Eduard Westra's Blog

Thesis accepted

Yes, it took a bit, but after a little over 4 years that I started my PhD that I submitted on May 25th 2007, my thesis has been accepted and my degree is to be conferred to me in December of this year. If you want to read my thesis, visit my publications page. I have a number of B5 sized hardcopies of my thesis. If you would like to receive a copy, just let me know.

Post-doc started

Last Sunday I officially started my post-doc with Brian Schmidt. First order of business was attending the ASA at Macquarie University in Sydney. This was good fun.

Thesis submitted

Yes, today the moment has arrived... I submitted my thesis! After 3 years, 9 months and 12 days I handed in my thesis at the Australian National University for examination!

Compiling external packages of IRAF on a 64bit machine

I have been struggling with this for ages, but I would really give my experience to other people what I found out to get some external packages running on a Fedora Core 6 machine, which is 64bit. For this, I downloaded the standard Redhat binaries (note, 32bit!) from iraf.net. When you want to compile the package that has *no* Redhat binaries for it, you can create them by making the following changes. They are quite general. I've tried them for ifocas and local (the local packages we run here). The recipe is the same for all (read below for pkg the package directory name, or the package name, whatever is appropriate):

Add in pkg/lib/mkpkg.inc to every *FLAGS line the following: -m32, so it should look like:

$set XFLAGS   = "$(XFLAGS) -p ifocas -m32"
$set XVFLAGS  = "$(XVFLAGS) -p ifocas -m32"
$set LFLAGS   = "$(LFLAGS) -p ifocas -m32"

(Ifocas is the used package in this example, but you catch my drift.)

Then in both pkg/src/mkpkg and pkg/mkpkg you add -m32 to each line starting with $link, i.e.:

$link   x_focas.o libpkg.a -lc -o xx_focas.e -m32

(so, that's two files to be changed!!)

Finally, if it does not exist, create pkg/bin.redhat

Fix up the $hlib/external.pkg as usual...

Then as iraf user go into cl and type:
cd pkg$
mkpkg -p pkg

And it should compile.

If you have a normal makefile, i.e. you need to type something like make or make ALL, just add -m32 to all *FLAGS and it should compile.

Hope this helps all desperate people out there with 64bit machines. Make sure, of course, that you have always the 32bit libraries installed of everything ;)

Gaussianity of the AAOmega cameras

In the AAOmega run of March 2006, we took some standard stars. They were reduced with drcontrol by changing the type from MFFLX to MFOBJECT. In a discussion with Rob Sharp (AAO) he mentioned that it was a good idea to test the Gaussianity of the standard stars throughout the frames. The reason is that the software uses a constant value for the FWHM of the Gaussian throughout the spectral image. This could have major impact on the relative flux calibration. (Absolute flux calibration with a fibre instrument is quite near to impossible.) The results are in the two figures.

The blue camera is quite constant at a FWHM = 3.05 pix and doesn't vary more than 1% between observations. The red arm, on the other hand, varies a lot more! It varies between a FWHM of 3.5 pix (towards the blue end) down to 3.1 pix and ~2% between observations. The measurements were done with some scripts I wrote in IDL. The FWHM and central pixel of the Gaussian are derived from the sum of 51 columns around the indicated points.

IDL boundary limts...

Consider the following code in IDL:

IDL> a = dblarr(4,4)
IDL> b=[4]
IDL> a[1,4]++
% Attempt to subscript A with  is out of range.
% Execution halted at: $MAIN$
IDL> a[1,b]++
IDL> print, a
       0.0000000       0.0000000       0.0000000       0.0000000
       0.0000000       0.0000000       0.0000000       0.0000000
       0.0000000       0.0000000       0.0000000       0.0000000
       0.0000000       1.0000000       0.0000000       0.0000000
IDL> b=[6]
IDL> a[1,b]++
IDL> print, a
       0.0000000       0.0000000       0.0000000       0.0000000
       0.0000000       0.0000000       0.0000000       0.0000000
       0.0000000       0.0000000       0.0000000       0.0000000
       0.0000000       2.0000000       0.0000000       0.0000000

This is quite annoying. I thought it was a bug (or feature) from IDL, but it seems that I didn't read the manual good enough. With the option compile_opt STRICTARRSUBS this behaviour is not allowed anymore...

Mari Minari from Support at Ittvis (formerly RSI) helped me out on this. Thanks!!

AAOmega spectrum of J114334.98-014433.7 (aka S11_5236 aka S11_13368)

As part of my thesis I am looking for line emitting objects. During follow up observations of low redshift (z ~ 0.24 - 1.2) emission line galaxies at the AAT with AAOmega, we also included the confirmed Lyman alpha emitter J114334.98-014433.7 (aka S11_5236 aka S11_13368; Westra et al. 2005). Below is the spectrum of this object. The exposure time is ~4 hours. The line can be clearly seen.

The spectrum was first presented at the ASA 2006.

Nasty MySQL statements

Just found out how to do a LEFT JOIN on combined tables with UNION. It's a bit tricky, but the point is that the result of a UNION is another table... See below:

SELECT `ID`, "whatever columns" FROM (
SELECT "whatever" FROM t1
UNION DISTINCT
SELECT "whatever" FROM t2
) AS newName1
LEFT JOIN
(
SELECT `ID` FROM t3
UNION DISTINCT
SELECT `ID` FROM t4
) AS newName2
ON `newName1`.`ID` = `newName2`.`ID` WHERE `newName2`.`ID` IS NOT NULL

Welcome

Hi all,

This is my attempt to blog whatever I'm doing. Most likely it will contain nothing useful and maybe at some point everything will be just canned, but I thought it might be a good idea to start something, so I can just spit my struggles with computing somewhere ;)

Eduard