It's been over a week since I've done anything with my brewing computer, and as if to make
the point, I ran out of stout. Got round to doing a little more work—I wonder why I'm
so reluctant. Put a connector on the Ethernet cable that I had installed over a year ago. It's still a pain to get the wires into the
connectors, but at least this time it worked first time. Also put the relay board in the
old UPS housing, and then went looking for the disk cage for the computer. Didn't find it,
so gave up instead.
A little more work in the garden, mainly weeding. This will keep me busy for a long time.
Also planted a few succulent cuttings in various places; they should fill out a lot of space
quickly.
More work on the brewing hardware today, at a snail's pace. Connected up the serial side of
things and gave up again. At this rate it'll take me another month before I start the next
brew.
More slow work on my brewing hardware. The hardware itself is now complete, but I need to
install the software. As the result of a number of problems, all my backups
of brewer were empty. Took the disk out of the old brewer and connected it in
the new machine. To my surprise, they're compatible; IDE (sorry, ATA (sorry, PATA)) has
been around for longer than I remembered. The old drives with separate control and data
cables have been gone for a really long time. Less to my surprise, the drive is dead:
Oct 3 14:32:23 brewer kernel: ad2: FAILURE - SETFEATURES SET TRANSFER MODE status=51<READY,DSC,ERROR> error=4<ABORTED>
Oct 3 14:32:23 brewer kernel: ad2: 405MB <WDC AC2420H 06.16K22> at ata1-master BIOSPIO
Oct 3 14:32:23 brewer kernel: ad2: FAILURE - READ status=59<READY,DSC,DRQ,ERROR> error=4<ABORTED> LBA=830759
Not a big issue—the only data I wanted was the configuration file—but I do
prefer to have the exact copy I started with. Still, no time for brewing today, so there's
always tomorrow...
One of the web communities I belong to is LinkedIn. As with many such communities, I don't really know why, and I'm certainly
not active. Today Edwin Groothuis told me that he had invited me to “connect with
him”. Just in time, it seems. When I went to the site, I found the invitation, along
with a (supposedly unrelated) suggestion on the right:
It's been 20 years since the events that led to the end of
the DDR and the reunification of
Germany—half the time that the DDR existed. How time flies! In Germany a whole
generation has grown up without any memory of the DDR.
Yvonne has allowed me to cook a cassoulet again, and spent most of the day preparing that.
In the process, discovered that
the Kransky sausages that Yvonne had
bought were in fact “Cheese Kransky”, but somebody had put a label over the
description. What good are they? The Wikipedia article says: “One
variation—the Cheese Kransky—is not considered to be an authentic member of the
Kransky family as it was created by Australian gourmet continental butchers”. I think
I can agree with that. Found some others in the deep freeze, but certainly it wasn't a
typical cassoulet. Still, the Yeardleys came over for dinner and consumed amazing
quantities, so it must have tasted OK.
More work on brewer today, and finally got it finished. Not that there was
much to do, of course, but it brought back to me how even the little details can take time.
The serial driver has changed since December 2000, and the device /dev/cuaa0 no
longer exists; instead I have /dev/cuad0 (and, for some strange reason on this
machine with a single serial port, /dev/cuad1). I suspect that they have always been
there, but they have different defaults. Took me a while to realise that all I really need
to do was to set the CLOCAL bit in the c_cflag.
Even then it didn't seem to do anything, and it took me a while to realise that I hadn't
told it where to display its status. A good reason to have kept the old config file. The
sample configuration file contains lots of comments, and is intended to be a good introduction to the configuration,
but it made it easy for me to miss things. I should consider restructuring it and
installing a stripped-down version as the base config file.
Summer's here! Or at least we put the clocks forward. Somehow it's more obvious this year,
and it's amusing to go into the laundry and see Lilac, our cat, sleeping in her basket, when
normally at this time she'd be scratching on one of the doors to get in.
More garden work. It was dry, and there was little wind—in the evening there was no
wind at all, something unusual. Finally did the planned weed spraying; the rest will have
to be by hand.
I've been getting this kind of message in the daily log message from my external web site
for some time:
Oct 5 00:41:11 w3 sshd[96071]: reverse mapping checking getaddrinfo for corporat201-24589042.sta.etb.net.co [201.245.89.42] failed - POSSIBLE BREAK-IN ATTEMPT!
Oct 5 00:43:50 w3 sshd[96084]: reverse mapping checking getaddrinfo for static-64-65-133-134.customer.sea.eschelon.com [64.65.133.134] failed - POSSIBLE BREAK-IN ATTEMPT!
Oct 5 00:50:37 w3 sshd[96171]: reverse mapping checking getaddrinfo for abts-north-static-171.184.160.122.airtelbroadband.in [122.160.184.171] failed - POSSIBLE BREAK-IN ATTEMPT!
Oct 5 00:51:42 w3 sshd[96178]: reverse mapping checking getaddrinfo for static.khi77.pie.net.pk [221.120.201.71] failed - POSSIBLE BREAK-IN ATTEMPT!
Oct 5 00:53:35 w3 sshd[96184]: reverse mapping checking getaddrinfo for 59.162.166.211.static.vsnl.net.in [59.162.166.211] failed - POSSIBLE BREAK-IN ATTEMPT!
Oct 5 00:56:37 w3 sshd[96213]: reverse mapping checking getaddrinfo for vp195031.kln.uac68.hknet.com [203.169.195.31] failed - POSSIBLE BREAK-IN ATTEMPT!
And yes, the IPs are genuine. If they're yours, don't complain: do something about it.
It's interesting to note that most of the IP addresses resolve to countries in South America
and Asia.
In principle, this should be harmless: I have my sshd set up to accept only public
key authentication—don't I? Checked the man page. Yes, defaults are public key only.
Still, it was worth checking. Set up a fake user name and password, connected
with ssh—and it accepted the password!
Further investigation revealed that the messages were only part of the
story. /var/log/messages contained further information. The following message
presumably relates to the first of the messages quoted above:
Oct 5 00:41:11 w3 sshd[96071]: error: PAM: authentication error for root from 201.245.89.42
It also proved that I hadn't read the man page very carefully:
yes, PasswordAuthentication defaults to no, but UsePAM defaults
to yes, and it effectively re-enables it. So the correct thing to do is:
# Change to no to disable PAM authentication
-#ChallengeResponseAuthentication yes
+ChallengeResponseAuthentication no
# Kerberos options
#KerberosAuthentication no
It's surprising that none of the documentation I read drew any attention to this seeming
compromise of security. Of course, it didn't do any harm; without exception, all of the
attacks were on root, and apart from the fact that root has a password
that is very difficult to guess, ssh to root is disabled anyway. Still, I
feel more comfortable now.
So finally I'm ready for my next brew. Almost. First I needed to crush the grain, and that
still takes far too long, over an hour. Started too late, and didn't get finished before
dinner.
What do I do now? I like to finish crushing the grain the day before. Another day's delay?
Or shall I do it first thing tomorrow? I'll decide tomorrow.
More pottering around in the garden. Weeds are the order of the day. One of
our Hibbertia scandens was so
overgrown that we could hardly find it. It should be climbing up a post, but it seems to
prefer to stay on the ground. We'll need to do some tying up.
Also did more arrangement on the verandah. Rearranged the hanging flower baskets again,
brought out the table that's been in the bedroom over winter, and had our first meal of the
spring outside. It's still a bit cool, but it's getting more pleasant.
Into the office this morning, turned on the monitors, and—nothing. Further
investigation showed that the system was still running, but the displays were blank.
Decided to reboot first and ask questions later. That in itself was an issue: I had set a
time bomb some days earlier by removing a USB card without umounting it. In versions
of FreeBSD before 8.0, this results in a panic
the next time the file system is accessed—by umount, for example, which
happened during the reboot. So I was in for another 2 hour fsck.
After reboot, found:
swap_pager_getswapspace(16): failed
swap_pager_getswapspace(16): failed
pid 37870 (Xorg), uid 0, was killed: out of swap space
pid 72545 (emacs), uid 1004: exited on signal 11 (core dumped)
pid 11462 (pbzip2), uid 0: exited on signal 6 (core dumped)
g_vfs_done():da0s1[READ(offset=16384, length=4096)]error = 6
deget(): pcbmap returned 6
g_vfs_done():da0s1[READ(offset=16384, length=4096)]error = 6
deget(): pcbmap returned 6
The last 4 lines were what the system managed while trying to umount the USB file
system before crashing. But the first lines show that I ran out of swap space. How did
that happen? I have 2 GB of memory and 3 GB of swap space—how can I run out there?
Still, that's the case, and I'll have to keep a more careful eye on it.
USB is really a pain. Later in the day I wanted to back up my photos, which I do to a
USB-mounted disk drive. Today when I plugged it in, the system froze. How? This time it
wasn't a panic. I suppose I should really give up on USB file systems on FreeBSD 7.x.
Tried it on kimchi, my NetBSD box. It
mounted! I had always thought that UFS file systems (especially UFS2) were not portable
between operating systems. Still, I don't have enough confidence that I'd actually write
to this file system with NetBSD: it's the only copy I have of my photos. Used
another FreeBSD system instead.
The computer problems gave me the excuse I needed to not brew any beer today. Finished
crushing the grain, and I should be ready for an early start tomorrow. As if to draw
attention to the urgency, I had two kegs to refill, and I only have one container of beer
left.
More garden work, tying up
the Hibbertia scandens that I
was working with yesterday. There's more light there now that we have demolished the
cathedral; hopefully they'll grow along
their guides now, and not along the ground.
Also more work in the Japanese Garden,
planting Euphorbias. The ones we
planted two weeks ago are not looking overly
happy, but it's pretty certain that they'll survive, so for the remainder it's the sooner,
the better:
Brew day today—I had run out of excuses to
postpone it, not to mention the even more serious danger of running out of beer.
Fortunately, things went well enough, apart from some minor problems with the temperature
probes. Starting early paid off: I was finished by mid-afternoon. Now to see whether I can
keep this up and brew another batch or two before the end of the month.
Somehow brew days are always a lot of stress, and I feel distinctly relieved when it's over.
Today was baking day (somehow that seems the wrong way round), and it was nothing in
comparison. I still needed to attend to the ambient temperature sensor, but basically
things are under way.
Baking isn't completely simple either. I've more or less settled on my recipe:
Mix 50 g starter, 100 g flour and about 100 ml water. Currently I'm using
“Manhattan light rye” mix, which is about 87% wheat, mainly because I have a
lot of it. At the moment I'm doing all the preparations at room temperature (round
20°).
When the fermentation stops, add 200 g flour and about 180 ml water. The time seems to
depend on the strain. I cultivated two separate starters from the dried starter that
Sue Blake sent me in April. They're
“interleaved”: each time I use the other strain, so I can compare them.
They seem to behave differently; this one (which I'm calling “strain 2” for
no particular reason) is slower, and it seems to create a more acid dough.
When the fermentation stops, I do it again: 200 g flour and about 180 ml water. I suppose
I should use less in the second step and more in the third, and I'll probably try that,
but this is what I'm doing at the moment.
Finally I save 50 g of the dough as the next-but-one starter and add 800 g of pure rye
flour and about 400 ml of water, 15 g of salt and 15 g of caraway seed. My bread pan is
most definitely not non-stick, so I cut baking paper to size (on our rolls, it's a
length of 32.5 cm for the bottom, with 5.5 cm cut off one end, and two lengths of 12x15
cm for the ends of the pan). I lay out the dough in the pan, smooth down and leave it
relatively wet, and leave to rise, which takes between 3 and 5 hours depending on the
strain.
Then I heat the oven to about 250°, spray the bread with water put it in the oven.
Today I tried leaving it on the grill and spraying frequently to get an effect similar
to the Hannover Gersterbrot. It took a lot of water, and somehow didn't have the same effect. That's not really
surprising; the real Gersterbrot is flamed before it goes into the oven.
After 15 minutes I drop the heat to 180° and bake for a total of 90 minutes, turning the
pan through 90° and spaying with water every 15 minutes. The turning is to compensate
for any irregularities in the oven temperature distribution.
More work in the garden, despite the cooler temperatures. I really need to find where my
hops are under the weeds (mainly grass) that have grown up over the winter. Rather to my
surprise, we've already filled up two bays of our three-bay compost heap, and I fear we'll
have the third one full before the first one has fully composted.
More trying out the ALDI barbecue in the evening, the third meal cooked with it. And the
ignition has failed already! It'll have to go back. I wonder if they will expect me to
dismantle it.
More fun with the wview weather station
software today, at least partially because I forgot to apply some of Steve Woodford's
patches. The result is that I have archive records with ridiculous values in them, and no
way to clean them out. Spent some time investigating that, discovering in the process that
the archive file format is not very conducive to such procedures. To be fair to the author,
he has since changed it, but it's now stored in a sqlite3 database (even if the rest uses, say, MySQL). I don't intend to follow that method, but I had
hoped for an easier way of sanitizing the data.
As it was, added a subdirectory cleanarchive to the utilities directory, and
spent most of the time trying to get these horrible GNU automake and friends to accept it.
Solution:
Create the directory, and copy another Makefile.am.
Add the name of the directory to the top-level configure.in. This was the part
that I didn't understand, and that caused so much cursing.
Sundance has now
left Waterloo University, and his next job is in Adelaide. Due to the
difference in the academic year, he now has until March to get there. He and Yana have decided to take the slow way home—by bicycle. They're
keeping a blog (back-to-front, of course), and I spent some time trying
to work out where they were. In the end I decided to create a Google Map to track their progress. The first
part of their journey will take them to Los Angeles (I think), a small matter of about 4000
km.
Apart from that, didn't do much. Did some work in the garden, messed around with
the fermenter—another overflow!—and did little else. It looks as if we'll have
plenty to keep us busy in the garden, though.
A horrible combination of bugs has been driving me crazy: the keyboard generates spurious “Ctrl locked
on” events, and the X server insists on interpreting
Ctrl-Alt-Backspace as “Terminate server”, even when
the xmodmap settings disable it. On the assumption that this is a keycode-dependent
setting, decided to remap the Backspace key to PrintScreen with the keyboard's
mapping tool, and then define that as Backspace in the xmodmap
settings—without “Terminate Server”, of course. And it worked!
Well, almost. Yes, the key still generates Backspace, at least in X—I still
need to check what happens in vty
mode—and Ctrl-Alt-Backspace no longer stops the server. But now
bash no longer interprets Alt-Backspace
as kill-word-backward. Emacs works correctly, though, and I suppose that's
the main thing.
House photo day today, and more playing around with fill-in flash. Got better results with
the Mecablitz 58 AF-1 O digital than last time, helped by the discovery that the head can rotate in a horizontal plane,
and that it's thus possible to point the sensor at the camera. I need to experiment
further.
Also some experiments with the polarizing filter, something that I haven't used much yet.
When it's overcast, the graduated grey filter works better, but on cloudless days like
today, the polarizing filter can create almost too-strong differences:
Looking out her office window, Yvonne saw an unusual sight:
a Koala with joey in a tree about 50 metres
away. Off in pursuit and got it a bit further down the paddock:
CJ along later with some posts for the Japanese Garden, but I wasn't happy with the
condition. They're not expensive new, and I think the difference is worth it. While he was
here, he told us of his recent journey to
the Birdsville races last month.
Seems lots of people are going on long journeys lately. This one went through the
Innamincka Reserve on roads that aren't even marked on Google Maps.
We're off to Anakie tomorrow to
visit the wildflower
show. Or are we? The venue is the Anakie-Staughton Vale Hall, and the directions are
very vague. Staughton Vale is some distance from Anakie, and I have no particular reason to
expect that the place will be well signposted. Google Maps was its usual helpful self:
Off early this morning to Anakie
to see the wildflower exhibition. Google Maps gave us a time of 80 minutes; I took a different way that they claimed
would take 87 minutes, as I discovered later; we left with 70 minutes to spare and were
there half an hour early—only 40 minutes, well less than half the time. That time
included finding the Hall, which was a couple of kilometres from the nearest indication on
the maps, and the first sign we saw was after the turnoff.
Waited around outside for a while while people tried to over-organize themselves; there were
tickets for everything, to get in ($1) and for the “workshops” (free). Finally
got in and heard the “workshop”, ostensibly about macro photography, which was
really a presentation about flower photography in general. Didn't learn too much, though he
did have the good idea of putting a box around plants for photography, both for wind
protection and lighting. I'll have to think of that.
Then a workshop about
identifying Acacias, where
we did learn something, and brought away a series of tests to identify the Acacias in
the Park. That's not that uninteresting; two of the three species that we know we have at
home are included,
including Acacia
myrtifolia, which was the subject of one of the examinations, and of which we needed
(and bought) another specimen as wind protection for the north bed. Also bought a CD and a
book.
Then off to look at the flowers in the park. We had already been along this road to get to
the Hall, and we hadn't noticed anything much:
Moved on to the Bert Boardman Recreation Area in Steiglitz, which everybody
pronounces Stieglitz. It seems there's a good reason: the town was founded by a Charles (if
I can believe the documentation) von Stieglitz, and presumably the name arose from the the
same kind of vague spelling that makes Google Maps claim that I live in “Kliens”
(and not Kleins) Road. The recreation area was not nearly as interesting as the Butchers
Road area, and I think the bus tours went there first. There were
some Xanthorrhoeas there,
presumably Xanthorrhoea
australis, and also
a Grevillea,
presumably Grevillea
chrysophaea, and not the
local Grevillea
steiglitziana:
On back home, stopping at a bridge to take some photos, and came across a lot of a ground
cover that we have at home, but which for some reason I didn't classify as a mystery plant. Here the ones on the side of the river,
followed by the ones in my garden:
Back home and processed our photos—Yvonne had taken
even more than I did. Then
decided to take some more photos in the garden and do some general garden work, including
planting one of the Xanthorrhoea quadrangulata (Mount Lofty Ranges Grass tree) that we grew from
seed years ago, and which seem to be going nowhere. They're certainly nothing like the size
of the ones in Steiglitz:
To the Yeardleys for dinner. David has a friend from his university days visiting him,
Peter, who lives somewhere in Ontario. It's funny how many people live so close together at
the other end of the world.
Spring is showing another attempt at coming. The overnight minimum temperature was only
about 13°, and it was very mild when we got up, so had breakfast on the verandah for the
first time this spring.
That kind of weather was conducive to gardening, of course, so spent more time weeding and
tying up roses. It's been less than two months since the
last rose of the last season bloomed (first photo), but already we have new ones:
The one in the second photo is a Climbing Iceberg, one of two that I needed to tie up to
climb up the side of the verandah. Also removed the Salvia that we bought at Diggers less than a year ago—for some reason it has died, not the only underperformer we
bought from Diggers.
In fact, the garden is booming.
The Osteospermum ecklonis
that seemed to be struggling in the winter have obviously won that struggle, and they're in
the process of taking over the north-east bed. Spent a few minutes removing a large part of
them, in the process revealing a number of bulbs that hadn't been able to find their way
out:
Somehow ended up doing nothing much today. The weather is cool, wet and windy again, but I
managed to do a certain amount of garden work—it's amazing how quickly grass takes
over if you let it.
My satellite connection has been relatively reliable lately: since the beginning of the
month I had “only” 5 outages and a total outage time of 398 seconds. Today that
changed completely. When I came into the office, the link had been down for about 4 hours,
and it continued all day long:
Date Outages Duration Availability
1255442400 20 16920 80.42% # 14 October 2009
Another day of rain, and once again didn't get much done. At least the rain suggests that
we might get some hay this year, though Yvonne has decided
not to cut it; we'll let the horses eat it directly from the stem.
Finally called up the ALDI service people
about the barbecue. It was clear that they're not used to this: I was transferred
to (silence), and after a while Jenny answered by accident, expecting to talk to
Rose. She couldn't help me either, and arranged for somebody to call back.
That somebody announced himself as Brett, but proved to be really
Tony—“Brett” is a handle, as he put it. I don't suppose it makes any
difference what name people use, though I admit that Brett sounds better than “Mockery”.
Tony didn't think that the problems I had were justification for return—in contrast to
ALDI's refund
policy, which obligingly showed exactly the barbecue on the page:
But it was clear that Tony worked for the service organization, not ALDI; he explained how I
could have paid $100 or $200 more buying it elsewhere, not exactly what I care much about.
Looks like I'll have to take it up with the local shop after all.
The tomatoes I planted months ago are now seedlings, but they look pretty weak; they can
barely stand up by themselves, and Yvonne tied them up to miniature stakes (really meat
skewers):
What's the problem there? Not enough fertilizer? Not enough light? Inappropriate pots?
Seeds not planted deep enough? In any case, time's getting on, and we have far more than we
can plant, so planted one each of Cherry tomato, Roma and Rouge de Marmande (which,
despite the name, seems to be an Australian variety) out in the veggie patch:
There's a good chance that the wind will break them; I think the area is sufficiently free
of frost that we won't have an issue there, though people in Victoria say you shouldn't
plant tomatoes until Melbourne Cup
Day.
Other plants in the veggie patch aren't looking too good either:
my Chinese Cabbage (which Wikipedia calls “Napa
cabbage”) is going to seed without ever having grown properly, and the Pok Choy (from
Diggers, who don't have a good track record
in our garden) isn't looking good either. On the other hand, more
conventional Brassicas, such
as Brussels sprouts
and Kohlrabi, seem to be doing OK:
I suspect they want far more water; my experiments show that they consist of about
95% water, so that makes sense. I think I'll postpone any further attempts until we erect
the greenhouse that David Yeardley has given us, and then we'll do it
with hydroponics.
Pulled out some more Osteospermums
from the north bed, and brought them over to the Yeardleys along with
some Gazanias, and took the opportunity
to look at the greenhouse. It was dismantled decades ago, and of course there's nothing
like instructions to go with it. Looks as if it'll be fun to put up. Confirmed, though,
that we can put it on the slab of the old pigsty, which will make things a lot easier.
The last two were taken with my Asahi Super Takumar 50 mm f/1.4 lens on bellows: the
ZUIKO DIGITAL
ED 50mm F2.0 Macro is almost certainly a better lens, but it gives up completely if it
doesn't have electronic communication with the camera. The lack of sharpness in the third
image is clear, and I think it's not just focus. Still, this photo is the first I've seen
that show the stepped shape of the sting.
It was new at the time, but 15 years later they have forgotten all about it, so the
link above is dead. It seems to be the Nikon D3S.
boasting ISO sensitivities of 12,800, 25,600, 51,200 and 102,400. What kind of numbers are
they? I've been with the ASA—no, ANSI—no, ISO scale for nearly 50 years now,
and they've always had sensible approximations to ensure that the numbers match the decimal
numbering system: we had 10 ASA, 100 ASA, 1000 ASA, even (arguably) 10,000 ASA. So why are
we suddenly going binary? It seems to be lack of understanding.
Spent some time discussing the matter on IRC, coming up with a surprising lack of
understanding of the underlying principles. Why did people come up with the current
numbers? They have nothing to do with the binary system; they're purely decimal. There are
ten steps between 1 and 10. But proportional increases aren't linear. If you took the
steps as 1, 2, 3 ... 9, 10, the first step would be much larger than the last, a factor of 2
compared to a factor of 1.111. Clearly equal steps have to be logarithmic: the difference
between each step is the tenth root of 10, 1.258925412. The following core of a little program shows the results:
for (x = 0; x < 1.1; x+= 0.1)
{
printf ("%6.1f\t%6.0f\n", x, y);
y *= tenthroot;
}
Clearly, this has nothing to do with binary. But by coincidence 3 steps is almost exactly 2
(not as exactly as shown here; it's actually 1.9952). This is the reason for the 1/3 EV
step in film sensitivities, which has now been adopted by digital cameras.
So what's wrong with departing from this method and using binary? The only way to look at
this in a binary perspective is by doubling; that causes three independent groups of
sensitivities, for example:
6400, 12800, 25600, 51200, 102400, 204800
8000, 16000, 32000, 64000, 128000
10000, 20000, 40000, 80000, 160000
Looking at the last members of each group, 102,400, 128,000, 160,000 and 204,800, the
relationships are no longer equal. They should all be 1.258925412, but in fact they're
1.25, 1.25 and 1.28.
Yes, that's not a serious problem—a more serious one will be when sensitivities
increase by another factor of 1000 (sorry, 1024): then the highest of those three groups
would be 209,715,200, 131,072,000 and 163,840,000. And that doesn't even make sense if you
stick to the decimal values 200,000,000, 125,000,000 and 160,000,000: the numbers are just
plain unwieldy. By this time it should be clear that the logarithmic method first pioneered
by Julius Scheiner makes
more sense: these last three values correspond to 84°, 82° and 83° DIN/ISO.
It's been raining a lot lately, but somehow found some time to do some work in the garden,
pruning and removing plants; we've decided to severely limit the number
of Osteospermum and remove all the
Carpobrotus
glaucescens in the main garden: they grow like fury, but so far they haven't flowered.
The weather's still pretty variable—we should really be happy that it's raining.
Managed to get some work done in the garden, though; the north bed has been through a lot of
change since we moved here, and the eastern side of the house will change a lot in the
coming months as well.
The trip to Anakie has me paying more attention to the
plants in the garden. Much of what we consider weeds could be some kind of interesting
wildflower. More macro photography. We two following plants that could conceivably be of
interest, though we will probably continue to consider them as weeds.
The first seems to be some kind of vine—I took some photos of it after returning from
Anakie, but not of the flowers:
The photos above were taken with the ZUIKO DIGITAL ED 50mm F2.0
Macro, pretty much at the limit of its focus. I can come much closer with the Asahi
Super Takumar 50 mm f/1.4 and the bellows, but I'm now relatively convinced that I'm hitting
some basic limitation of the lens. This time I turned it around with a reversing ring,
which should improve things, but it seems that there's a minimum circle of confusion, and it's
limiting the resolution. Here an extract from the previous image followed by a full-size
image with the Super Takumar:
The Super Takumar only barely manages to be sharper than the Zuiko. It's much more obvious
(I could hardly say “clearer”) with yet another attempt with nettles, this time
perpendicular to the leaf. The second image is a detail from the first which displays at
natural resolution as a “thumbnail”. It's clearly very unsharp, and it's not primarily due to focus.
More house photos stuff today. Every time I try to do something out of the ordinary with my
Mecablitz 58 AF-1 O digital, I run into problems. Today it was the photos of the
verandah. The light was subdued and just right for a photo with a lot of flash fill-in.
And I couldn't get the unit to fire out of the line of sight of the wide-angle lens (18 mm
equivalent on a 35 mm camera). Why can't these things have a proper cable connection? In
this case, it dawned on me that I could mount the flash on the camera in this particular
case, but that's not the point.
Ended up with a whole lot of photos, and my current method of naming them—partially
automatically from a list of photo names, and partially manually by entering
names—proved less successful. In the Good Old Days we made contact prints of entire
films and selected what we wanted. I suppose the GUI generation do something similar, but
they tie it down with so much mouse pushing and other stuff that I don't want to know. At
any rate it seems unusual that people actually give descriptive names to their photos when
using GUI tools, maybe because it requires typing.
Ended up writing a little PHP script that created a web page with really small images
(150x200), which took much less time than I had expected. Now if I only had the EXIF data
showing as well. MythWeb has nice
popups of the programme details, so went to investigate how they did that.
Javascript, of course, something that I don't know much about. Spent some time dissecting a
web page and removing the stuff I didn't need, and even got some kind of display, but
clearly I need different formatting. More work required.
While I was at it, also played around with my “slide
show” page showing the changes in the garden. It's all well and good, but it
only showed one image at a time. Added code to display a second image and change it
independently. That wasn't so easy, though it looks simple enough. The problem is
that HTML has such a really horrible concept of state: I have a central function to display
photos, and it handles the links. But every time I add some additional information (here
the variables pos2 and dir2 to manipulate the position and direction of
movement of the second image), I need to modify the central display function to pass them in
every link. That's not a question of the difference between GETDATA
and POSTDATA; it's just the fact that it needs to know at all. If HTML had
structure data types, I could just add a value to the structure and pass it without the
central function needing to know anything. As it is, as soon as I do something as simple as
trying to change the size of the second image, it goes away, because
the showphoto function doesn't know the variables.
Still, this page shows a basic weakness in my display functions: all the additional
information is there whether you want it or not, at least for the larger photos. Here's the
current state for a “small” image:
I seem to have forgotten to include it.
I probably need to provide the EXIF information as a Javascript popup and make a drop-down
menu for the other items.
My unfinished work with the “contact prints” yesterday gave me no rest, and
spent most of the day working on it and learning precious little about Javascript. Got the
output looking fine, but it appeared at the bottom of the screen, even if that part of the
page wasn't currently being displayed, something that MythWeb doesn't do. Looked at the
insides of the MythWeb scripts,
which proved to be hundreds of lines of stuff that looked much more complicated than what I
needed.
Finally asked on IRC and got a reply from Edwin Groothuis, who took the obvious step of
asking Google and came up with this script. Downloaded it and played around with it a lot, and got it to work, sort
of. Clearly I need to understand Javascript better.
Yvonne took a liking to the little orange flowers I
photographed a couple of days ago, and left the glass on the dining table. Today they
opened—they're about 5 mm in diameter—so I had the chance to take more macro
photos and investigate this problem with unsharpness. Was it related to the small aperture
(still only f/16, the smallest that the Super Takumar can do)? Took a number of photos with
three different lenses, the others being the ZUIKO DIGITAL ED 50mm F2.0
Macro and my 135 mm f/2.8 Exaktar.
The photo with the Zuiko looked nice enough, but I couldn't really come as close as I
wanted:
The Super Takumar came close enough, but I'm still not happy with the sharpness. It beats
the Zuiko (here first a detail of the previous image, then two full images taken with the
Super Takumar):
But it's still not really sharp. Did some comparison photos with different
apertures, and then with the Exaktar. The results suggest that the aperture isn't the
issue. Here details taken at f/16, f/8 and f/5.6 respectively:
About the only obvious difference is the depth of field, along with a little burn-out in the
last image. And the Exaktar did no better; the same size section of the closest I could get
shows the same kind of blurring:
For the time being I'm going to have to assume that it's a limitation of the lenses, which
almost certainly weren't designed to provide this sharpness.
Still more weeding. Sometimes I despair of getting this stuff fixed up. Yvonne is planning some red posts for her Japanese garden, and spent
some time preparing the rather dubious looking raw material.
More work on the Javascript stuff today, and finally got it working acceptably, though I
can't get the windows to stop moving with the cursor. Now the question is whether the
approach is worth the trouble, or whether the continual popups get on people's nerves:
More problems with the WH-1080 weather station today: something went wrong with the
communication between station and computer, and it took me a few hours to notice. During
this time I got a continual stream of:
Oct 19 13:52:09 kimchi wviewd[18269]: <1255920729129> : readStationData: preposterous rain rate! raw 440, delta 65532
...
Oct 19 16:43:01 kimchi wviewd[18269]: <1255930981749> : readStationData: preposterous rain rate! raw 440, delta 65532
Restarting wview got rid of it, so
it looks like there's room for improvement. In the meantime I have 1 spurious mm of rain
and flat values for all the other parameters, presumably because the entire data record was
discarded:
A couple of months ago I bought a lens and had it sent USPS First Class Mail.
Despite the contradictory tracking information, the lens arrived quickly, but even
after it arrived the tracking information did not
acknowledge that it had been posted. Was reminded of this incident today, and checked. The
tracking information had been written out to backup store, but was retrieved automatically
in a couple of hours, where I could discover that it still hasn't been sent:
One of the things I do in the morning is to check the TV programme and set recordings with
MythWeb. Not today: cvr2
was down, and it wouldn't come up. Further investigation showed that it was the same
problem that I had had three months ago: the system had
written the (XFS) root file system superblock
back offset by 64 bytes:
Last time this happened, I had assumed that this was a random hardware glitch. But it's the
only problem I've had with this computer, and it's happened twice now, with exactly the same
offset. I've got to assume that it's a software bug, and the clear workaround is to get rid
of XFS. Pity: I thought this was probably the best file system around. It may be, of
course, that this only affects a particular release: a Google search didn't find anything obvious except my last incident.
Still, the relatively copious notes I made last time were of great help. It's clearly worth
keeping this kind of information, even if it's boring for the casual reader. The next step
I did was xfs_repair, as before.
root@naan:/etc/network# xfs_repair /dev/sdb1 Phase 1 - find and verify superblock...
bad primary superblock - bad magic number !!!
attempting to find secondary superblock...
...............................................................................
This takes a long time—I think it reads the entire disk, printing out many dots.
Finally it finishes:
...............................found candidate secondary superblock...
verified secondary superblock...
writing modified primary superblock
sb realtime bitmap inode 18446744073709551615 (NULLFSINO) inconsistent with calculated value 129
resetting superblock realtime bitmap ino pointer to 129
sb realtime summary inode 18446744073709551615 (NULLFSINO) inconsistent with calculated value 130
resetting superblock realtime summary ino pointer to 130
Phase 2 - using internal log
- zero log...
ERROR: The filesystem has valuable metadata changes in a log which needs to
be replayed. Mount the filesystem to replay the log, and unmount it before
re-running xfs_repair. If you are unable to mount the filesystem, then use
the -L option to destroy the log and attempt a repair.
Note that destroying the log may cause corruption -- please attempt a mount
of the filesystem before doing this.
As on the previous occasion, the mount failed, but the superblock had been recreated:
Ran xfs_repair -L immediately. It ran much more quickly:
root@naan:/etc/network# xfs_repair -L /dev/sdb1 Phase 1 - find and verify superblock...
sb realtime bitmap inode 18446744073709551615 (NULLFSINO) inconsistent with calculated value 129
resetting superblock realtime bitmap ino pointer to 129
sb realtime summary inode 18446744073709551615 (NULLFSINO) inconsistent with calculated value 130
resetting superblock realtime summary ino pointer to 130
Phase 2 - using internal log
- zero log...
ALERT: The filesystem has valuable metadata changes in a log which is being
destroyed because the -L option was used.
- scan filesystem freespace and inode maps...
- found root inode chunk
Phase 3 - for each AG...
- scan and clear agi unlinked lists...
error following ag 0 unlinked list
- process known inodes and perform inode discovery...
- agno = 0
b6471b90: Badness in key lookup (length)
bp=(bno 24144, len 16384 bytes) key=(bno 24144, len 8192 bytes)
b6471b90: Badness in key lookup (length)
bp=(bno 47296, len 16384 bytes) key=(bno 47296, len 8192 bytes)
- agno = 1
After that, lost+found was once again full of files, but I was able to reboot, and it
didn't seem to have impacted the way the system runs. I'll use it like that until the next
version of Ubuntu comes out (8 days to go, it
seems), and then I'll install a new system with a different file system.
CJ and Sue along this morning to look at the greenhouse, and started trying to work out how
it fits together. Made surprising progress. Yvonne went
over to the Yeardleys to pick up the screws, but they had been lost in the course of the
years, so we'll have to get some new ones. Not much chance of starting tomorrow now.
Followed up on a label we found on the door of the greenhouse, and found the manufacturer
on the web. Unfortunately they
don't have any online manuals for erecting the thing.
The weather was surprisingly warm today—it went over 30°, and we ate all our meals on
the verandah. Somehow it was a little too warm, though—wouldn't it be nice if
the temperature would stay in the mid-20s?
My vote is to drop the pop-ups. I find them annoying ... and there's precious little
information in them of any interest (or value) to me. I might occasionally click a picture
to see a bigger version, but I'll never care what lens was used, or what the exposure
settings were.
Of course, there's precious little of interest to most people in any part of my diary. I do
like the idea of the popups, but the way they follow the cursor annoys me. The instructions tell me that all
sorts of things are configurable, including the position, the font and the size. But none
of that works for me. I think I should look at some of the alternatives.
One unexpected problem came from the W3 validator, which tells me:
Line 322, Column 54:
document type does not allow element "tr" here
The element named above was found in a context where it is not allowed. This could mean
that you have incorrectly nested elements -- such as a "style" element in the "body"
section instead of inside "head" -- or two elements that overlap (which is not allowed).
This is part of a Javascript assignment:
<script type="text/javascript">
exif_info [0] = ["Image ct-1.jpeg", "<tr><td>Author:</td><td>Greg Lehey</td></tr><tr><td>Date taken:</td><td>Thursday, 1 October 2009, 9:34:00</td></tr><tr><td>Exposure:</td><td>1/250 sec, f/10.0 (EV 14.6), 100 ISO, Normal saturation</td></tr><tr><td>Camera:</td><td>OLYMPUS IMAGING CORP. E-30</td></tr><tr><td>Lens:</td><td>Olympus Zuiko Digital ED 12-60mm F2.8-4.0 SWD</td></tr><tr><td>Focal length:</td><td>32.0 mm (35 mm equivalent: 64 mm)</td></tr><tr><td>Focus:</td><td>0.76 m (0.69 - 0.85)</td></tr><tr><td>Meter mode:</td><td>Center-weighted average</td></tr><tr><td>Flash:</td><td>Fired, Fill-in (+0.7 EV)</td></tr>"];
</script>
And yes, this is really all on one line, because Javascript wants it that way. But why
should the validator be looking into Javascript strings? This looks like a bug to me. If
it's part of any HTML specification, it's an extreme misfeature.
Low cost web development
My spam filters are getting most spam now, and also a considerable quantity of false
positives. One message that got through recently was (somewhat trimmed):
Date: Tue, 20 Oct 2009 11:05:27 +0530
From: "Team Ads" <enpershyaam@gmail.com>
Subject: Web Site Designing & SEO (internet marketing) for low cost - Coim
batore, INDIA
Content-Type: text/plain; charset="iso-8859-1"
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
" http://www.w3c.org/TR/1999/REC-html401-19991224/loose.dtd">^M
= ^M
^M
^M
= ^M
^M
DOES you company need
any of ^M
= these SERVICES?
^M
^M
^M
= ^M
^M
^M
^M
Team Ad's. Coimba= tore, INDIA (FREELANCER) ^M
Web = Designing and SEO for low cost ^M
Mob: +91 99422 20302, 92624 01244 ^M
mail us - ^M
admi= n@teamads.com
______________________________________= _____________
Our Best Services: ([1]htt= p://www.teamads.com) ^M
- Web Desig= ning NEW ^M
OFFER bel= ow !!!
- ^M
= Web Development (PHP - Web Portals, Classifieds etc,)
- Web ^M
This was really just text/plain; there was no HTML version. I think I can safely
answer this one with “no”. About the only thing he's done “right”
is to get through my spam filter.
CJ and Sue along today. We had planned to put the greenhouse together, but we still don't
have the screws. Took one out as a sample. It's amazing how primitive the things look,
with straight slits and square nuts:
That's not ideal for modern tools; maybe we can find something else.
Instead of the planned work, CJ did some general spring cleaning, removing and burning the
remaining wood from the pig sty that I dismantled two years
ago, and also pulling out dead and dying trees, including the Hakea laurina which fell over some months back. Despite the
forecast, it was also sunny and the winds were low, so spent some time spraying weeds. I
can see another long session of mulching to follow.
Also mowed the lawn for the first time this spring, after I finally got the thing to start.
The problem seems to be that the spark plug has lost the ferrule onto which the cable fits,
and the connection is pretty loose. The carburettor also seems to be leaking, and various
screws in inaccessible places need tightening up. Time to have it serviced.
One thing that can't be blamed on the condition of the mower was that I somehow managed to
catch a garden hose in it and drag it some distance. The hose was connected to a tap, and
it pulled out the tap and its plastic riser:
As mentioned in the
HTML 4 Recommendation's note about
specifying
non-HTML data in element content, end tags are recognized within
SCRIPT elements, but other kinds of markup--such as start tags and
comments--are not. This is an unintuitive quirk of SGML for elements defined to have
CDATA content.
Call it a quirk if you will; for me it borders on insanity. The result, along with another
issue that the author of the Javascript had ignored, was that I had to change my text from
this relatively straightforward text:
The HTML comment seems to be needed too, and the close HTML comment needs to be commented in
Javascript, with the // form that I hate so much, since otherwise it wouldn't close
properly. What a mess! But at least it validates again, though I've written a page about the topic to which my “Valid XHTML” icon
at the bottom of these pages links.
In the process, did some more checking and once again ran afoul of the naming conventions
for values of id and name in anchor tags. I had been using the photo name
for part of the tags for linking to photos, but they can contain illegal characters.
Strangely, it was relatively simple to remove the photo name from the name, and it
simplifies the URLs.
We're having visitors again. Today the first arrived, Hanna Barth
from Karlsruhe, and she and Yvonne spent most of the afternoon at Chris' place. Chris along later
for dinner, and of course she had to show Hanna how to maltreat cats:
Only a year ago I put a 500 GB disk in dereel, my new main machine, and there was
more than enough space for everything I needed. But already the writing was on the wall: my
photos are taking up more and more space, and by today there were 212 GB of photos, and free
space was down to 3 GB. Time for a new disk.
The problem is, I want to rebuild the machine soon, preferably with some virtualization
layer, but that won't be until FreeBSD8.0-RELEASE at the earliest. In the meantime I don't want to make any too
far-reaching changes. But I have a second disk on the machine, /dump, which had
about 60 GB free, so I copied some of the less interesting archive stuff there. No problem.
Checked that the data got there, then started deleting the old files. And exactly at that
moment my time bomb struck: about 2 weeks ago, I converted my camera copy scripts to use
mtools, but I still had
a mount call in there, and I didn't notice it until after disconnecting the camera.
It was clear that the system would freeze up at some point, but it was strange it happened
right now. fsck. fsck. fsck.
It's funny how spam comes in waves. Lately there hasn't been much, but today I got one that
puzzled me: it had a SpamAssassin score of about 30, but it still got delivered. Further
investigation made it clear why:
Received: from [82.114.73.162] (unknown [82.114.73.162])
by w3.lemis.com (Postfix) with ESMTP id C771B3BA57
for <grog@lemis.com>; Thu, 22 Oct 2009 08:10:18 +0000 (UTC)
Received: from 82.114.73.162 by gort.ebay.com; Thu, 22 Oct 2009 09:10:18 +0100
Received by gort.ebay.com? Not a hope.
That's a faked header, clearly aimed at people like me who have so much trouble with broken
mail from eBay that they have a separate procmail rule for it. You might argue that
that's not eBay's fault, but it is: if their mail wasn't such a complete and utter mess, I
wouldn't have that problem.
tvremote lives again
Also downloaded an ISO of FreeBSD7.2-RELEASE and installed it on a USB stick for the laptop which I used to
call tvremote. Its last use was for the kitchen, but we still haven't found a good
name for it. This time I was able to boot, but I've decided to leave the stick read-only
and NFS mount most stuff from dereel. Still don't have X running properly, but most
stuff went relatively smoothly. I should keep a copy of the stick and use it for multiple
machines; brewer looks like another candidate.
I've come to the conclusion that flash exposure is still a dark art. For some time I've had
my flash exposure compensation set to +0.7 EV, but the exposure of last night's photos was
all over the place, up to 2 EV difference. Some of that may be because I haven't waited for
the flash to be fully charged, something I should pay more attention to. But some of the
photos were decidedly overexposed, and that by more than 0.7 EV, something that
undercharging wouldn't explain. It doesn't seem to be a problem peculiar to the Mecablitz 58 AF-1 O digital: it happens with the built-in flash unit of the
E-30 too, and it
happened with the E-510 as well.
UFRaw improvements
When I started using ufraw, I had
big problems with the settings. They're still very
wrong: typically, the exposure is off by about 1.3 EV, and the saturation and contrast don't
match what the Olympus software or the in-camera JPEGs produce. In addition, the automatic
exposure adjustment created photos that were uselessly “underexposed”. But at
least that has changed: now automatic exposure compensation works (and shows values round
+1.3 EV, as I had empirically discovered). That was particularly helpful with the flash
shots.
A bit more garden work, mainly weeding. Also decided to rearrange
the Gazanias, and split one bush into
about 6 smaller plants. Gradually things are taking shape.
More attention to my disk reshuffle today; the system had crashed before deleting all the
files from /home, so deleted the remainder. Well, that was the intention. The
system crashed again. This must be related to issues I know about soft updates. But
it's a real pain, and it took me a couple of hours to sort things out; first
run fsck, which took over an hour, then reboot the machine into single user
mode—why doesn't it respond correctly when I try to interrupt the boot?—turn off
soft updates on the file system, remount, delete the files, umount again, set soft
updates, then go back to multi-user mode. What a pain. It must be possible to catch this
situation before it causes a crash.
Spent more time looking at the new kitchen computer today, trying to get X running. The
mouse had the same issue that it has always had, and once
again I could turn to this diary for a quick solution. But there were other issues.
Firstly, I wanted to use the software installed on dereel rather than installing
copies on the laptop. That requires using ldconfig, whose man page tells me:
FILES
/etc/ld.so.conf Conventional configuration file containing
directory names for invocations with -aout. /etc/ld-elf.so.conf Conventional configuration file containing
directory names for invocations with -elf.
That's completely incorrect. Those files no longer exist. After some time, found the
correct information in /etc/defaults/rc.conf, including:
That worked well enough after finding out how. But I still couldn't get X to work.
Configuring X has been a fairly constant problem over the last 20 years. When I first
installed X on BSD, about 17 years ago, it took me forever to create a correct
configuration. But then people gradually improved the situation, even to the point that you
usually don't really need a configuration file to get a basically functional X display. In
the cases where it doesn't work out of the box, X -configure usually does
the job.
This time X -configure didn't work either: I got a message “Can't assign
video memory”. I've seen that, too, with my other Dell laptop, but not with this one.
On further investigation, discovered that it had created a configuration file for two
screens, something which this el-cheapo laptop most definitely doesn't have. Removed that,
but it still didn't work. /usr/local/Xorg.0.log showed that both keyboard and mouse
had been disabled.
Discussions on IRC suggested that this was the HAL issue. If I understand this correctly,
HAL is Linux's answer to FreeBSD's devfs, and newer versions of X require it. From the
man page:
hald is a daemon that maintains a database of the devices connected to the system system in
real-time. The daemon connects to the D-Bus system message bus to provide an API that
applications can use to discover, monitor and invoke operations on devices. For more
information about both the big picture and specific API details, refer to the HAL spec which
can be found in /usr/local/share/doc/hal/spec/hal-spec.html depending on the distribution.
It's not clear where hald comes from, but it got installed on my box. But it's not
on the FreeBSD web site, there's no
information on how to start it at boot time, and /etc/defaults/rc.conf doesn't
mention it. From IRC I got the suggestion that I set the following in /etc/rc.conf:
What a mess! But why? If I install the same version of the X server on Linux, it works
fine. As far as I can see, there's something basic missing in the FreeBSD configuration.
How can people release software like this?
Apart from that, didn't do much. More weeding—will it ever end?—and of course
my Saturday photos. The sky was pretty much cloudless, so I used the polarizing filter
instead of the graduated neutral density filter. I'm not sure that it isn't too strong: in
some cases the sky takes on an unpleasant steely grey colour:
Laurel Gordon came along today, brought by her son Leah and by a somewhat roundabout
way—they lost their way coming from Ballarat, and they were almost
in Geelong before they realised they had
taken the wrong way.
More photos in the evening, of course, and did some more playing around with flash. I've
known about bounce flash for decades, but for some reason I haven't used it much since I've
taken up photography again. Today I changed that, with good results (left direct flash,
right bounce):
More work trying to get X to work on my laptop today. Michiel Overtoom sent me a message
suggesting adding the following line to the ServerLayout section of
Xorg.conf:
Option "AutoAddDevices" "Off"
Tried that and got some improvement, but there are still lots of problems, partially
related, it seems, to my decision to put a lot of the software on an NFS mount. More
experimentation needed. I should also consider trying the new Ubuntu when it comes out.
These aren't my own photos, clearly. They'll come later. For now the question is where to
plant them.
In fact, the remaining photos are my own photos: the images I originally chose
have suffered from webrot. The Itea ilicifolia image also disappeared, but sadly the
plant did too, and since I've forgotten what it should have looked like, there is no image
any more.
It's been 2 years since I bought my “new” scanner, and somehow I haven't used it
much. Spent a lot of time scanning in some of my oldest negatives, in the process coming to
various conclusions:
Photography has come a long way since 1964. The following photo was taken on
Kodak Plus-X Pan, 125 ASA (now ISO). The grain is quite noticeable even at normal
screen width:
I think that my current camera would get about the same “grain” at ISO 3200.
The photos aren't very good, of course; they're some of the first I ever took, with a
camera with no rangefinder and no exposure meter. I also needed to learn something
about composition.
It's almost impossible to get a clean scan. All sorts of crud gets onto the negatives.
I think that this happened to me at the time, too, when making prints. All that's gone
with digital imaging, but here I think I'll have to put up with it. The scanner
software has a “dust reduction” facility which it uses somewhat grudgingly:
it resets the box before every scan, and I have to set it again. It doesn't seem to
work overly well.
Scanning photos takes a long time. These were done at 1200 and 1800 dpi, and
each negative took over a minute. I made a conscious decision not to take up even more
time by cropping and other postprocessing. If I find photos worth the trouble, I can
still do that.
On the other hand, it would be nice to manually input the exposure data as EXIF. I
should consider the options there.
CJ along today to do some various work in the garden. First, Yvonne has prepared a series of posts for her “Japanese” garden, which
required knocking into the ground:
The problem is that a number of the brackets for the roof are broken, and I don't know where
to find replacements. I suppose I'll have to call up the manufacturer. They're made out of
plastic, and they don't look like an easy thing to fake:
Leah and his aunt came to pick up Laurel today. The immediate reason for her visit was that
she has bought our horse float, which we hardly use any more:
Before we got that far, we discovered that Laurel's car has a new kind of electricity
connection to the float. The only one I have ever seen, here or in Europe, is a circular
connector with three rows of alternate pins and sockets. This one was a rectangular one
with a single row of 7 contacts. They went off into Ballarat and fortunately found an
adaptor with no difficulty.
Also went theough the garden with Laurel again to identify new plants. She thinks that
Mystery flower 18 could be a Kalanchoe, which seems
possible; it looks similar to
the Kalanchoe blossfeldiana:
I also understood her to have said that Mystery flower 17 could be
a Crocosmia, but that doesn't seem to
match the appearance. More investigation needed.
Spent a lot of time scanning photos taken in 1964 and 1965 today. I suspect I have never seen most
of them: photographic paper was expensive, and many of these photos were of very poor
quality. I developed them all myself, with the possible exception of the very first; the
marginal notes in the negative album describe the film (frequently off-cuts of commercial
film such as Kodak 30 ciné) and the developer (usually Promicrol). There's also a surprising amount of physical damage to the negatives,
suggesting that the problems I recall getting the film into the spiral were worse than I
remember. But amongst the mess there were a few interesting photos, including a couple of
views of Kuala Lumpur that I haven't
seen before, and which I turned into panoramas. The first is a view of the
“Padang” (now called “Merdeka Square”) from Federal House
(now Rumah Persekutuan), and the second of Mountbatten Road (now called Jalan Tun
Perak). The black square at bottom left of the second image is because of the differing
formats of the two images I used to make the panorama (left landscape, right portrait). In
view of the historical interest, I decided not to crop it. These photos would have been
taken about the end of August 1964.
The Clematis at the south end of the
verandah is growing like fury, but I saw a group of very unhappy looking leaves, which
appear to have a fungal infection:
Summer's here! Only a week ago the daytime high temperatures were in the mid-teens, but
today we hit 32°. Didn't feel overly motivated as a result, but did plant a couple of
plants: the “Phyllis Bide” climbing rose and the “Monsieur Tillier”
bush rose. Used fresh compost from the compost heap, which is now looking very good. I
filtered it through the filter I had built for this purpose, but there were very few solid
twigs in it.
The “Monsieur Tillier” is supposed to be quite big, and the place I put it
(replacing an Erysimum) looked smaller
than appropriate, and I ended up moving it about 40 cm; better now than later.
More toys arrived today: a polarizing filter for the macro lens, also a
10 dioptre close-up for the same lens. I've established that the lens is good, but it's a
pity it only comes down to a magnification of 2:1 (corresponding in subject size to 1:1 on
35 mm). With the close-up lens it appears to come approximately twice as close. How do you
calculate these things? The way I learnt it, for thin lenses the law is 1/f = 1/u
+ 1/v, where u is the distance from the object to the “centre” of
the lens, v is the distance from the focal plane to the “centre” of the
lens, and the magnification is v/u, usually written as the ratio u:v. So
at 2:1, u is twice v. For a 50 mm lens, that makes v = 66.7 mm
and u = 133.3 mm.
What happens when you add a 10 dioptre lens in front? For thin lenses, you simply
add the
dioptres. A dioptre is 1000 / f, so 50 mm is 20 dioptres, and 30 dioptres are
33.3 mm. So with v = 66.7 mm, u shrinks to 66.7 mm, and the magnification
goes to 1:1.
So much for the theory, and for my assumptions: modern lenses are anything but thin, and the
supplementary lens is a long way from the optical centre of the main lens. Put the lens on
and tried it out, and it roughly seems to be the case; I'll test in more detail some other
time.
Also a little bit of playing around with the polarizing filter, which certainly showed the
advantage in the area I was expecting:
Apart from that, more scanning of images. I kept careful notes of the photos I took, and I
had them until a couple of years ago, when I moved here. Now I can't find them any more,
and it's a real pain trying to guess which photo was taken when. Given that I classify the
photos by date, this is a serious issue. I should mount a search.
Up this morning with the realization that I had planted our ”Monsieur Tillier“
in the wrong place—thus the issue with the exact placement. So ended up replanting it
a second and hopefully the last time.
Also did some other planting, though we're already running out of space. Planted the
Haworthia that we
bought in April and which had been sitting in a pot until
its roots were firm enough, so that we could plant the
dwarf Salix in the pot it was in, and also
planted the Libertia
grandiflora in two different places near the
bigger Silver Birch. Then I ran
out of ideas and of places, so gave up. We still have 3 roses to plant, and I keep
forgetting which goes where.
We've already noted that Piccola has been chasing magpies, and that the magpies delight in
luring her on. Today things were different: we found a young magpie in one of the trees,
apparently too young to fly, and it was screaming for help. That attracted Piccola, of
course, who actually managed it up into the tree and got hold of it. Fortunately we managed
to shake her off and bring her inside: there were other magpies around, and they're
particularly aggressive in such conditions. I don't have much hope for the bird; hopefully
it will die before Piccola gets hurt.
More scanning today, including some comparisons of old colour negatives. I had already
scanned one particular film with the Canon scanner, but I didn't notice until I had tried again with the Epson. That at
least gave me the opportunity to compare the rendition. In both cases, the colours of the
scanned image were pretty terrible (and on the Epson the option “recover
colors” seems to make no difference), but the Ashampoo photo optimizer (Ashampoo doesn't believe in
useful links) makes them a lot better—a little better with the image of the Epson than
the Canon, which also managed to lose the bottom of the image. The dull greens are probably
because of the film (Ektachrome X, I
think).
The following photos show the results of the scans, first Epson, then Canon. The left-hand
ones are the raw output of the scanner software, and the right-hand ones after optimization.
I've written in the past about Australian officialdom's
obsession with “speeding”, but more or
less given up hope of any improvement. Somebody pointed me to an article in the Sydney Morning Herald which sounds surprisingly like my own opinion. Is there still hope?
The original link
http://www.smh.com.au/drive/motor-news/is-speed-really-a-killer-20091030-ho9e.html died.
I've replaced it with one that looks like being the same. In particular, the date and
filename part of the URL match, as do the quotes below.
Still, there are some interesting statements. From the constructive selective quote
department:
It believes speed is put down as a primary cause when factors such as ... police pursuits
and other criminal activity are major contributing factors.
In reality, the statement is just badly formulated:
It believes speed is put down as a primary cause when factors such as alcohol, drugs,
unregistered vehicles, unlicensed drivers, police pursuits and other criminal activity are
major contributing factors.
“It” is the National Motorists
Association Australia, if the title on the home page is to be believed. It's not
clear how unlicensed vehicles and drivers can contribute directly to road deaths; I suspect
this is just inaccurate quotation. The NMAA clearly has considerable objections to the
current situation. It also seems to be another site with variable links, so this link may
die.
Perhaps comparing apples with oranges, the RTA argues that the British road system (which is similar to ours) is safer. It
says the German road toll is 50 per cent higher overall than in Britain, where all
motorways have mandatory speed limits.
It's interesting that this is in contradiction to the information that the Australian Government published and which I quoted back in
January.
This link is now dead. I don't know whether it was because it contradicted official
policy, or whether it was just a victim of web site reorganization, but I suspect the
latter.
I've updated the table to add the statistics for the UK:
Australia
France
Germany
UK
Victoria
Deaths per 100,000 people
7.7
7.7
6.2
5.4
6.6
Deaths per 10,000 vehicles
1.1
1.3
0.9
1.0
0.9
Deaths per 100 million km
0.8
0.8
0.7
0.6
0.6
The way I read the differences between the UK and Germany, British vehicles drive further
and are driven by a smaller percentage of the population. The deaths per 100,000 people is
clearly the least dependent on road behaviour than the others, and the differences are
inconclusive: the UK has more deaths per unit vehicle (so the UK drivers have fewer cars),
and they have fewer deaths per unit distance. But there's no way this can be in agreement
with the RTA's claim that the German “road toll” is 50% higher. The RTA is a
State Government agency, so they must be aware of it. Without seeing any proof, I'd claim
that this is another unfounded claim.
There's also a poll on
the topic—not very detailed, and clearly opinions rather than arguments, but it's
still interesting to see the (current) results: 61% of respondents thought that higher speed
limits would lower the “road
toll”, 31% thought “only if the roads were up to it”, which sounds
like a more cautious (and sensible) answer in favour. Only 5% said “No, Australian
drivers aren't skilled enough”, and only 3% toe the official line “No, speed
always increases fatalities”.
My 404 document for the web site sends me mail when an internal link fails, so that I can
fix it as quickly as possible. I'm continually amazed by the number of false positives.
Direcway has a crawler which seems to ignore
HTML syntax:
The thing is, there is no such link on that page, and I'd most certainly never
create a page with a space in its name. It's part of the Javascript in the page:
I've put in all the incantations to keep the XHTML spec happy, and the W3 validator seems happy, so I must assume that this is a bug with direcway (who, it seems, now call
themselves HughesNet).
Another really hot day today—the overnight low was about 18°, and it rose to 34° for
much of the afternoon. Once again it kept me from doing much in the garden, and I spent
most of the day in the office scanning negatives. It's amazing how much better the photos
got when I bought the Pentax SV.
Chris along for dinner in the evening. Jana is vegetarian, so I had another excuse to cook
Indian food.
Do you have a comment about something I have written? This is a diary, not a
“blog”, and there is
deliberately no provision for directly adding comments. It's also not a vehicle
for third-party content. But I welcome feedback and try to
reply to all messages I receive. See the diary overview for more details. If you do
send me a message relating to something I have written, please indicate
whether you'd prefer me not to mention your name. Otherwise
I'll assume that it's OK to do so.
