't fix it, so we'll embrace
it" philosophy.

> OpenVMS has some useful features here with cluster rolling upgrades, and
> it's commonly why many applications increasingly use multiple hosts,
> multiple DCs and related. And why tools such as Chaos Monkey can be
> used to trigger failures.
>
> When you're doing DT, it helps to be triggering your own disasters, and
> recovering from them.
>
> There was a senior software engineer an eon or two ago, and one that was
> quite fond of poking off-line and power buttons in server rooms. He's
> been posited as an important reason why various parts of the software he
> was working on functions as well as it does, too. His initials are
> mysteriously embedded in one key subsystem of that software, too, though
> I've heard denials of that...

:-)

Johnny

Bill Gunshannon

unread,
Dec 30, 2012, 7:26:09 PM12/30/12
to
In article <nospam-1A764C....@news.chingola.ch>,
All the years I listened to him I thought he could. I understood he had
moved to Germany when he was forced to leave his birthplace, British East
Africa, for whar I assume are obvious reasons. I was utterly amazed when
I learned he could not speak any German.

Bill Gunshannon

unread,
Dec 30, 2012, 7:45:09 PM12/30/12
to
In article <akc4d1...@mid.individual.net>,
bill...@cs.uofs.edu (Bill Gunshannon) writes:
> In article <nospam-1A764C....@news.chingola.ch>,
> Paul Sture <nos...@sture.ch> writes:
>> In article <akb274...@mid.individual.net>,
>> bill...@cs.uofs.edu (Bill Gunshannon) wrote:
>>
>>> One of the best examples of the ease of pronunciation for German is
>>> Roger Whitaker. He sings a number of his songs in German as well as
>>> English and yet, according to the usual sources, does not understand
>>> or "speak" a word of German. But, given the lyrics he can pronounce
>>> them in time to the music perfectly. :-)
>>
>> I never knew that he couldn't speak German, and his official website
>> confirms it.
>>
>> <http://www.rogerwhittaker.com/bio.htm>
>>
>> (I could have sworn I had seen or heard him being interviewed in German.
>> Obviously not.)
>
> All the years I listened to him I thought he could. I understood he had
> moved to Germany when he was forced to leave his birthplace, British East
> Africa, for whar I assume are obvious reasons. I was utterly amazed when
> I learned he could not speak any German.
>

A most interesting link to his bio, which is very different from what
I remember reading in Hör Zu or some such 40 some years ago.

Paul Sture

unread,
Dec 30, 2012, 9:23:41 PM12/30/12
to
In article <akc4d1...@mid.individual.net>,
bill...@cs.uofs.edu (Bill Gunshannon) wrote:

> In article <nospam-1A764C....@news.chingola.ch>,
> Paul Sture <nos...@sture.ch> writes:
> > In article <akb274...@mid.individual.net>,
> > bill...@cs.uofs.edu (Bill Gunshannon) wrote:
> >
> >> One of the best examples of the ease of pronunciation for German is
> >> Roger Whitaker. He sings a number of his songs in German as well as
> >> English and yet, according to the usual sources, does not understand
> >> or "speak" a word of German. But, given the lyrics he can pronounce
> >> them in time to the music perfectly. :-)
> >
> > I never knew that he couldn't speak German, and his official website
> > confirms it.
> >
> > <http://www.rogerwhittaker.com/bio.htm>
> >
> > (I could have sworn I had seen or heard him being interviewed in German.
> > Obviously not.)
>
> All the years I listened to him I thought he could. I understood he had
> moved to Germany when he was forced to leave his birthplace, British East
> Africa, for whar I assume are obvious reasons. I was utterly amazed when
> I learned he could not speak any German.

Yes, I had a college friend whose parents had to leave Kenya, losing
their farm and money there in the process.

According to the English Wiki on Whittaker, he did manage to speak
German in the end, so maybe I was right about the interview.

That's not to detract from his achievement of course, even though I
didn't particularly like his songs.

"In March 2006, Whittaker announced on his website that a 2007 Germany
tour would be his last, and that he will limit future performances to
"occasional concerts". Now more fluent in German, he was seen singing
and was interviewed in German on Danish TV in November 2008."

--
Paul Sture

Richard B. Gilbert

unread,
Dec 30, 2012, 11:08:15 PM12/30/12
to
On 12/28/2012 9:28 AM, MG wrote:
> On 28-dec-2012 15:12, Paul Sture wrote:
>> Whatever floats your boat, and Phillip clearly enjoys VMS. I am taking
>> note of Hoff's warnings about career opportunities passing you by if you
>> don't move on.
>
> Ever heard of sarcasm? Also, what career opportunities?! Now you're
> simply being facetious.
>
>
>> While Phillip was posting about free disk space I was researching to see
>> if I could grab more than one audio stream to disk simultaneously (with
>> mixed success). Oh, and testing deduplicated backups.
>
> Yes, okay...?
>
>
>> P.S. It's "pastime" not "pass-time".
>
> Mea culpa, I only ever heard it pronounced. English isn't exactly a
> phonetic language.
>

English was the result of Norman men at arms trying to make time
with English Barmaids! That is "Why can't the English learn to speak?"

And the Americans created their own dialect!

Phillip Helbig---undress to reply

unread,
Dec 31, 2012, 7:40:39 AM12/31/12
to
In article <nospam-1A764C....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> I never knew that he couldn't speak German, and his official website
> confirms it.
>
> <http://www.rogerwhittaker.com/bio.htm>
>
> (I could have sworn I had seen or heard him being interviewed in German.
> Obviously not.)

You might have seen him SPEAKING German, in the sense that a parrot can
speak German.

Phillip Helbig---undress to reply

unread,
Dec 31, 2012, 7:51:11 AM12/31/12
to
In article <nospam-EE3512....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> True although in Switzerland a lot of French words have crept into
> everyday use. English too: a Spanish speaking girl in my German class
> pronounced "homepage" as "homma pagga" when she first came across it (in
> the context how was she to know that it was English?).

The standard English expressions for a web document are "web page" and
"web site", with the implication that the latter contains several of the
former. In German, "Webseite" is often used, which translates as "web
page". However, since it sounds something like "web site", it is often
confused with that. Seite means "page" but this is derived from the
meaning "side". Thus, I've seen "web page" translated as "web side",
which makes me cringe. Can I become a beefsteak, please?

> > Some languages are very close to being easy in both directions:
> > Norwegian, Macedonian, Spanish, Italian (and others as well).
> > Macedonian even has one sound per letter and one letter per sound.
>
> IIRC Dutch (oops I mean Netherlands) was quite easy in that respect,
> again with the occasional French word cropping up.

Yes, German is also very regular, but one has to know the rules, which
are a bit strange. First, spelling follows pronunciation, rather than
vice versa, so changes in pronunciation which exist in other languages
are reflected in spelling. For example, in (dialectical; the standard
is "lese") German, "I read" is "ich les", which is pronounced something
like "leece". Plural would be "wir lesen", pronounced something like
"lezen". In Dutch, one has "ik lees" and "wij lezen". So, the fact
that s is pronounced like z between vowels is reflected in the spelling,
as well as the fact that the e is long in both cases. (To understand
this, for one-syllable words ending in a consonant, Dutch assumes they
are short unless the vowel is doubled, whereas German assumes they are
long unless the final consonant is doubled. This implies that there is
never a final double consonant in Dutch.)

Jan-Erik Soderholm

unread,
Dec 31, 2012, 9:21:45 AM12/31/12
to

Phillip Helbig---undress to reply wrote 2012-12-31 13:51:
> In article <nospam-EE3512....@news.chingola.ch>, Paul Sture
> <nos...@sture.ch> writes:
>
>> True although in Switzerland a lot of French words have crept into
>> everyday use. English too: a Spanish speaking girl in my German class
>> pronounced "homepage" as "homma pagga" when she first came across it (in
>> the context how was she to know that it was English?).
>
> The standard English expressions for a web document are "web page" and
> "web site", with the implication that the latter contains several of the
> former. In German, "Webseite" is often used, which translates as "web
> page". However, since it sounds something like "web site", it is often
> confused with that. Seite means "page" but this is derived from the
> meaning "side". Thus, I've seen "web page" translated as "web side",
> which makes me cringe. Can I become a beefsteak, please?

:-)

The swedish word "sida" means both "side" and "page". So we use the
same word for "besides of..." ("på sidan av...") as for "on page 5..."
("på sidan 5..."). :-)

Jan-Erik.

We say website (for the whole "site") and websida (for a single page).

Also common is "hemsida" ("home page"), usualy for a smaller site
with just a few "pages"...

Happy New Year

/Janne.



Paul Sture

unread,
Dec 31, 2012, 10:25:11 AM12/31/12
to
In article <kbs1nv$1f5$1...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

> In article <nospam-EE3512....@news.chingola.ch>, Paul Sture
> <nos...@sture.ch> writes:
>
> > True although in Switzerland a lot of French words have crept into
> > everyday use. English too: a Spanish speaking girl in my German class
> > pronounced "homepage" as "homma pagga" when she first came across it (in
> > the context how was she to know that it was English?).
>
> The standard English expressions for a web document are "web page" and
> "web site", with the implication that the latter contains several of the
> former.

"Home page" was in common use at one time. Whereas nowadays it means
the landing page for a site. From http://en.wikipedia.org/wiki/Home_page

"In the 1990s the term was also used to refer to a whole web site,
particularly a personal web site (perhaps because simple web sites often
consisted of just one web page)."

> In German, "Webseite" is often used, which translates as "web
> page". However, since it sounds something like "web site", it is often
> confused with that. Seite means "page" but this is derived from the
> meaning "side".

This can be quite confusing in my experience. It can be hard to tell if
someone is talking about one page or a whole web site.

> Thus, I've seen "web page" translated as "web side",
> which makes me cringe. Can I become a beefsteak, please?

As ever, loads of howlers crop up when translating...

> > > Some languages are very close to being easy in both directions:
> > > Norwegian, Macedonian, Spanish, Italian (and others as well).
> > > Macedonian even has one sound per letter and one letter per sound.
> >
> > IIRC Dutch (oops I mean Netherlands) was quite easy in that respect,
> > again with the occasional French word cropping up.
>
> Yes, German is also very regular, but one has to know the rules, which
> are a bit strange. First, spelling follows pronunciation, rather than
> vice versa, so changes in pronunciation which exist in other languages
> are reflected in spelling. For example, in (dialectical; the standard
> is "lese") German, "I read" is "ich les", which is pronounced something
> like "leece". Plural would be "wir lesen", pronounced something like
> "lezen". In Dutch, one has "ik lees" and "wij lezen". So, the fact
> that s is pronounced like z between vowels is reflected in the spelling,
> as well as the fact that the e is long in both cases. (To understand
> this, for one-syllable words ending in a consonant, Dutch assumes they
> are short unless the vowel is doubled, whereas German assumes they are
> long unless the final consonant is doubled. This implies that there is
> never a final double consonant in Dutch.)

Informative thanks. What used to do my head in was the way that the
Dutch where I lived dropped 'n' in pronunciation, particularly at the
end of words, which can make the understand of plurals versus singular
difficult. However that prepared me for Switzerland where they do the
same (but in the middle of words too; a couple of places I am due to
visit tomorrow are written as Wettingen and Mellingen, but the (English
phonetical) pronunciation is Wettiga and Melliga).

--
Paul Sture

Paul Sture

unread,
Dec 31, 2012, 10:43:49 AM12/31/12
to
In article <kbs147$h9$4...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:
ROFL. When I first got to Switzerland I had a pretty good German
accent, but sorely lacked a decent vocabulary. Many a time folks would
assume I was fluent and that was sadly not the case, so until I could
get some serious study in, I had to "lose" the accent.

Nowadays folks ask if I am from Holland, so something in my accent must
have stuck from the time I was there. Then again when I was in France,
folks would ask if I was from Belgium. Perhaps they couldn't believe
that an Englishman was capable of speaking in a non-English accent :-)

(and yes, I do cringe at some of the accents English native speakers
come out with in otherwise totally fluent German or French)

Happy New Year!
Es Gutes Neues!
En guete Rutsch!
Bon Année!

--
Paul Sture

Bill Gunshannon

unread,
Dec 31, 2012, 11:43:41 AM12/31/12
to
In article <kbs1nv$1f5$1...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply) writes:
> In article <nospam-EE3512....@news.chingola.ch>, Paul Sture
> <nos...@sture.ch> writes:
>
>> True although in Switzerland a lot of French words have crept into
>> everyday use. English too: a Spanish speaking girl in my German class
>> pronounced "homepage" as "homma pagga" when she first came across it (in
>> the context how was she to know that it was English?).
>
> The standard English expressions for a web document are "web page" and
> "web site", with the implication that the latter contains several of the
> former. In German, "Webseite" is often used, which translates as "web
> page". However, since it sounds something like "web site", it is often
> confused with that. Seite means "page" but this is derived from the
> meaning "side". Thus, I've seen "web page" translated as "web side",
> which makes me cringe. Can I become a beefsteak, please?
>
>> > Some languages are very close to being easy in both directions:
>> > Norwegian, Macedonian, Spanish, Italian (and others as well).
>> > Macedonian even has one sound per letter and one letter per sound.
>>
>> IIRC Dutch (oops I mean Netherlands) was quite easy in that respect,
>> again with the occasional French word cropping up.
>
> Yes, German is also very regular, but one has to know the rules, which
> are a bit strange.

I guess I'm just lucky. I never found the rules to be strange.

> First, spelling follows pronunciation, rather than
> vice versa,

Doesn't that conflict with the previously stated concept of Meyer, Meier
and Mayer all being pronounced the same way?

> so changes in pronunciation which exist in other languages
> are reflected in spelling. For example, in (dialectical; the standard
> is "lese") German, "I read" is "ich les", which is pronounced something
> like "leece". Plural would be "wir lesen", pronounced something like
> "lezen". In Dutch, one has "ik lees" and "wij lezen". So, the fact
> that s is pronounced like z between vowels is reflected in the spelling,
> as well as the fact that the e is long in both cases. (To understand
> this, for one-syllable words ending in a consonant, Dutch assumes they
> are short unless the vowel is doubled, whereas German assumes they are
> long unless the final consonant is doubled. This implies that there is
> never a final double consonant in Dutch.)

I actually got quite good at playing Scrabble in Dutch (well, I guess
it was actually Flemish as it was with my peers in the Belgian Army
when we were in the field in Germany.) I can't remember enough to
do that any more, but I can still read and understand it when I hear it.

Paul Sture

unread,
Dec 31, 2012, 11:43:31 AM12/31/12
to
In article <kbqh8r$fcr$1...@Iltempo.Update.UU.SE>,
Johnny Billquist <b...@softjar.se> wrote:

> On 2012-12-30 23:36, Stephen Hoffman wrote:
> > On 2012-12-30 21:50:45 +0000, Johnny Billquist said:
> >
>> >>
> >> Just want to point one thing out, in all this bickering.

Can we call it an exchange of ideas rather than bickering? :-)

> >> Run an OS X box for a month without rebooting, and it will be paging
> >> heavily, and the results are much worse than an old VMS box.
> >> Todays machines (well OSes) are designed for regular reboots.

FWIW I have been running OS X since 10.1 (Yuk!), but in later versions
Apple have changed the paging/swap stuff with each major version, and
generally speaking on the same hardware that has led to better
performance.

Until Leopard, logging off and on again would trim the swapfile to and
used memory reasonable levels, but I tried that again with 10.8.2 Server
and a lot of memory was still in use. Yes the algorithms changed again
with Lion/Mountain Lion, but adding the server component very likely
brought other things into play.

> > For some cases, uptime is great. Uptime isn't always a panacea, and it
> > can sometimes be a siren. Everything's working... Or appears to be...
>
> True. It's just that I like it.

But as I discovered in the mid-nineties, when running a mission critical
set of systems, regular reboots were a Good Thing. For example, if you
had changed your startup procedures / scripts, those changes were still
fresh in the minds of the folks who had done those.

> >> The memory is used rather aggressively for disk caching, and
> >> eventually you'll be out of memory, and your processes will be paging
> >> a lot, since disk caching is regarded as a more useful usage of memory
> >> than for processes.
> >
> > Interesting. Was that with 10.8? Or 10.7? Or earlier?
>
> Running 10.8.2 now, but it's been this way for a few years actually. I
> have the same issues in Linux, as well as NetBSD. All Unixes seems to
> have taken the same path here...

As said above I have only noticed this with 10.8 using the Server stuff.

> >> I hate having to reboot machines twice a month. (Or course, mandatory
> >> updates means you often have to reboot that often anyway...)
> >
> > Boxes that are up for months or years are a production risk, even when
> > the operating system itself is working perfectly. If it's production
> > servers, then Chaos Monkey
> > <http://techblog.netflix.com/2012/07/chaos-monkey-released-into-wild.html>
> > and similar can help spot production errors, avoiding single-point
> > problems, and can help keep your servers online when Bad Things happen.
> > Better to force the reboots and restarts and software updates — on a
> > schedule, or automated, or entirely randomly — when staff is available
> > to deal with the fallout.
>
> While I agree in a way in the concept of having software services
> designed to be resilient to failure of individual machines, and so on,
> when it comes to my personal workstation it does not impress me.
> I don't care that my neighbors machine is up and running if I need to
> reboot.

If I cast my mind back to IBM times, one thing I was supposed to do was
write up recovery procedures in a way that humble operators could
understand.

Then I was supposed to switch off a disk drive and sit on top of it so
that the operators could not change the disk in that drive, and see if
my instructions worked...

> >
> > There was a senior software engineer an eon or two ago, and one that was
> > quite fond of poking off-line and power buttons in server rooms. He's
> > been posited as an important reason why various parts of the software he
> > was working on functions as well as it does, too. His initials are
> > mysteriously embedded in one key subsystem of that software, too, though
> > I've heard denials of that...
>
> :-)

A nice story, but it also reminds me of the data entry guy who would
make mistakes that nobody in the computer department would think of
doing. We should have sold his "expertise" as an application tester :-)

--
Paul Sture

Phillip Helbig---undress to reply

unread,
Dec 31, 2012, 12:12:23 PM12/31/12
to
In article <kbs71p$acn$1...@news.albasani.net>, Jan-Erik Soderholm
<jan-erik....@telia.com> writes:

> The swedish word "sida" means both "side" and "page". So we use the
> same word for "besides of..." ("på sidan av...") as for "on page 5..."
> ("på sidan 5..."). :-)

There is another German word for page, Blatt, which is used more for
just loose pieces of paper, as opposed to numbered pages in a book. It
is also common in newspaper titles, e.g. Hamburger Abendblatt. Blatt
also means blade (with which it is cognate), as in a saw. (In the case
of a knife, though, one would say "Klinge".) It also means leaf as in
leaf from a tree.

The Dutch blad is for both pages in a book and leaves from a tree, but
there are two plurals---bladen and bladeren---depending on which is
meant. (German also has this occasionally. "Wort" is "word"; the
plural "Wörter" means literally many words, but "Worte" is more
metaphorical, e.g. "große Taten sollten großen Worten folgen" ("great
deeds should follow great words").)

> We say website (for the whole "site") and websida (for a single page).

The first English, the second Swedish.

Phillip Helbig---undress to reply

unread,
Dec 31, 2012, 12:38:49 PM12/31/12
to
In article <nospam-115F5D....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> "Home page" was in common use at one time. Whereas nowadays it means
> the landing page for a site. From http://en.wikipedia.org/wiki/Home_page

I think of it as the main page for a site, such as one's personal home
page. Some people use(d) the term for a list of bookmarks. :-(

> > > IIRC Dutch (oops I mean Netherlands) was quite easy in that respect,
> > > again with the occasional French word cropping up.
> >
> > Yes, German is also very regular, but one has to know the rules, which
> > are a bit strange. First, spelling follows pronunciation, rather than
> > vice versa, so changes in pronunciation which exist in other languages
> > are reflected in spelling. For example, in (dialectical; the standard
> > is "lese") German, "I read" is "ich les", which is pronounced something
> > like "leece". Plural would be "wir lesen", pronounced something like
> > "lezen". In Dutch, one has "ik lees" and "wij lezen". So, the fact

Should be "Dutch is also very regular, of course".

> Informative thanks. What used to do my head in was the way that the
> Dutch where I lived dropped 'n' in pronunciation, particularly at the
> end of words, which can make the understand of plurals versus singular
> difficult.

This is common in some dialects of Dutch (and also in southwestern
dialects in German). Less common in German is to pronounce the "n" in
the "en" ending for infinitives, but leave out the "e".

> However that prepared me for Switzerland where they do the
> same (but in the middle of words too; a couple of places I am due to
> visit tomorrow are written as Wettingen and Mellingen, but the (English
> phonetical) pronunciation is Wettiga and Melliga).

Similar to southwest German dialect here.

Back to VMS. I have now replaced all my grey BA356 with blue ones.
During the first change, I shut down the node while I powered off the
box. For the second one, I forgot. However, apart from some expected
mount verification, there was no problem. How easy is it on other
systems to power down the system disk then power it back on? The third
was already blue, but I powered it down as well (no shutdown) so that I
could put it in one of the standup cabinets (all 3 now have standup
cabinets; the first always had, the second and third I set up today).
Now for a few full copies to replace the remaining grey disks.

Jan-Erik Soderholm

unread,
Dec 31, 2012, 1:27:27 PM12/31/12
to
Phillip Helbig---undress to reply wrote 2012-12-31 18:12:
> In article <kbs71p$acn$1...@news.albasani.net>, Jan-Erik Soderholm
> <jan-erik....@telia.com> writes:
>
>> The swedish word "sida" means both "side" and "page". So we use the
>> same word for "besides of..." ("på sidan av...") as for "on page 5..."
>> ("på sidan 5..."). :-)
>
> There is another German word for page, Blatt, which is used more for
> just loose pieces of paper, as opposed to numbered pages in a book.

Sure, "blad" in swedish. And one "blad" ("sheet"?) has two "sidor"
("pages" or "sides"? Not sure there...).

> It
> is also common in newspaper titles, e.g. Hamburger Abendblatt.

Svenska Dagbladet is a major newpaper in sweden.

> Blatt
> also means blade (with which it is cognate), as in a saw. (In the case
> of a knife, though, one would say "Klinge".) It also means leaf as in
> leaf from a tree.

Same here. :-)

Swedish has three major sources of foreign words, German, French and
English.

Phillip Helbig---undress to reply

unread,
Dec 31, 2012, 3:21:25 PM12/31/12
to
In article <kbslef$7sb$1...@news.albasani.net>, Jan-Erik Soderholm
<jan-erik....@telia.com> writes:

> > There is another German word for page, Blatt, which is used more for
> > just loose pieces of paper, as opposed to numbered pages in a book.
>
> Sure, "blad" in swedish. And one "blad" ("sheet"?) has two "sidor"
> ("pages" or "sides"? Not sure there...).

Definitely sides. A page is either a sheet of paper in a book, or one
numbered side of such a sheet, but not as part of a piece of paper.

> Swedish has three major sources of foreign words, German, French and
> English.

Probably, German mainly from the time of the Hanseatic League, French
from the time of Bernadotte and English since English became a common
language after the second world war.

Swedish words which look the same in German or English, however, might
reflect a common Germanic root ("hand" for example). Also, there are
words in English which originated with the Vikings (mainly Danes and
Norwegians---the Swedish Vikings mainly went east and then south, as far
as Constantinople---but all Scandinavian languages are close, and were
closer back then) and are different in other Germanic languages. "Sky"
is one example, which, however, means a type of cloud in Swedish. (I
can see a Viking and an Englishman pointing at the sky and discussing
terminology and the Englishman thinking he means the whole sky when the
Viking is pointing to clouds.) Another example is "take the knife and
cut the steak", which in modern Swedish is "taga kniven och kutta
steken" (really) whereas in modern German it is "nimm das Messer und
schneide den Braten". :-) (One can say "Steak" in modern German, but
this is a recently borrowed English word.)

Richard B. Gilbert

unread,
Dec 31, 2012, 5:09:57 PM12/31/12
to
On 12/30/2012 6:10 AM, Paul Sture wrote:
> --
> Paul Sture

Any language is easy if you are between the ages of two years and
about five or six years. We are programmed to learn one or more
languages in those years.

I was born during World War Two and travel to both Europe and the
"far east" was severely restricted! I was introduced to Spanish in
ninth grade. German was what I wanted but very few people care what you
want. "You must study a foreign language. Your options are French or
Spanish." I chose
Spanish.

The following year I attended a different school.
"No, you can't waste that year of Spanish!" "No, you may not
study both Spanish and German!" <sigh!!!!> My teachers
could not explain WHY Spanish needed both "Ser" and "Estar", both
of which translate to the English the verb "to be"!

When I finished U.S. Army Training, the powers saw fit to send me to
Japan. My second overseas assignment was in Turkey.

For a few years, I could order a beer in five languages:
English, Spanish, German, Japanese, and Turkish!

glen herrmannsfeldt

unread,
Jan 1, 2013, 4:38:35 AM1/1/13
to
Johnny Billquist <b...@softjar.se> wrote:

(snip)
> Just want to point one thing out, in all this bickering.
> Run an OS X box for a month without rebooting, and it will be paging
> heavily, and the results are much worse than an old VMS box.
> Todays machines (well OSes) are designed for regular reboots.
> The memory is used rather aggressively for disk caching, and eventually
> you'll be out of memory, and your processes will be paging a lot, since
> disk caching is regarded as a more useful usage of memory than for
> processes.

I am reading this on a PPC Mac Mini with the full 1GB, running 10.5.8,
which has now been up 127 days.

It isn't the OS that has paging problems, but the web browser. Every few
days (depending on use) I stop and restart Safari. Then use the
"open all windows from previous session" menu item to get back where
I was. Safari seems a little better than Firefox in memory usage.

Much less often, I close and reopen Adobe reader.

> I hate having to reboot machines twice a month. (Or course, mandatory
> updates means you often have to reboot that often anyway...)

Even my Win2K machine stays up months at a time.

-- glen

Bill Gunshannon

unread,
Jan 1, 2013, 1:41:43 PM1/1/13
to
In article <h_mdnQl39b7-kH_N...@giganews.com>,
"Richard B. Gilbert" <rgilb...@comcast.net> writes:
> On 12/30/2012 6:10 AM, Paul Sture wrote:
>> --
>> Paul Sture
>
> Any language is easy if you are between the ages of two years and
> about five or six years. We are programmed to learn one or more
> languages in those years.

Reality tends to prove otherwise.

>
> I was born during World War Two and travel to both Europe and the
> "far east" was severely restricted! I was introduced to Spanish in
> ninth grade. German was what I wanted but very few people care what you
> want. "You must study a foreign language. Your options are French or
> Spanish." I chose
> Spanish.
>
> The following year I attended a different school.
> "No, you can't waste that year of Spanish!" "No, you may not
> study both Spanish and German!" <sigh!!!!> My teachers
> could not explain WHY Spanish needed both "Ser" and "Estar", both
> of which translate to the English the verb "to be"!
>
> When I finished U.S. Army Training, the powers saw fit to send me to
> Japan. My second overseas assignment was in Turkey.
>
> For a few years, I could order a beer in five languages:
> English, Spanish, German, Japanese, and Turkish!

In High School, after my mandatory two years of Latin (which tells you
how old I am!) I was told I had no apptitude for languages so was not
allowed to study any. After High School I went in the Army. I qualified
every school but the Defense Language Institute as my DLAT score said,
you guessed it, no apptitude for languages. One year into my first Army
tour I was sent to Germany. I chose to not live in the barracks (and
EM Club) like most of my peers and spent most of my time among Germans.
After about 6 months I could converse freely in restaurants and clubs.
I spoke with enough of a dialect that on trips away from my base (I
went often to Garmisch for Winter Sports) people could tell wehere I
was from in Germany. By the time I left Germany, I could speak, read,
write and thoroughly understand German. (and a little Italian, French
and Russian) Came home, went back to Germany (again with the Army) 4
years later. This time I was not stationed with Americans so I was
even more immersed in Germany. Developed my German considerably. Came
back 3 years later. 15 years later I started taking German in college.
Instead of the usual "What I did this summer" essays I was writting
philosophical papers on subjects like Rassismus and "Die Mauer im Kopf".
Today I am fluent in German, listen to German radio everyday in my home
and all this from someone with no apptitude for languages. Oh yeah, while
no where near fluent I get by in French and can still understand some
Italian, Russian and Spanish. I tried Arabic, but there is a real killer!!

Paul Sture

unread,
Jan 1, 2013, 2:25:57 PM1/1/13
to
In article <kbuaqr$h4p$1...@speranza.aioe.org>,
glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:

> Johnny Billquist <b...@softjar.se> wrote:
>
> (snip)
> > Just want to point one thing out, in all this bickering.
> > Run an OS X box for a month without rebooting, and it will be paging
> > heavily, and the results are much worse than an old VMS box.
> > Todays machines (well OSes) are designed for regular reboots.
> > The memory is used rather aggressively for disk caching, and eventually
> > you'll be out of memory, and your processes will be paging a lot, since
> > disk caching is regarded as a more useful usage of memory than for
> > processes.
>
> I am reading this on a PPC Mac Mini with the full 1GB, running 10.5.8,
> which has now been up 127 days.

I had similar uptimes on my PPC iBook and PowerBook. As I have
previously mentioned a logout/login sequence would bring them back to
normal speeds if the swapfiles got too large and things got bogged down
(sometimes just quitting everything and waiting would achieve the same
results, sometimes not).

> It isn't the OS that has paging problems, but the web browser. Every few
> days (depending on use) I stop and restart Safari. Then use the
> "open all windows from previous session" menu item to get back where
> I was. Safari seems a little better than Firefox in memory usage.

I use a Safari Extension called Sessions, which automatically gives me a
Save / Restore dialogue on startup. I can't find it in the Apple Safari
Extensions library, but there is another one which appears to offer
similar functionality.

With the original 2GB RAM supplied with my Intel Mac mini, I was having
sufficient performance problems with Safari that I switched to using
Firefox. Until a couple of versions ago FF would quite often throw a
wobbly on restoring open tabs at startup (and display an apology that it
couldn't do this). I haven't seen that for a while so assume that
problem has been addressed.

Although Mozilla no longer support the PPC platform, you might be
interested in these links:

Anewer PPC version of Firefox -

http://www.floodgap.com/software/tenfourfox/

and a newer PPC version of Thunderbird -

http://sourceforge.jp/projects/tenfourbird/


> Much less often, I close and reopen Adobe reader.

I don't have that on any of my systems.

>
> > I hate having to reboot machines twice a month. (Or course, mandatory
> > updates means you often have to reboot that often anyway...)
>
> Even my Win2K machine stays up months at a time.

I had long uptimes with NT4, but that was on hardware specifically put
together with that O/S in mind and it wasn't connected to the internet
except for short periods. I also did my surfing on VMS or Linux.

--
Paul Sture

Paul Sture

unread,
Jan 2, 2013, 11:38:39 AM1/2/13
to
In article <kbsij9$eih$1...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

> Back to VMS. I have now replaced all my grey BA356 with blue ones.
> During the first change, I shut down the node while I powered off the
> box. For the second one, I forgot. However, apart from some expected
> mount verification, there was no problem. How easy is it on other
> systems to power down the system disk then power it back on?

That's an interesting question. VMS is the only system I have come
across in recent years which can survive a system disk going off line.

When I was running a Mac laptop booted to an external disk and no UPS
for example, with a short power cut the laptop would survive on its
internal battery, but of course the external disk would go offline.

If I was doing any processing at the time the laptop would lock up. If I
wasn't doing anything I might get lucky, or not. I concluded that if
any system disk I/O was required (logging, paging, timestamps etc)
during the time the disk was offline I was looking at a locked system.

With a completely cold start for both disks and system I also found the
system could beat the disk in coming online, in which case a system
power cycle was required.

I put your question to a Solaris/Red Hat guy yesterday and that prompted
him to tell me the tale of a couple of clustered Solaris systems which
have precariously long uptimes. But that's for the other thread :-)

--
Paul Sture

Phillip Helbig---undress to reply

unread,
Jan 2, 2013, 1:32:09 PM1/2/13
to
In article <nospam-40BC21....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> > Back to VMS. I have now replaced all my grey BA356 with blue ones.
> > During the first change, I shut down the node while I powered off the
> > box. For the second one, I forgot. However, apart from some expected
> > mount verification, there was no problem. How easy is it on other
> > systems to power down the system disk then power it back on?
>
> That's an interesting question. VMS is the only system I have come
> across in recent years which can survive a system disk going off line.

Several years ago, while sorting cables at home, I accidentally pulled
the SCSI cable from the system-disk shadow set on a VAX. Apart from
some mount-verification messages, no effect. Literally the next day at
work I was re-arranging some cables under my desk and pulled out the
mouse from a PC. Had to reboot. No joke. Nuff said. :-)

glen herrmannsfeldt

unread,
Jan 2, 2013, 1:34:51 PM1/2/13
to
Paul Sture <nos...@sture.ch> wrote:

(snip)
> That's an interesting question. VMS is the only system I have come
> across in recent years which can survive a system disk going off line.

> When I was running a Mac laptop booted to an external disk and no UPS
> for example, with a short power cut the laptop would survive on its
> internal battery, but of course the external disk would go offline.

I presume you don't count diskless machines with the system (root
partition) NFS mounted.

With the traditional (and still recommended) hard mount, an NFS client
will wait forever for its server to come back.

(I once sold a machine that others still had mounted. Had to reboot
those clients, as the server was never coming back.)

-- glen

Bob Koehler

unread,
Jan 2, 2013, 4:48:12 PM1/2/13
to
In article <kc1uf9$2bi$1...@online.de>, hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply) writes:
>
> Several years ago, while sorting cables at home, I accidentally pulled
> the SCSI cable from the system-disk shadow set on a VAX. Apart from
> some mount-verification messages, no effect. Literally the next day at
> work I was re-arranging some cables under my desk and pulled out the
> mouse from a PC. Had to reboot. No joke. Nuff said. :-)

Years ago, had lots of VAXStation 3100 of various models. Unplugged
things on a regular basis, then just plugged them in (no external
drives on these).

Got a DECStation 3100 (MIPS). Unplugged the keyboard, then plugged it
in.

Had to force a reboot.

David Froble

unread,
Jan 2, 2013, 5:28:11 PM1/2/13
to
It's the attitude, or sometimes worse. Do you plan for possible errors
and implement methods to attempt to recover from them, or, do you say
"either it works right or it dies".

Some years ago I was working with a programmer at one of the clients. I
asked what conditions were set for some of the possible errors,
including a timeout for messaging over the internet. The response was,
"oh, you'll always get a response". As her boss rolled his eyes, I
unplugged the RJ45 plug and held it up to her face and replied "even now?".

Even after you've "seen it all", you'll still be surprised at some of
the attitudes out there ....

Paul Sture

unread,
Jan 3, 2013, 7:34:49 AM1/3/13
to
In article <dm13RI...@eisner.encompasserve.org>,
I had a test rig which was a pair of Alpha PWS systems clustered
together. Due to lack of rack space I shared the console, keayboard and
mouse between them, simply swapping the cables from one to the other on
the rare occasions it was required.

The first time I swapped the cables with a Sun guy observing he gasped
and said "You couldn't do that on a Sun".

--
Paul Sture

Paul Sture

unread,
Jan 3, 2013, 7:39:11 AM1/3/13
to
In article <kc1ukb$16h$2...@speranza.aioe.org>,
I was thinking of machines booting from local disks (and for the
purposes of that a Fibrechannel disk at a remote site can look like a
local disk), but thanks for the reminder that NFS can do this.

--
Paul Sture

Paul Sture

unread,
Jan 3, 2013, 9:05:31 AM1/3/13
to
In article <kc1uf9$2bi$1...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:
I once got away with knocking a whole bank of disks offline for a few
seconds. I was expecting at least something in the logs but there was
nothing.

I have had the mouse thing more recently with a Windows 7 system, when
plugging in a new mouse. The strange thing was that the new mouse
functioned quite adequately to navigate and perform the requested
restart...

--
Paul Sture

Simon Clubley

unread,
Jan 3, 2013, 9:19:40 AM1/3/13
to
On 2013-01-03, Paul Sture <nos...@sture.ch> wrote:
>
> I have had the mouse thing more recently with a Windows 7 system, when
> plugging in a new mouse. The strange thing was that the new mouse
> functioned quite adequately to navigate and perform the requested
> restart...
>

Was the mouse a PS/2 or USB mouse ?

If it's the latter, then plugging in a USB HID mouse while powered up
_should_ be a fully supported operation.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

Paul Sture

unread,
Jan 3, 2013, 9:46:10 AM1/3/13
to
In article <kbqfko$k21$1...@dont-email.me>,
Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:

> On 2012-12-30 21:50:45 +0000, Johnny Billquist said:
>
> > Run an OS X box for a month without rebooting, and it will be paging
> > heavily, and the results are much worse than an old VMS box.
> > Todays machines (well OSes) are designed for regular reboots.
>
> For some cases, uptime is great. Uptime isn't always a panacea, and
> it can sometimes be a siren. Everything's working... Or appears to
> be...
>
> > The memory is used rather aggressively for disk caching, and eventually
> > you'll be out of memory, and your processes will be paging a lot, since
> > disk caching is regarded as a more useful usage of memory than for
> > processes.
>
> Interesting. Was that with 10.8? Or 10.7? Or earlier?

My experience of 10.3 to 10.5 (PPC client versions) was that OS X would
allocate new swapfiles in chunks of increasing sizes when required,
something like this:

128 MB swapfile0
128 MB swapfile1
256 MB swapfile2
512 MB swapfile3

and so on, with the precise sizes varying as successive versions
introduced refinements to the algorithm (and usually produced better
performance on the same hardware).

The swapping/paging algorithm would actively try to reduce these when
memory demands allowed. In many cases simply by quitting all
applications and waiting a while, OS X would trim those swapfiles back
to just one or two. If some process happened to be occupying the larger
swapfiles, this wouldn't happen, but normally a logout/login would
restart such processes in a way which allowed aggressive trimming to
occur. If you had restarted the any of the network components they
could be sitting "up there" and the trimming might not happen of course.

As I understand it, memory reclamation got a rework in 10.7 with the aim
of keeping apps around after exit but I skipped that version so have no
first hand experience of its behaviour.

"Something is deeply broken in OS X memory management (Lion performance
problems part 3)"

<http://workstuff.tumblr.com/post/20464780085/something-is-deeply-broken-
in-os-x-memory-management>

and

"Mountain Lion update: possibly fixed."

<http://workstuff.tumblr.com/post/28556080639/mountain-lion-seems-to-have
-partially-addressed-the>

--
Paul Sture

Paul Sture

unread,
Jan 3, 2013, 10:15:24 AM1/3/13
to
In article <kbqfko$k21$1...@dont-email.me>,
Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:

> Boxes that are up for months or years are a production risk, even when
> the operating system itself is working perfectly. If it's production
> servers, then Chaos Monkey
> <http://techblog.netflix.com/2012/07/chaos-monkey-released-into-wild.html>
> and similar can help spot production errors, avoiding single-point
> problems, and can help keep your servers online when Bad Things happen.
> Better to force the reboots and restarts and software updates — on a
> schedule, or automated, or entirely randomly — when staff is available
> to deal with the fallout.

A friend who is a *nix system administrator has the following problem:

2 Solaris boxes, one with an uptime of 3 years, the other with an uptime
of 6 years. These are configured together using Veritas Cluster Server
(VCS) such that for limited time periods one box should be able to cope
with the workload of both.

One of the mirrored system disks has failed on one of those boxes.
New disk ordered [1], but when it arrived they couldn't get the old disk
out.

Somewhere in this mix an extra file system was added to System A, but it
won't mount on System B (reason unknown).

A project to replace these systems was initiated about 4 years ago, but
got dropped in favour of something else.

My friend is now in the position of ensuring that the Unix team in
charge of these systems don't perform a reboot, in case either of them
or at worst case, both, don't come back. Oracle[2] have been called in
to have a look but they want to take the systems down to have a look at
them, and the customer is afraid that the boxes won't come up again. A
quick look at Wiki suggests that the version of Solaris mentioned
dropped off support last March.

I detected another factor at play here. The guy who set these systems
up has since left for pastures new, but there appears to be a refusal to
believe he could have missed something, simply because he had such a
good track record of getting things right. This reaction came when I
suggested checking firmware levels - it's surprising what can come
crawling out of the woodwork...

[1] Or as I understood it, "New 10 year old disk ordered".
[2] I still can't quite get used to thinking Oracle as a hardware company

--
Paul Sture

Stephen Hoffman

unread,
Jan 3, 2013, 10:47:59 AM1/3/13
to
On 2013-01-03 15:15:24 +0000, Paul Sture said:

> ...it's surprising what can come crawling out of the woodwork...

If you don't test or if you miss a test case, certainly.

Apple just got bagged by a corner case in iOS date handling with their
Do Not Disturb function, and the details of that are still just
filtering out.

VMS was getting bagged by date handling for years, and it's still far
too messy for my preferences. VMS date handling is still arguably
fundamentally broken, too. Unix got this one (more) right with UTC
everywhere, rather than localtime. But I digress.

There's reportedly a very evil PCSI bug floating around in an error
path, too — make sure your PCSI patches are current, or Very Bad Things
can happen.

I've chased critters in NEWUSER over the years. If you're not current
there, please upgrade.

Bugs happen.

That's why we test, after all.

Long server uptimes are the antithesis of testing.


--
Pure Personal Opinion | HoffmanLabs LLC

Paul Sture

unread,
Jan 3, 2013, 10:50:44 AM1/3/13
to
In article <kc441r$npf$1...@dont-email.me>,
Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote:

> On 2013-01-03, Paul Sture <nos...@sture.ch> wrote:
> >
> > I have had the mouse thing more recently with a Windows 7 system, when
> > plugging in a new mouse. The strange thing was that the new mouse
> > functioned quite adequately to navigate and perform the requested
> > restart...
> >
>
> Was the mouse a PS/2 or USB mouse ?

I am pretty sure I have seen it happen with both types.

> If it's the latter, then plugging in a USB HID mouse while powered up
> _should_ be a fully supported operation.

The last time it happened was when adding a USB mouse to a system which
already had a PS/2 mouse connected.

P.S. I have an "interesting" KVM set up here, to share a monitor. Being
an old KVM switch it only supports PS/2 keyboard and mouse. It gets its
power from the keyboard port on one PC, and won't work unless a real
mouse and keyboard are connected, but partly functional old ones do :-)

I use more modern and nicer USB keyboards and mice on the 2 systems.

--
Paul Sture

Paul Sture

unread,
Jan 3, 2013, 12:49:53 PM1/3/13
to
In article <kc497f$s4p$1...@dont-email.me>,
Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:

> On 2013-01-03 15:15:24 +0000, Paul Sture said:
>
> > ...it's surprising what can come crawling out of the woodwork...
>
> If you don't test or if you miss a test case, certainly.

One positive result of the conversation with my friend was that it
prompted him to dig out test plans which were created but not executed
due to other work taking priority...

> Apple just got bagged by a corner case in iOS date handling with their
> Do Not Disturb function, and the details of that are still just
> filtering out.
>
> VMS was getting bagged by date handling for years, and it's still far
> too messy for my preferences. VMS date handling is still arguably
> fundamentally broken, too. Unix got this one (more) right with UTC
> everywhere, rather than localtime. But I digress.

When I first learned that NTFS stores file dates as UTC I could
immediately see the disadvantages of VMS clusters across time zones.
However NTFS has its own problems:

<http://en.wikipedia.org/wiki/NTFS#Universal_time>

"As a result, especially shortly after one of the days on which local
zone time changes, users may observe that some files have timestamps
that are incorrect by one hour. Due to the differences in implementation
of DST in different jurisdictions, this can result in a potential
timestamp error of up to 4 hours in any given 12 months."

This has implications for anything using date based file comparisons
such as rsync and the likes of Dropbox. When I tested a Truecrypt disk
image with Dropbox I noticed that Truecrypt leaves the container image
date untouched when its contents are changed. Dropbox noticed this when
changes were made on more than one system and saved inconsistencies in
separate files, but you need to be aware of this.

> There's reportedly a very evil PCSI bug floating around in an error
> path, too — make sure your PCSI patches are current, or Very Bad Things
> can happen.
>
> I've chased critters in NEWUSER over the years. If you're not current
> there, please upgrade.
>
> Bugs happen.
>
> That's why we test, after all.
>
> Long server uptimes are the antithesis of testing.

Yes, yes and yes.

--
Paul Sture

Bill Gunshannon

unread,
Jan 3, 2013, 12:58:55 PM1/3/13
to
In article <nospam-E941F4....@news.chingola.ch>,
And yet, 6 months ago, long uptimes were the holy grail and one that
was claimed to be the domain of VMS. Go figure.

bill
(Who has never had a problem attaining long uptimes on a number of
disparate systems but never really saw the value therein as long as
my users were satisfied with systems availability. :-)

Phillip Helbig---undress to reply

unread,
Jan 3, 2013, 1:14:19 PM1/3/13
to
In article <nospam-1EA2F8....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> I had a test rig which was a pair of Alpha PWS systems clustered
> together. Due to lack of rack space I shared the console, keayboard and
> mouse between them, simply swapping the cables from one to the other on
> the rare occasions it was required.
>
> The first time I swapped the cables with a Sun guy observing he gasped
> and said "You couldn't do that on a Sun".

:-)

With ALPHAs no problem. With VAXes, at least some wanted to have the
console always switched on, at least during boot. (ALPHAs will even
boot with the console switched off or not even present.)

Phillip Helbig---undress to reply

unread,
Jan 3, 2013, 1:15:11 PM1/3/13
to
In article <nospam-83BCBC....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> I have had the mouse thing more recently with a Windows 7 system, when
> plugging in a new mouse. The strange thing was that the new mouse
> functioned quite adequately to navigate and perform the requested
> restart...

Better than the "keyboard not found...press any key to continue"
message.

glen herrmannsfeldt

unread,
Jan 3, 2013, 1:17:11 PM1/3/13
to
Simon Clubley <clubley@remove_me.eisner.decus.org-earth.ufp> wrote:

(snip)
> Was the mouse a PS/2 or USB mouse ?

> If it's the latter, then plugging in a USB HID mouse while powered up
> _should_ be a fully supported operation.

I never tried on Windows, but once had two mice and a trackball on
the same OS X machine at the same time. And yes, two added after the
system was running. Works fine. All three can move the cursor and
select things.

-- glen

Bob Koehler

unread,
Jan 3, 2013, 1:36:16 PM1/3/13
to
In article <kc4hpq$ao7$5...@online.de>, hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply) writes:
>
> With ALPHAs no problem. With VAXes, at least some wanted to have the
> console always switched on, at least during boot. (ALPHAs will even
> boot with the console switched off or not even present.)

Reminds me of the time our LA36 console died on one of our 11/780.
We had no spare LA36, and none of our VT100 had the current loop
interface that the consoles used.

We had two competely different setups with two completely different
sets of disk platters (RP06), so we had to be able to reboot.
Different applications, different data, different SYSGEN parameters,
differet UAF contents, sometimes different VMS versions, ...

I showed the operations folks how to tell that SHUTDOWN.COM had
probably finished, and how to use the front panel toggle switches
to force a reboot.

Ran just fine that way until the next day, when the parts came in.

Consoles? We don't need no stinkin' consoles!

glen herrmannsfeldt

unread,
Jan 3, 2013, 1:37:32 PM1/3/13
to
Paul Sture <nos...@sture.ch> wrote:

(snip)

> A friend who is a *nix system administrator has the following problem:

> 2 Solaris boxes, one with an uptime of 3 years, the other with an uptime
> of 6 years. These are configured together using Veritas Cluster Server
> (VCS) such that for limited time periods one box should be able to cope
> with the workload of both.

Solaris/SPARC or Solaris/x86?

> One of the mirrored system disks has failed on one of those boxes.
> New disk ordered [1], but when it arrived they couldn't get the old disk
> out.

> Somewhere in this mix an extra file system was added to System A, but it
> won't mount on System B (reason unknown).

> A project to replace these systems was initiated about 4 years ago, but
> got dropped in favour of something else.

> My friend is now in the position of ensuring that the Unix team in
> charge of these systems don't perform a reboot, in case either of them
> or at worst case, both, don't come back.

Favorite reason for SPARC systems not comming back is that the battery
in the TOY clock has failed. Ones I remember won't boot at all with it
down. For x86, I don't know.

> Oracle[2] have been called in
> to have a look but they want to take the systems down to have a look at
> them, and the customer is afraid that the boxes won't come up again. A
> quick look at Wiki suggests that the version of Solaris mentioned
> dropped off support last March.

I have before made copies of the file systems of running machines
(I don't remember SunOS or Solaris), fsck them (it is a copy of a
running system, so obviously not shut down) and then booted.

That is one that I would try here. (Assuming a spare system to
get running.)

Used SPARC boxes are pretty cheap, used x86 even cheaper.

-- glen

Bob Koehler

unread,
Jan 3, 2013, 1:38:53 PM1/3/13
to
In article <kc4hv7$d04$2...@speranza.aioe.org>, glen herrmannsfeldt <g...@ugcs.caltech.edu> writes:
>
> I never tried on Windows, but once had two mice and a trackball on
> the same OS X machine at the same time. And yes, two added after the
> system was running. Works fine. All three can move the cursor and
> select things.

"Somehow" my Mac decided it was OK to listen to the Magic Mouse down
the hall. Took me a couple minutes to work through System
Preferences and tell it not to, while the fellow down the hall was
doing work.

Stephen Hoffman

unread,
Jan 3, 2013, 2:12:33 PM1/3/13
to
On 2013-01-03 18:37:32 +0000, glen herrmannsfeldt said:

> Favorite reason for SPARC systems not comming back is that the battery
> in the TOY clock has failed. Ones I remember won't boot at all with it
> down. For x86, I don't know.

VAX and Alpha have that feature, too.
Alpha loads AlphaBIOS, which gets you nowhere with booting VMS.
VAX prompts for the time. Or boots with a wacky time on older versions, IIRC.
Haven't encountered a dead battery in an Itanium. Yet.

Phillip Helbig---undress to reply

unread,
Jan 3, 2013, 2:41:39 PM1/3/13
to
In article <kc4l71$gai$1...@dont-email.me>, Stephen Hoffman
<seao...@hoffmanlabs.invalid> writes:

> Alpha loads AlphaBIOS, which gets you nowhere with booting VMS.

That is the case on the PWS. Didn't older ALPHAs always boot properly,
even without a battery?

> VAX prompts for the time. Or boots with a wacky time on older versions, IIRC.
> Haven't encountered a dead battery in an Itanium. Yet.

Isn't there a timeout after which it stops asking for the time?

Stephen Hoffman

unread,
Jan 3, 2013, 3:05:03 PM1/3/13
to
On 2013-01-03 19:41:39 +0000, Phillip Helbig---undress to reply said:

> In article <kc4l71$gai$1...@dont-email.me>, Stephen Hoffman
> <seao...@hoffmanlabs.invalid> writes:
>
>> Alpha loads AlphaBIOS, which gets you nowhere with booting VMS.
>
> That is the case on the PWS. Didn't older ALPHAs always boot properly,
> even without a battery?

AFAIK, no. When the NVRAM loses its brains, it loses the console type
setting, and when it loses the console setting, the console program
finds a zero, and when you find a zero in the console type setting, you
get AlphaBIOS and when you get AlphaBIOS you get no OpenVMS bootstrap.
Which is not one of the better designs, IMO. It'd have been better to
have failed into a "help me" mode, and not into a a mode that's
indistingushable from a valid console set-up.

>> VAX prompts for the time. Or boots with a wacky time on older versions, IIRC.
>> Haven't encountered a dead battery in an Itanium. Yet.
>
> Isn't there a timeout after which it stops asking for the time?

If there's a bogus time or other related bogus setting, I'd hope the
console would stay stuck, and AFAIK it does.

There are settings to suppress the time prompts, but I've not seen
those applied here, nor would I suggest trying that, given getting
weird dates associated with stuff can cause — for instance — CMS to
lock out library access.

My general philosophical preference for unrecognized or weird errors is
to generate diagnostics where possible and exit quickly when something
goes wrong or goes weird, and let the human sort things out.
Recovery from weird or unknown errors is perilous at best. There are
more general cases — not involving the BB_WATCH or the TOY used for the
system time — where generating diagnostics is difficult or infeasible,
such as low-level and out-of-quota failures in the image activator.
But I digress.

With VAX, I don't know all of what would happen — those old VAX
consoles varied a whole lot more than did the Alpha and EFI consoles —
beyond the boostrap time prompting that would usually ensue when a
weird time was encountered. (VAX didn't store a useful time value in
the TOY clock, so it depended on the TOY value and a value retrieved
from the system image to generate the actual system time, which led to
decades of weirdness.)

If you have an old box around, pull the battery and try it.

John Wallace

unread,
Jan 3, 2013, 4:34:21 PM1/3/13
to
On Jan 3, 7:41 pm, hel...@astro.multiCLOTHESvax.de (Phillip Helbig---
undress to reply) wrote:
> In article <kc4l71$ga...@dont-email.me>, Stephen Hoffman
>
> <seaoh...@hoffmanlabs.invalid> writes:
> > Alpha loads AlphaBIOS, which gets you nowhere with booting VMS.
>
> That is the case on the PWS.  Didn't older ALPHAs always boot properly,
> even without a battery?
>
> > VAX prompts for the time.  Or boots with a wacky time on older versions, IIRC.
> > Haven't encountered a dead battery in an Itanium.  Yet.
>

As you rightly point out, not all Alphas had AlphaBIOS, Even on those
that don't (or on those with half-flash that had *either* AlphaBIOS or
SRM but never both at the same time), I'm not sure what happened in
the event of a flat battery - what about boot device info and such?


> Isn't there a timeout after which it stops asking for the time?

If the system does actually get as far as booting, then if the TOY
clock time appears invalid, VMS will prompt for a time and if one is
not entered it will wait for TIMEPROMPTWAIT (measured in
microfortnights) before continuing.

Paul Sture

unread,
Jan 3, 2013, 4:22:30 PM1/3/13
to
In article <kc497f$s4p$1...@dont-email.me>,
Stephen Hoffman <seao...@hoffmanlabs.invalid> wrote:

> Apple just got bagged by a corner case in iOS date handling with their
> Do Not Disturb function, and the details of that are still just
> filtering out.

Apologies if the following links have been posted here before, but IMHO
they are worth a revisit:

One of the comments I read when looking up the iOS date bug was a
pointer to "Falsehoods programmers believe about time":

<http://infiniteundo.com/post/25326999628/falsehoods-programmers-believe-
about-time>

and a follow up:
'More falsehoods programmers believe about time; łwisdom of the crowd˛
edition'

<http://infiniteundo.com/post/25509354022/more-falsehoods-programmers-bel
ieve-about-time-wisdom>

From the latter link:

71. You can determine the time zone from the state/province.
72. You can determine the time zone from the city/town.

I have variations on those:

* You can determine what human language I want your app to present from
the date and time format used on the client.
* You can determine what human language I want your app to present from
the date and time format used on the server.

I have given up with Windows apps here. The easiest way to get them to
default to English _and_ get something close to the date and time
formats I desire is to select UK English (and then modify the time zone
manually to reflect the one I am actually in).

Rinse and repeat for web sites offering downloads for software available
in different languages.

Rinse and repeat for web sites offering downloads which try to determine
my preferred GUI language from the IP address rather than the
information offered by the browser.

(Both posts at the above links were inspired by the post "Falsehoods
Programmers Believe About Names"
<http://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about
-names/> )

--
Paul Sture

David Froble

unread,
Jan 3, 2013, 5:42:08 PM1/3/13
to
Some do, some don't.

VAXstation 4000 systems will halt, sooner or later, if the line to the
console terminal is unplugged ..

Usually sooner ..

David Froble

unread,
Jan 3, 2013, 6:02:55 PM1/3/13
to
Paul Sture wrote:

> When I first learned that NTFS stores file dates as UTC I could
> immediately see the disadvantages of VMS clusters across time zones.

Ya know, nothing is ever perfect, and usually even worse at first.

When VMS was initially implemented, there were no clusters, and very
rudimentary networking. It's been a long time, but I'm not sure that
ethernet was working in 1978.

It's sort of hard to prepare for something that doesn't yet exist. No
USB in 1978 either ....

No Y2K issues in 1978 either. Not too many were worried about a 2 digit
year.

Yeah, it would be a better approach to keep all dates and time in GMT,
and have local offsets. Don't know how hard it would have been to
change VMS without breaking existing applications. Not to mention
applications that already took care of multiple time zones in the
application code.

Thus the delimma of older software. What to do when conditions change?
Can changes be incorporated without breaking things? Or do you just
say "bust everyone, this is the new way"?

Paul Sture

unread,
Jan 3, 2013, 6:36:40 PM1/3/13
to
In article <kc51bt$343$1...@dont-email.me>,
My very first experience of a VAXcluster was a pair of VAX 4000-300
systems.

One day the power went, the UPS then generator kicked in and I was still
happily working away.

The mains came back and the systems promptly gave no response.

The problem was that the VT220 lookalike consoles weren't connected to
the UPS and when they powered on they each sent a break...

As a sidenote, these were pretty crappy VT220 emulations, so had been
consigned to the computer room because nobody wanted them on their own
desks.

--
Paul Sture

Paul Sture

unread,
Jan 3, 2013, 7:02:59 PM1/3/13
to
In article <kc52iu$db5$1...@dont-email.me>,
David Froble <da...@tsoft-inc.com> wrote:

> Paul Sture wrote:
>
> > When I first learned that NTFS stores file dates as UTC I could
> > immediately see the disadvantages of VMS clusters across time zones.
>
> Ya know, nothing is ever perfect, and usually even worse at first.
>
> When VMS was initially implemented, there were no clusters, and very
> rudimentary networking. It's been a long time, but I'm not sure that
> ethernet was working in 1978.

Ethernet may have been on the drawing board in 1978, but according to
Wiki it was only commercially available in 1980 and standardised as IEEE
802.3 in 1985.

http://en.wikipedia.org/wiki/Ethernet

In that time frame I only used comms stuff over modems though, and even
as late as 1997 I didn't have a NIC on any home system - something that
allowed me to use a cheap backup solution incidentally, since the PC
backup software I was using assumed that if a NIC was present you were
running a server and had to pay a lot more.

> It's sort of hard to prepare for something that doesn't yet exist. No
> USB in 1978 either ....
>
> No Y2K issues in 1978 either. Not too many were worried about a 2 digit
> year.

I started my IT career in 1977 and it was kind of assumed that you would
need a rewrite to upgrade to the next box anyway. Nobody back then
seemed to think application solutions would last decades.

> Yeah, it would be a better approach to keep all dates and time in GMT,
> and have local offsets. Don't know how hard it would have been to
> change VMS without breaking existing applications. Not to mention
> applications that already took care of multiple time zones in the
> application code.
>
> Thus the delimma of older software. What to do when conditions change?
> Can changes be incorporated without breaking things? Or do you just
> say "bust everyone, this is the new way"?

I did do wholesales changes to code back in the 1970s and 1980s to
embrace new features, but I was one of the few who dared do that. It
helps if you work in a smallish team. Once you find yourself in a
corporate environment with many departments, that becomes so much more
difficult.

--
Paul Sture

Paul Sture

unread,
Jan 3, 2013, 7:46:50 PM1/3/13
to
In article <kc4mtj$f5h$1...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

> In article <kc4l71$gai$1...@dont-email.me>, Stephen Hoffman
> <seao...@hoffmanlabs.invalid> writes:
>
> > Alpha loads AlphaBIOS, which gets you nowhere with booting VMS.
>
> That is the case on the PWS. Didn't older ALPHAs always boot properly,
> even without a battery?

I don't recall any problems with non-PWS Alphas, but maybe I got lucky.

The problem with the PWS was that when changing from AlphaBIOS to SRM,
the option offered to save the "new" setting involved a power cycle,
which would of course revert to AlphaBIOS. I don't know if this worked
with a completely flat battery, but with a battery with a little life
left, if you hit the reset button instead, you could get away with it.

> > VAX prompts for the time. Or boots with a wacky time on older versions,
> > IIRC.
> > Haven't encountered a dead battery in an Itanium. Yet.
>
> Isn't there a timeout after which it stops asking for the time?

I don't recall one. VAX 11/750s were notorious for losing time so I
usually had SETTIME=1 so that VMS prompted for time at startup. AFAICR,
it just kept asking...

--
Paul Sture

Paul Sture

unread,
Jan 4, 2013, 5:59:49 AM1/4/13
to
In article <kc4hrf$ao7$6...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

That one is a gem, and at least you can have a good laugh at it.

--
Paul Sture

Paul Sture

unread,
Jan 4, 2013, 6:04:21 AM1/4/13
to
In article <kc4j5b$gqf$1...@speranza.aioe.org>,
glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:

> Paul Sture <nos...@sture.ch> wrote:
>
> (snip)
>
> > A friend who is a *nix system administrator has the following problem:
>
> > 2 Solaris boxes, one with an uptime of 3 years, the other with an uptime
> > of 6 years. These are configured together using Veritas Cluster Server
> > (VCS) such that for limited time periods one box should be able to cope
> > with the workload of both.
>
> Solaris/SPARC or Solaris/x86?

I am assuming Solaris/SPARC.

<snip>

> Favorite reason for SPARC systems not comming back is that the battery
> in the TOY clock has failed. Ones I remember won't boot at all with it
> down. For x86, I don't know.

<snip>

>
> I have before made copies of the file systems of running machines
> (I don't remember SunOS or Solaris), fsck them (it is a copy of a
> running system, so obviously not shut down) and then booted.
>
> That is one that I would try here. (Assuming a spare system to
> get running.)
>
> Used SPARC boxes are pretty cheap, used x86 even cheaper.

Many thanks. I'll pass the info on.

--
Paul Sture

glen herrmannsfeldt

unread,
Jan 4, 2013, 7:35:02 AM1/4/13
to
Paul Sture <nos...@sture.ch> wrote:

(snip)

>> Better than the "keyboard not found...press any key to continue"
>> message.

> That one is a gem, and at least you can have a good laugh at it.

If I remember that one from the MS-DOS days, I then plugged in the
keyboard, pressed F1 (I think that is what it wanted) and it worked!

Anyone remember:

C:> FORMAT C:
press any key when ready?

In the early versions, control-c and control-break counted as
any key.

Then again, there were add-on keycaps saying "ANY" that you could
put on your keyboard.

-- glen

Paul Sture

unread,
Jan 4, 2013, 8:03:06 AM1/4/13
to
In article <kc6i9m$s3p$2...@speranza.aioe.org>,
glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:

> Paul Sture <nos...@sture.ch> wrote:
>
> (snip)
>
> >> Better than the "keyboard not found...press any key to continue"
> >> message.
>
> > That one is a gem, and at least you can have a good laugh at it.
>
> If I remember that one from the MS-DOS days, I then plugged in the
> keyboard, pressed F1 (I think that is what it wanted) and it worked!
>
> Anyone remember:
>
> C:> FORMAT C:
> press any key when ready?
>
> In the early versions, control-c and control-break counted as
> any key.

Oh dear...

> Then again, there were add-on keycaps saying "ANY" that you could
> put on your keyboard.

When Compaq was still Compaq, they had a support page devoted to
explaining what the "any key" was.

--
Paul Sture

Phillip Helbig---undress to reply

unread,
Jan 4, 2013, 8:21:28 AM1/4/13
to
In article <kc6i9m$s3p$2...@speranza.aioe.org>, glen herrmannsfeldt
<g...@ugcs.caltech.edu> writes:

> Then again, there were add-on keycaps saying "ANY" that you could
> put on your keyboard.

German speakers and/or those familiar with Germany might be interested
to know that the "Alt" key can be replaced with a "K�lsch" key.

Bill Gunshannon

unread,
Jan 4, 2013, 8:43:32 AM1/4/13
to
In article <kc6l0o$6ah$1...@online.de>,
> to know that the "Alt" key can be replaced with a "Kölsch" key.
>

Where does the beer come out?

bill

Paul Sture

unread,
Jan 4, 2013, 8:42:55 AM1/4/13
to
In article <kc6l0o$6ah$1...@online.de>,
hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
wrote:

> to know that the "Alt" key can be replaced with a "Kölsch" key.

My dictionary gives me "beer from Cologne".

--
Paul Sture

Phillip Helbig---undress to reply

unread,
Jan 5, 2013, 6:26:10 AM1/5/13
to
In article <nospam-95EBE1....@news.chingola.ch>, Paul Sture
<nos...@sture.ch> writes:

> In article <kc6l0o$6ah$1...@online.de>,
> hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply)
> wrote:
>
> > to know that the "Alt" key can be replaced with a "Kölsch" key.
>
> My dictionary gives me "beer from Cologne".

Right. And "Alt" is beer from Düsseldorf. Düsseldorf and Cologne, like
Frankfurt and Offenbach, or Oxford and Cambridge, have a long-standing
friendly rivalry.

Bob Koehler

unread,
Jan 7, 2013, 9:49:56 AM1/7/13
to
In article <kc4mtj$f5h$1...@online.de>, hel...@astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply) writes:
> In article <kc4l71$gai$1...@dont-email.me>, Stephen Hoffman
> <seao...@hoffmanlabs.invalid> writes:
>
>> Alpha loads AlphaBIOS, which gets you nowhere with booting VMS.
>
> That is the case on the PWS. Didn't older ALPHAs always boot properly,
> even without a battery?
>
>> VAX prompts for the time. Or boots with a wacky time on older versions, IIRC.
>> Haven't encountered a dead battery in an Itanium. Yet.
>
> Isn't there a timeout after which it stops asking for the time?

There is a SYSGEN parameter which controls the behaviour. Default
setting is to prompt until answered.

0 new messages