Black Coffee (Deprecated)

Workflow... Nikon Super Coolscan 4000, Photoshop 7, and the Epson 2100

I've been terribly slow in turning my photos from our trip to Kerala into presentations, either web or printed.

Last week I arbitrarily started on a scrapbook based on the photos I took around Kochi's Chinese Fishing Nets. When I printed the first photo at a modest 8" width, I was dissapointed to see that it wasn't pleasingly sharp.

Where's the problem coming from?
I used Kodak SW-100 film, which has very fine grain, a Canon 100-400mm lens with image stabilization, scanned the slide w/ Nikon Super Coolscan 4000's maximum resolution (4000 dpi), did next to no modifications in Photoshop, and printed it out with high resolution on the Epson 2100.

All of those steps were reasonable; where did the workflow break down and give me medicore sharpness?

Start with the source. I checked the slide with a 4x loupe. It looked quite sharp; an excellent image. But 4x isn't much amplification. So I dragged out my Leica Pradovit slide projector and viewed the slide as a three-foot wide image. There was the problem! The image just lacked natural sharpness to begin with. In low afternoon light I'd been standing 50 feet away. Although the image is ok, it's not tripod-sharp.

Comparing the print and the digital image of the slide to what I was viewing on the projector screen, the sharpness complaint was clearly originating with the photo, not some later step in the workflow. (This is not to say that I'm not introducing sharpness problems later on in the chain, just that the source image was the pareto cause)

Correcting the image
So to make this a presentable image, I'm going to have to use Photoshop's sharpening capabilities. There are scores of lousy, conflicting tutorials presenting a huge gamut of techniques ranging from super-simple to enormously tedious.

So my current project is to figure out the right way to sharpen this image into its most presentable state.

Resolution
Almost universally, tutorials have warned to apply the unsharp mask on an image that is already adjusted to the final output resolution of the printer.

Ok. Fine. What should that be, then?

Epson says my printer has maximum resolution of 2880x1440 dpi. So this is suggesting that the dots are twice as wide or twice as tall? Why is its printing resolution an area unit (2880x1440) when the Nikon Super Coolscan 4000's resolution a one dimensional unit (4000 dpi) ?

Photoshop :: Image :: Image Size also presents resolution as a one dimensional unit. (in the case of my scanned image, it calls it '4000 pixels per inch')

My pessimistic guess is that I cannot even assume that Photoshop's "Pixels per inch" is the same as my printer's "dots per inch"

What I need to figure out is the function that maximizes the quality of the Epson printer print, while not creating a needlessly bloated image (where photoshop creates 'fake' picture detail just because I forced it to be a certain resolution) given the following inputs and outputs

inputs
sc_scan_resolution: the resolution I tell the Nikon 4000 (sc) to scan with. (in dpi)
ps_pixel_width&height: photoshop settings under "Photoshop::Image::Image Size" units "pixels"
ps_document_width&height units "inches"
ps_document_resolution: units "pixels per inch"

DJ 31-Die on the Decks Tonight

So I (31die) spent a "fun" hour or two trying to figure this out:

I was downloading as usual, and of a sudden I got a pop-up message indicating a delayed write to one my my 200GB drives failed. Uh-oh.

Poke at it for a while and realize that I had put enough data on it that I was cresting the 128GB point, which is the most current limit on HD size. Western Digital addresses this by including a controller card that understand big drives, and pretends to your system that the drives are SCSI -- I guess the addressing limitation is only from IDE drives.

So I had installed both the 200GB drives on said controller card, while two 80GB drives are on the motherboard, as well as the DVD and CD-RW. The whole point of this confuration was to allow me to use the monster drives.

Poke and dig at it for a while, and it turns out there is a WinXP/SP1 registry variable that lets you enable "48-bit LBA", which I guess is the fix that extends IDE capacity yet again. I have a MB bios which supports it, so I enabled. Of course, I shouldn't have to do this, because the controller card is supposed to hide all this unpleasantness. Sure enough, it didn't fix anything.

WD tells you to install their driver for the card, but when I installed the card in the system WinXP happily recognized it. I tried to update the driver to the one WD provided, but WinXP informed me that the driver it was using was already the most current or most suitable for my hardware or whatever.

As an experiment I decided to bully the OS into using the latest WD (actually a rebraded Promise Technology) drive for the IDE controller card. I told it to install the specific driver I was pointing it at, whether it thinks it knows better than me or not.

Lo and behold, everything works now. Verrrrrry tedious. I have a feeling it comes down to something stupid, like the current WD driver for the card isn't signed by Microsoft, so Microsoft will always prefer the outdated signed driver or something.

Anyway, it works, and it is a fix for a problem other people are having I think. I'm tossing it at you as a somewhat incoherent story with the thought that maybe you can jam it up as a "31die relates this rambling story in the hopes it may help others" post to black-coffee or something.

I'd put it on the eldie-blog, but apparently that isn't even indexed by search engines...

Suspect in killings, abduction surrounded

CNN.com - Suspect in killings, abduction surrounded - Apr. 20, 2003

FBI Special Agent Linda Vizi said Hixson was a former Navy SEAL who "may have training with weapons" and was definitely considered "armed and dangerous."

I guess Pocono Lake police are hoping that Hixson is one of the Navy SEALS who did not actually receive training with weapons, and instead plans to sit-up his way to safety.

Make Your Sunday Even More Depressing

"Indians in government service [in India] and most other offices can legitimately stay away from work for 202 days a year. The 202 days are made up of the 104 Saturdays and Sundays in year, 26 religious and national holidays, 30 days' privilege, 12 days' casual leave and 30 days' sick leave."

--Singapore Straits Times, April 20, 2003

"On the Ground in Iraq, the Best Compass Is in the Sky" by Seth SCHIESEL NY Times

On an afternoon early this month, in the desert near Najaf, Iraq, elements of an elite United States Army unit received word of a column of almost 60 vehicles, including about two dozen tanks, moving along a nearby road.

Some of the soldiers thought it could be Saddam Hussein's Republican Guard.

Then a general in his Humvee leaned over to a computer console that is part of a satellite-based navigation system called FBCB2. He tapped in the military grid coordinates where the mystery force was located. Then on the screen, up popped the little blue symbols that represent friendly units, rather than the red icons that the United States military uses to designate enemy forces.

It was not the Republican Guard. It was a separate United States division.

During the cold war and even the 1991 Persian Gulf war, satellite technology was not an everyday part of the lives of foot soldiers or even generals. But in the Iraqi desert, satellite technology - specifically the Global Positioning System, or G.P.S. - has become a fundamental and pervasive navigation tool for ground forces.

G.P.S. gadgetry has become almost as much a part of army life as shovels and cigarettes - whether integrated into vehicles in advanced systems like FBCB2 (often referred to as "blue-force tracker"), used in the hand-held receiver known to soldiers as the Plugger, or even bought off the shelf.

"Primarily the way that G.P.S. technologies have changed the way the army can perform its mission is it has given us a more accurate way to navigate the battle space," said Lt. Col. William S. Harborth, the Army's product manager for Global Positioning System technology.

That means the devices simply help soldiers figure out where they are. Perhaps even more important, the ability to define location precisely can help soldiers figure out where other units are.

"Increased accuracy is more important because if you know better where you are, you can ensure that you reduce fratricide," Colonel Harborth said. "In the old days, there was some human error in determining your location on the ground."

Satellite navigation continues to be crucial for long-range weapons like cruise missiles, and G.P.S. is essential in the sort of unmanned aircraft that saw their first broad deployment in Afghanistan. In contrast, the main such tool among ground troops, the Plugger, is in some ways less sophisticated than gear found at Wal-Mart or in rental cars - its utility in traversing the open desert diminished as forces entered urban areas, for example, since roads and landmarks are not programmed into it.

Still, the increasing use of satellite-based systems for navigation - and for "situational awareness," in military parlance - is one of the biggest changes in United States ground operations since the 1991 gulf war.

During that war, the Global Positioning Satellite network was in its infancy, and among front-line units, a single G.P.S. receiver might be allotted to an army company, perhaps numbering 180 soldiers in the infantry. Now the Army says that it has more than 100,000 Pluggers (the name is derived from the initials for their full name, PLGR for Precision Lightweight G.P.S. Receiver). In Iraq, the leader of each combat squad, which might include nine soldiers, often has a Plugger at hand; in some Army units, Pluggers are even more numerous.

The Marines have adopted the technology more cautiously. Matthew Brandt, the Marines' project manager for G.P.S., said the corps had purchased only about 5,400 of the units and generally deployed them at the platoon level. (A platoon might include three to five squads.)

That may be one reason that at least some marines are carrying their own civilian-grade G.P.S. devices from home. The civilian devices, made by companies including Garmin International, are typically smaller than Pluggers and, though not quite as precise as Pluggers, are apparently sufficient for everyday purposes.

Those purposes can be as trivial as finding the chow line. Before the shooting started in Iraq, some soldiers in front-line units were using their Pluggers to navigate through the dark and sand to the mess tent.

As with most technologies, however, satellite navigation is only as useful as the human intelligence guiding its use. For instance, in late March an American military detachment was sent to pick up some prisoners near Najaf. The soldiers were told the coordinates of the captives.

Their Plugger unit worked fine and the soldiers reached the coordinates. But they did not find the prisoners there. Instead, they came close to a mortar attack. The human intelligence had failed, not the device.

And even with the growing use of satellite navigation devices, there are gaps. A prominent setback for the Army in the early days of the war was the ambush of members of the 507th Maintenance Company near Nasiriya, Iraq, in which eight soldiers were killed. A private captured in the confrontation, Jessica D. Lynch, was later rescued, and five others taken prisoner were found alive north of Baghdad on Sunday.

The Nasiriya episode, which occurred while the soldiers were traveling in a convoy of trucks and other vehicles, was initially attributed to their having taken a wrong turn off a major highway. The Army has refused to comment publicly on precise details of the incident, and more recent accounts indicate that the convoy was ambushed after having stopped to repair vehicles.

But a technology expert with the American forces in the region and a civilian expert on military G.P.S. both said it was unlikely in any case that the captured unit had a G.P.S. device on board.

While Plugger units are almost ubiquitous among front-line combat units, they remain less common among units like maintenance companies, which are not generally meant to engage the enemy.

Even soldiers who have Pluggers are relying on devices that are in some ways primitive compared with their civilian counterparts. It is a curious position for the Pentagon, the driving force behind the creation of the constellation of 24 G.P.S. satellites in the 1980's and 90's.

The Plugger devices remain largely unchanged since their initial deployment in 1994 (although their cost has fallen from about $2,000 each to less than $1,000), and for many purposes, the relatively scant information they provide is sufficient. Soldiers can specify their destination, and the unit will tell them what direction to go. Using encrypted satellite signals reserved for government use, they are accurate to within roughly 10 yards, compared with 20 to 25 yards for civilian devices.

Built for resilience in combat, they are big (roughly the size of a small shoebox), heavy (about 2.75 pounds) and have a small text-based display incapable of showing maps or other information. In general, the units, which are made by Rockwell Collins, display only location, velocity (if the unit is moving) and time.

Civilian G.P.S. devices like the NeverLost system in Hertz rental cars, in contrast, are often able to display maps and other information.

The advanced graphical FBCB2 system used in Army combat vehicles, in contrast, allows commanders to electronically "see" broad swaths of a battlefield. In the version of FBCB2 known as "blue-force tracker," far-flung United States units not only receive their location information from G.P.S. but also communicate with one another using other classified satellite systems. Other versions of FBCB2 units receive their location from G.P.S. but communicate with one another using land-based radio.

(In either case, the system is connected by cable to a Plugger, which serves as the actual location-detection device. In fact, more than half of the Pluggers in the Army are not used in a hand-held mode. Rather they are used as "slave'' location-detection devices for other systems, which include air-defense batteries in addition to FBCB2.)

FBCB2, which has been in development since 1997, has been deployed in practically every tank and Bradley fighting vehicle in the Fourth Infantry Division, said Michael Lebrun, deputy director in the Army's command, control, communication and computers office. Elements of the Fourth Infantry, which in some ways is the most technically advanced of the army's infantry divisions, are on the way to Iraq.

The FBCB2 system displays the location of similarly equipped units in the area as blue icons. When any of the units spot enemy forces, they enter their location into the system. They are then displayed as red icons, and that information is relayed to other FBCB2 trackers.

Mr. Lebrun said that over the last seven or eight months, FBCB2 was deployed to other army divisions, though generally company by company. A tank company might include three platoons, each with four tanks.

For foot soldiers without access to the FBCB2, however, satellite navigation usually means getting their location from the Plugger and then using a paper map to plot their location manually.

That is why the Pentagon is ordering a new generation of hand-held G.P.S. devices, to be known as DAGR, pronounced "Dagger," for Defense Advanced G.P.S. Receiver. Rockwell Collins is competing with Raytheon for the right to produce the new system, which is scheduled to reach everyday soldiers next year. The Pentagon is to pick the winning company in September.

"Plugger is about 12 years old, and if you can make an analogy to the commercial electronics marketplace, just think about your cordless phone you had at home 10 years ago versus now," said Mark Youhanaie, Raytheon's strategy director for G.P.S. products. "Now, we can make these receivers more accurate. We can acquire the satellite signal more quickly. It has higher jam immunity, and we can give you that all in a package that is a quarter of the size of the old Plugger system."

For now, it appears that the Rockwell Collins contender is a bit smaller than Raytheon's, while Raytheon's boasts a bigger screen. Whichever company wins, however, the Dagger will weigh only about a pound and will be much smaller than the Plugger. Perhaps most important, the new devices will allow soldiers to see not just lines of coordinate numbers, but also a map that shows their location in relation to objects like minefields, rivers and enemy positions. The units will also incorporate graphical user interfaces.

Drawing a comparison to generations of computer operating systems, Steve Jones, the Rockwell Collins marketing manager for land navigation products, said that "Plugger is DOS, and Dagger is Windows."

By plugging the Dagger system into a military radio, soldiers may be able to display their location on the screens of nearby Dagger units or more advanced FBCB2 systems, Mr. Jones said.

The Dagger devices, which are meant to initially cost about $2,000 each, will be more advanced than the Plugger in other ways as well. While the Plugger receives its encrypted signals at 1,575 megahertz, the band also used for civilian G.P.S. devices, the Dagger will also be able to pick up signals at the government-only 1,227-megahertz band, allowing for additional accuracy. The 1,227 band is now used largely for military aircraft, cruise missiles and other airborne systems, military officials say.

The new system will also track all 12 G.P.S. satellites in each hemisphere at once. The old units can only track five satellites at once, and signals from four satellites are required to establish a three-dimensional position. In addition, current G.P.S. receivers are somewhat vulnerable to enemy equipment that beams false G.P.S. signals to indicate the wrong location, a technique known as spoofing.

The Dagger is meant to include classified technology that will help the device verify that the signal it is receiving is actually coming from a United States G.P.S. satellite.

It is still unclear just how many of the new devices will reach United States soldiers. "The plan was to replace all of the Pluggers in one year,'' said Mr. Brandt of the Marines, "and of course that depends on how much money Congress decides to give us, which is never certain."

But no matter how many are ultimately deployed, the new devices are meant to give the soldiers perhaps the most precious commodity on the modern battlefield besides life itself: information.

"The key is greater situational awareness for our soldiers so we bring them home alive," Colonel Harborth said. "That's it."

Best Ride

Just finished up the most exhilirating bike ride I've ever had. I did the Bukit Timah circuit in a bit less than forty minutes at nightfall after a torrential monsoon rain had soaked and saturated the track completely.

My bike really proved its mettle in the conditions. I set the shock absorption to roughly 'medium' settings (giving you an idea how sophisticated my 'tunings' are...) and took off. With a mediocre bike light on, powered by crappy Korean batteries, I could sort of make out the trail between the light and the falling sun, but there were large stretches where I was making downhill banking turns at high speed and I would feel the suspension taking massive blows, but was able to keep on going, unphased. It was really impressive. It was thrilling to be going too fast to really know what was going on, instead letting the bike sort out issues for me.

Today has been a big biking day for us. We originally planning to goto Johor State to ride at some plantation called Plentong. But when I woke up, I wasn't really in the mood for a ride over the Causeway. Sitting in the cool morning air reading the Economistand drinking strong coffee was more pleasant. So afterwards we took our bikes for a second attempt to find the Jalan Kayu trail.

We found it. I'll do a full trail writeup at some point, but basically it consisted of riding two miles through a two to six inch deep clay mud that smelled like human feces, followed by ten or fifteen miles of riding on sidewalks back to homebase. The lameness of this trail was attenuated by the unseasonably cool weather. Stupidly I snakebit my tire about two miles from the finish, so I stole Ling's bike and went ahead to collect the car and recover Ling. Nice cardio trip, but awful as an adventure exploration.

Throughout the trip I was having more trouble with my brakes, this time the rear brake seemed to have lost 80% of its power--basically useless. It felt like the brake control was pretty soft as well, so I figured it was a hydraulic problem, not a pad problem.

We took the bike over to Treknology 3 in Holland Grove (I recommend going to this branch -- they are very friendly here and quick to service the bikes) to have the tire fixed and the brake investigated.

Their diagnosis was brake pad contamination. They said they cleaned the pad with isopropyl alcohol and recommended I do the same for the disk. Frankly I suspected they were wrong in the case, but I couldn't test the brake because there was a sumatran monsoon soaking everything, and these aggressive knobby tires I have are hell on wet concrete.

Well, from there we stuffed ourselves with Margaret Road dumplings and then went home, repacked, and went off for my aforementioned nighttime run. Turns out the brakes were fine, and their solution accurate. Not sure what the contaminant was. Perhaps the disk got hit with some chain lube when I last lubed the chain or something. At any rate, it grabs ferociously once again and all is well.

HK to Check Temperatures of Arriving Air Travelers

Hong Kong Chief Executive Tung Chee-Hwa

Hong Kong will check the temperatures of inbound air travelers from next week to curb the spread of a deadly flu-like virus, the territory's chief executive, Tung Chee-hwa, said on Wednesday.
Air travelers leaving Hong Kong will be subject to temperature checks from Wednesday midnight, Tung told reporters a week after he promised a rescue package to help ease financial pain caused by Severe Acute Respiratory Syndrome (SARS).

He also said the government will consider measures to strengthen health checks on sea and land travelers.

High fever of above 38 degrees is one of the symptoms of SARS which has killed 61 people and infected over 1,200 in Hong Kong, the worst affected area outside mainland China.

Other symptoms include chills, muscle aches, a dry cough and breathing difficulties. The virus has killed around 160 people and infected over 3,400 worldwide.

It's Beetlejuice Baby!

Beetlejuice

Tiny headed dwarf Beetlejuice (JollyDwarf.com) came in this morning to ramble on and on about nonsense. Beet said that he was already drunk on Vodka this morning. Howard spent a few minutes asking Beet some questions and he got the good old ''Me!?'' answers from him. Howard would ask Beet a question and Beet would answer ''Me!?'' and then go off on a tangent for a few seconds. They asked him about everything from Hank the Dwarf to what he shaves with. Howard also said that Beetle will be shooting a porn movie soon where he sticks his head inside of the woman's private parts. You can find out more info about that at PicklePilot.com. Howard says he's worried about Beet doing that because it's dangerous. Beet says he'll be wearing a condom on his head when he does it.

KAP Rig todo for Outback2003

From 31die:

Here's an initial, overly ambitious, list of KAP revisions:

* Get a replacement spar for the heavy-lift kite

Spar is 77" long, 5/8" square hardwood. Might make sense on this
one to call "Into the Wind" and see if I can just buy a replacement.
Mike -- this kite is/was available from them, wasn't it?

* Move the camera TX unit back the pan-tilt assembly

Putting it on the picavet created a lot of grief, even if it did
help the range.

* Look at changing the pan servo to use a servo-city gearbox.

We'd have to check what the weight looks like (probably minimal) and
if it gives us the range of motion we want. I think the problem
with it is that it offers the same range of motions as before (120
degrees or whatever you nominally get from a servo) with more
torque, lower speed. The torque/speed change is useful, but we also
want to get to > 360 degree rotation. What I'm looking for here is
something that is an all-in-one solution, so we don't have to screw
with the external potentiometer again.

This change may be in the category of needless -- the current
solution does work, although it is mercurial. Camera TX improvements
are probably more important.

* Change camera TX patch antenna (mickey mouse ear?) to optimal
coaxial stub

I think this is a definite. The only thing I'm worried about is
that the web site made a lot of noise about this modification being
very sensitive to getting the dimensions spot-on. I don't know if
the people having trouble were off by some gross error, but I doubt
my ability to any better than they did. I can do experiments to
figure out of I've made things better/worse, but I wouldn't know how
to improve the result, short of just going back to the old antenna.

* Add amplifier to camera TX unit

I think makes a lot of sense. The risk here is that involves adding
something without damaging the TX circuit. Presumably if I'm
careful I can manage this task.

* Unify power supply with radio RX

This is already done on the current version. The camera TX battery
box is 6V, so I put a tap on that to run the servo RX

* add DC-to-DC power regulator

This would replace the probably very shitty DC-to-DC regulator
included in the camera TX box. Probably necessary if I add the
amplifier to the camera TX -- the existing regulator likely doesn't
have enough power. This is kind of an annoying job -- a handful of
little components. There are plans on the web for it though, and it
should be manageable.

* replace XCam2 CCD with high quality CCD

At least one guy recommends a $100 Panasonic CCD -- apparently the
CCD has much faster response time (so the camera won't get blinded
by changes in orientation to the sun) as well as a better lens. The
XCam2 lens has huge amounts of distortion. I'm not sure about this
mod -- I think it might be in the category of "getting more
expensive faster than it is getting better." I'll think about it
some more.

* camera RX power supply

The current camera RX power supply is an 8-pack of AA batteries.
I'm not sure that these necessarily provide enough current for it.
Something beefier would probably be good, although I don't really
know what to do here. Could just go with C batteries instead of AA
batteries. Could introduce a 12V nicad, which would need a charger.

* camera RX antenna

The RX antenna probably needs as much improvement as the TX antenna,
and since it doesn't have to fly we can do something more
aggressive. There were plans for some sort of antenna that's
supposed to work much better. One issue is I think the result is
more directional. This could be okay, or it could be a pain.
Presumably the kites macro-scale direction changes pretty slowly,
assuming the camera man is standing in the same area as the flyer.

* ground unit display

It would be nice to have a bigger ground-unit display than the
CamCorder's display/viewfinder. Issues here are that it would need
to still be visible in bright sunlight, which generally seems to be
the kiss of death for LCDs. Maybe a small CRT?

* more kites

We very rarely have trouble with too much wind. A kite that's
compatible with low winds and docile, etc. would be a useful
investment. I think the thing here is a parafoil soft kite, but I'm
not sure.

Looking over the list, I think all of these make sense to do, with the
possible exception of the servo-city gearhead fo the pan-axis, and the
replacement CCD. I'll start putting together a list of materials.

Jalan Kayu Bike Trail

So I kept hearing about some trail that goes from Jalan Kayu to Seletar Reservoir. A few years ago, when the area was under construction, I walked along the mosquito infested street and remembered seeing a very distinct trail that led off into the woods. I figured this was the trail everyone referred to.

Well, it sure as hell wasn't, or, if it was, it is impenetrable right now. I rode fifty feet into the forest before I saw a shabby encampment of plastic chairs, tarpaulins, and various rubbish. There were no people there, but I stirred up a hornet's nest of about a dozen+ wild dogs. And these were not sheepish mongrels, either... at least one had the chest color of a brown dobermann pinscher, although it was clearly bred with some other nasty creature. So ended that exploration.

(FYI: the best way I've found to deal with dogs like that is to aggressively bark back at them, and act like you want to attack them! I hopped off my bike, held it in front of me, and growled and moved one step towards them (they flinch back) and then discretely three steps back, keeping up the act till I figure I am outside their threat perimeter, and then get the hell away!)

I rode around some more, including up Seletar Farmway West 4 to a dead end that showed signs of maybe being a trail once, but it needed more bushwalking than I was in the mood for (I was looking for a nice, brisk MTB ride--not parang hacking) and it seemed like every fish farm I passed had another three mangy dogs that wanted to chase me.

I surveyed to the end of Jalan Kayu, which ends at Seletar Air Base, looking for other places this trailhead might be. I never found anything that looked like a candidate. I did, however, find an easy trail that ran alongside the SLE for maybe two miles before it runs into a canal. Sort of neat, although short. I think it might be possible to take some other paths along this canal, and I'll be researching that sometime soon.

Anyone know more about this trail?

Using Oziexplorer

I started playing around with OziExplorer and my newly-repaired Garmin Etrex Vista this evening. I went on a bike ride through some less-manicured areas of Singapore yesterday and wanted to work out a trail map.

Since there wasn't heavy jungle canopy, the GPS worked fine and I got a good track and set pertinent waypoints.

With only a few moments of twiddling (well-guided by the OziExplorer help file) I had downloaded the tracks and waypoints from my Garmin into OziExplorer.

I don't have a base map of Singapore (where can I find a digital base map of Singapore?) so I just called OziExplorer to create a blank map. Simple.

[My user experience so far suggests that OziExplorer was written by a GPS-Navigation enthusiast who added functions as he thought, "Gee, it would be useful to have function X." So far he's thought of everything that I want to do. The downer is, it's all organized in a somewhat non-traditional menuing and icon structure. It's idiosyncratic, but I am sure I can get used to it. (The icons are real eyesores, though -- way, way too busy) ]

I pulled up the base map and the map size was way too big. It had world map Lat/Long proportions. I started throwing away a few left-over waypoints from Canada, USA, Taiwan and Europe and reduced the map info to just local Singapore geodata. Then used the button called, "Rescale Map." It pruned the size, but still the few waypoints and tracks I had (a meandering, short 9 mile bike ride that probably was 3 miles diameter max) were still too small in scale to the map. They were piled right on top of each other.

Repeated rescans wouldn't make it any smaller, and my zoom was maxed-out at 750%. What am I supposed to do?

Consulting the help file, I found this troubling (?) explanation:

Why doesn't OziExplorer have zooms larger than 750%

OziExplorer is Raster software (that is it uses images for the map, even the blank map is just an image). Zooming in on a raster image does not improve accuracy. The best a raster map can be calibrated to is +or - 1 pixel, when you zoom in at say 500% (x5 times) each pixel in the map image is increased 5 times, the accuracy of a position is now +or - 5 screen pixels. There seems little point in adding more zoom levels, as the calculations cause pixel rounding the liklehood of causing positional errors (on the zoomed screen that is) increases as the zoom level increases.

I do agree the blank map zooming could be improved as an image need not be used for that but to do that I have to rewrite quite a bit of the code and am saving that for a while.

So I don't know... Am I stuck from ever displaying a 3-mile diameter set of tracks in size large enough to read comfortably? It is sounding like I'm stuck.

Let's put it this way, I would be happy to sacrifice some accuracy in an image of my mapping data if I could zoom it enough to nicely fill a 8x12" paper map. It feels like OziExplorer is assuming that I demand absolute cartographic precision or nothing else. All I really want in this case is a trail map I can refer to now and then -- not a rigorous map.

If I'm forced to, I can always just save the map to an image, and then inflate it with Photoshop, but this is such a gross solution, and I'd prefer to do all my work inside OziExplorer.

Ideas?

Thule Rack

Despite what we were told, Thule racks are sold in Singapore. In fact, Pro-RV 4x4 has a very good stock of Thule components, installs the racks for you (for free) and their prices are comparable to what you'd pay in the USA. Don't be confused by their model number system being different from the USA part numbers.

The racks are very solid. We got a roof rack setup which clamps the bike, not the fork. Thus no need to remove the wheels. If I yank on the bike, it makes the car wiggle.

A few observations:

1) Lowering the front suspension of our Trek Liquid 20's alters the frame geometry enough that the rack gets a good, firm grip on it. At max-suspension, the geometry is a bit weird and it doesn't sit as nicely on the rack.

2) The leashes that anchor the wheels to the rack are not quite long enough. I have an especially high combination of tire thickness and rim thickness. Consequently, I can only get the leash to snap into the last detente. I'd feel more comfortable if there was some more slack.

I've submitted this #2 question to Thule... We'll see what feedback they give.

Serious as a Heart Attack

So a few weekends ago, while Ling was away in Laos, I bought two 2003 Trek Liquid 20 mountain bikes. These are very high-end bicycles -- heavier, more ruggish versions of the Trek Fuel, a cross-country racing mountain bike.

Technologically, the bikes are like lightweight motorcylces, having both front and back suspension systems, adjustable up to 5" of travel and rebound control over the action of the shock system. The frame is made of special alloys and carbon fibers. Most notoriously, it has disc brakes, not rim/caliper brakes like a traditional bicycle.

I had no idea how problematic this braking system can turn out to be. Admittedly, the brakes are strong, very smooth, and give a very nice, linear response on downhills. They should perform better in wet, muddy conditions also. But what I didn't realize when I was busy dragging the bikes around Singapore was how easy it is to knock these things out of precision.

So how do you load a bike in your car? You take the wheels off, throw it in the hatchback, and you're on your way.

With disc brakes, however, you take the wheels off, and if you fuck up, you accidentally squeeze a brake handle. When you do that, the automatic adjustment feature of the brake goes haywire because the disc is missing, and the piston pushes itself out the whole way. You'll barely be able to get the wheel back in between the brake pads, and even if you do, the friction is enormous.

So what do you do? You check out Hayes instructions on their hydraulic disc brakes and see their explanation in how to push the pistons back into the bore with a box wrench, reinstalling the wheel, and allowing the brakes to reset properly.

Then you try to do it yourself. And it doesn't work at all. You can almost push the pistons back in flush, but not quite, so there is still friction.

Why is it so hard to get them back in? Probably because you flipped the bike upside on its seat and handlebar to do the work. In doing so you allowed air to leak from the top of the brake fluid reservoir back into the line. Now the brakes need bled.

If you've lucked out and not poisoned the line, you still probably have reinstalled the wheel slightly out of perfect alignment, so the micro tolerances are off, and you're going to have pulsating drag.

So two weeks later, Ling still hasn't got a chance to ride her bike. The first weekend she was sick, and last weekend her brakes were spoiled. It took me several days of frustrated tinkering before I had the shop sort out the problem.

They urged in the strongest possible terms: keep the wheels on, and don't turn the bike upside down. Makes it great for travelling... both car and plane. Not only are the brakes at risk, but oil drains to bad places in the front Psylo shock if it goes upside down.

Hate to be learning this stuff the hard way. The literature that accompanies the bike is horribly sparse boilerplate. It would be nice if I could get some serious technical instruction on the components. Even generic bicycle maintenance books aren't sufficient. It's like a need a 'shop guide' to the Liquid 20! That and probably some specialty tools. Even my bike shop was delayed in tuning the bike because they lacked a peculiarly small wrench fitted for the brakes.

So I'll go collect the bike from repair tomorrow. We'll have fun playing with them this weekend. They sound bad, and they are, but also bad in a good, thuggish way. They're horribly capable. In climbing terms, riding a Liquid 20 is like a pair of shoes magically allowing a 5.8 climber to lead 5.11. The improvement is that noticeable. And since the bike is so goddamned expensive, I won't be replacing it for a long time, so I will have plenty of time to become familiar with its its systems, their repair, and maintenance. So they'll work out fine, it's just that this was my first slap of reality in owning high-end gear.