SVG has a bunch of properties for laying out text. Two really handy ones are “dominant-baseline” and “baseline-shift”, which together let you move text up and down a bit. “dominant-baseline: central” seems particularly useful because it lets you lay out text where 0, 0 is the middle of the text, not the baseline underneath the letters. Ie: visually center things vertically.

Unfortunately those properties don’t work in many SVG implementations. They work in Safari and Chrome, but not in Mobile Safari, IE, or Firefox.

A workaround is to use the dy parameter on the text nodes to move them around. dy: .35em is about like centering. I don’t think this works via CSS, though, only in the DOM.

I’ve added some new code to my wind history project to show winds based on month of year. With all 12 months selected the new code should produce similar results to the old code. I don’t really have any proper tests for this stuff, though, so I’m just kinda looking at the pictures and saying “yeah, looks about right”. Here’s a before and after: do the wind roses look the same to you? (Note that I moved the location of one wind rose, San Jose. That’s a bug fix).

Update: Seth was nice enough to register the images and do a difference calculation for me. Which is awesome and sad all at once :-P

Firefox, Chrome, Safari, and the ugly duckling: IE9 in a Windows VM.

I’m adding geolocation to my windhistory project. It’s a very easy API. Some quick notes based on a call to getCurrentPosition() with a few options: { maximumAge: 1000*60*60, timeout: 1000*2, enableHighAccuracy: false}

Chrome on MacOS seems to use IP address to identify what city I’m in. It puts me at 37.775/-122.419, the middle of San Francisco, whether my WiFi is on or not.

Safari on MacOS seems to require WiFi location. It won’t geolocate at all if wifi is turned off; the Javascript API returns error -2 (no position available). With WiFi on it seems to use the same airhook style location as the iPhone. It puts me at 37.749/-122.428, which is just about where I’m sitting. MacOS has privacy System Preferences for location services. The geolocation API stops working if you disable location services entirely, but it’s not necessary to have “Safari” checked in order for Safari to geolocate you. That seems like a bug, even if Safari itself is doing some privacy protection.

Firefox 6 on MacOS also seems to use WiFi location, 37.749/-122.428. Didn’t test what it did without wifi.

MSIE 9 on Windows 7, running in a VirtualBox VM on my Mac, gives me a slightly different location of 37.748/-122.416. I’ve got no idea where it got that idea, but it’s a pretty good fix. Certainly not wifi! Maybe Microsoft has a geolocation service guessing based on external IP?

Firefox 4 on Windows 7 (in the VM) puts me at 37.775/-122.419, just like Chrome.

Bottom line: Chrome seems to use an IP database, Safari and Firefox use the Mac API, and who knows what IE is doing.

Playing with my new Mac, this is my first time writing Objective C code in 15 years, and C or C++ code in nearly 10. Some quick thoughts after finishing Your First Mac Application.

• Header files? In 2011? That require forward declarations of classes? Ugh.
• Objective C properties are handy syntactic sugar. I have no idea what “synthesize” really does, but I like the result.
• Static typing is awkward. I’m prefering dot style for properties
  self.track.volume = sender.floatValue;
  But if the type of “sender” is id that doesn’t work. I have to fall back to
  self.track.volume = [sender floatValue]
  and rely on Objective C magically not type checking method names on id. Objective C has duck typing, sorta, but not everywhere.
• Manual memory management continues to suck. Apple has papered over it in various ways, with retained properties and autorelease and now ARC. But the fact that the programmer is even aware of reference counts is a serious flaw in the language. Seriously, every other OO language for 20 years has had garbage collection. Hell, even C languages have had automatic garbage collection via a conservative collector like the Boehm collector. WTF is wrong with Apple that their development language requires manual memory management?
• Xcode 4.1 is a pretty solid IDE. I could drop into it with just a basic tutorial and understand most of what was going on quickly. The Interface Builder stuff so far seems OK, although we’ll see how it works with real code. I’m impressed with how one can do complex user interface constraints entirely graphically.

I created my first Macintosh application today, following Apple’s excellent docs. I didn’t write any code yet, just the default Cocoa application template. I’m always fascinated what the size of empty apps is; it often indicates how simple and clean the underlying architecture is.

I did “Build for Run” and stripped the binary. The compiled app is 44k. The binary is 14k, the NIB is 26k. It takes 10M of memory (80+M if you count shared).

Xcode created 13 files in my source directory: 14,000 lines of source totalling some 400k. Almost all of that is an XML bloated NIB file I’ll never look at as text. There’s 51 lines of ObjC code in 3 files. Pretty much all that is boilerplate; the important bits are

int main(int argc, char *argv[]) {
    return NSApplicationMain(argc, (const char **)argv);
}
@interface TrackMixAppDelegate : NSObject <NSApplicationDelegate> {
    NSWindow *window;
}
@property (assign) IBOutlet NSWindow *window;
- (void)applicationDidFinishLaunching:(NSNotification *)aNotification {
    // Insert code here to initialize your application
}

It’s certainly nicer than the 400+ lines of unmaintainable garbage that Visual Studio’s wizard generates doing a similar task. I’m optimistic that the NIB system will work out well. MacApp’s version of this was a fucking disaster in the OS 6 days, but NextStep Interface Builder was always pretty impressive.

I forget this about once a year. These two kinds of markup do not render equivalently:

<div>
<span></span>
</div>

<div><span></span></div>

The problem is the whitespace in the line breaks; those get rendered. In Chrome, they seem to add 1 pixel of spacing.

Doing some research on Digital Elevation Models, for making terrain maps.

SRTM

(Wikipedia, OSM Wiki) sourced from a space shuttle overflight in 2000. Comes in several flavours and sources.

Sources:

• NASA source (download link). Original data. Version 2.1 is their latest, cleaned up by NGA. The NASA has some data voids, particularly at mountains and lakes. Documentation of the dataset.
  License: Data said to be fully open for any use; don’t have a specific reference.
• CGIAR-CSI Data processed to remove voids. SRTM3 only. They publish it as “version 4.1”, version number reflects their data cleaning. Also available as an Amazon EBS volume.
  License: non-commercial with attribution.
• HydroSHEDS. SRTM data with voids filled, SRTM3 only. Also vector data on watersheds. Technical documentation.
  License: “free for non-commercial use”

• Resolution: 3 arc seconds, roughly 90m, z=10
• Format: .HGT files as SRTM, 1201×1201 per one degree square file.
• Coverage: global land mass from 60N to 56S. (Missing Antarctica, significant Northern parts of North America, Europe and Asia).  Coverage map.
• Size: 14546 files of 2.8 megs each, or 39 GB. 2:1 compression with gzip.

Viewfinder Panoramas

Eclectic collection of more detailed / patched data than SRTM in certain spots. Unclear licensing.

ASTER GDEM

Joint NASA/Japan METI project, produces a 30m DEM. (Wikipedia, OSM Wiki) Has coverage between 83N to 83S and some holes filled compared to SRTM. Some question on quality; the data comes with a “quality assessment file”. May be free to use, but some concern on the OSM Wiki that its license is not consistent.

OpenDEM

Well intentioned but small project to provide better DEM data. Some useful notes.

GMTED2010

USGS project. Global data: 90S to 83N. 7.5 arc second (30m). Looks to be a combination of SRTM data and other sources. License is presumably public domainish. The metadata page says “Metadata access constraints: No restrictions, All GMTED2010 data products and associated metadata are publically available”.

I got interested in these Fisk maps of the Mississippi, so I’m making a slippy map. Here’s what I did.

1. Downloaded the rectified TIFFs from the Army. These have already been rotated and georeferenced and are pretty high resolution. The p22* files are the maps I worked with.
2. Created pyramidal overviews for the TIFFs so I could explore them in QuantumGIS more quickly. Not strictly necessary, and in fact may be harmful: gdal2tiles seems to pick up the JPG compression.
   gdaladdo -r average –config COMPRESS_OVERVIEW JPEG –config PHOTOMETRIC_OVERVIEW YCBCR –config INTERLEAVE_OVERVIEW PIXEL -ro p22_s12.tif 2 4 8 16 32
3. Created a virtual raster layer including properly annotated sections of no data
   gdalbuildvrt -srcnodata “255 255 255” fisk.vrt *.tif
4. Generated slippy map tiles
   gdal2tiles.py -z 4-13 fisk.vrt

I’m trying to understand the consolidated.db iPhone location database, which means understanding how iOS stores and managing data. Here’s some notes.

Resources

1. @aallan’s original announcement
2. iPhone Tracker (Objective C code)
3. Technical instructions from howto.wired
4. nphonetrack (.NET code)
5. iPhoneLs.py
6. iphone-tracker.py
7. Quick start for sqlite3

.mode csv
.output wifilocation.csv
select Timestamp, Latitude, Longitude from WifiLocation order by Timestamp;
.output CellLocation.csv
select Timestamp, Latitude, Longitude from CellLocation order by Timestamp;

sqlite> .tables
CdmaCellLocation                   CellLocationCounts
CdmaCellLocationBoxes              CellLocationHarvest
CdmaCellLocationBoxes_node         CellLocationHarvestCounts
CdmaCellLocationBoxes_parent       CellLocationLocal
CdmaCellLocationBoxes_rowid        CellLocationLocalBoxes
CdmaCellLocationCounts             CellLocationLocalBoxes_node
CdmaCellLocationHarvest            CellLocationLocalBoxes_parent
CdmaCellLocationHarvestCounts      CellLocationLocalBoxes_rowid
CdmaCellLocationLocal              CellLocationLocalCounts
CdmaCellLocationLocalBoxes         CompassCalibration
CdmaCellLocationLocalBoxes_node    Fences
CdmaCellLocationLocalBoxes_parent  Location
CdmaCellLocationLocalBoxes_rowid   LocationHarvest
CdmaCellLocationLocalCounts        LocationHarvestCounts
Cell                               TableInfo
CellLocation                       Wifi
CellLocationBoxes                  WifiLocation
CellLocationBoxes_node             WifiLocationCounts
CellLocationBoxes_parent           WifiLocationHarvest
CellLocationBoxes_rowid            WifiLocationHarvestCounts

CREATE TABLE CellLocation (
  MCC INTEGER,
  MNC INTEGER,
  LAC INTEGER,
  CI INTEGER,
  Timestamp FLOAT,
  Latitude FLOAT,
  Longitude FLOAT,
  HorizontalAccuracy FLOAT,
  Altitude FLOAT,
  VerticalAccuracy FLOAT,
  Speed FLOAT,
  Course FLOAT,
  Confidence INTEGER,
PRIMARY KEY (MCC, MNC, LAC, CI));

CREATE TABLE WifiLocation (
  MAC TEXT,
  Timestamp FLOAT,
  Latitude FLOAT,
  Longitude FLOAT,
  HorizontalAccuracy FLOAT,
  Altitude FLOAT,
  VerticalAccuracy FLOAT,
  Speed FLOAT,
  Course FLOAT,
  Confidence INTEGER,
  PRIMARY KEY (MAC));