Progress
So! Train! 3:40 hours ... used on writing my cycling monthly column, talking about the Tour de France (I need some scape also) and adding new things from the paper to my talk.
Afternoon discussing and analyzing and discussing again the differences on the parameters from postage stamps with 0.03 "/pixel and 0.05 "/pixel. And also priorities!
Got a new version of our library ... good and a nice discussion with a friend that's a programer about how to structure the new code.
At home, re-structuring the talk and checking if it fits the given time. Tomorrow I'll do it again.
Anniversary post
Three years ago I started this science blog as a suggestion from David Hogg in a conference in Marseille. The first post "
And so it begins", on a Babylon 5 style the 27th of June 2005.
This blog has helped me in several ways. Recollecting the work of the day, checking when did I do this or that, checking what did I do that day, and showing me, after sometime that I could have been better, much better. That's what I'm into, making things better in a new life, with new motivations.
Enough with "there was a crisis" and so on ...
Last days were busy and complicated outside of science and also in science!
3 projects at once, they all should be done at the same time (due to my "running after the lost time" scheme ...). No excuses, "just do it" (Nike should pay me for that hehehehe).
Photometric table for a "side project" and resolution comparison.
The comparison is how "CAS" parameters change for the same objects on ACS 0.05 and 0.03 pixel size images. That's important for this galaxy structure paper. We also have to be sure on how different filters (F606W and F814W) affect the results (unfortunately part of the data is on one filter, part in the other ...).
Also Elif Kutdemir's paper got the final acceptation this week. Cool! Congratulations for the first paper. (to me also, I'm also in the paper)
Coding
Solved a couple of coding problems on a routine for our spectroscopy and my office-mate Joerg Dietrich finished the first version of an important library for our project ...
Late for a few days ...
Few crazy days without posting. Crazy due to scientific and non-scientific reasons.
Thursday and friday were used to finish re-reducing the new FORS images, now correctly, generating the photometric catalogue in all bands, matching with the ACS photometry and selecting the postage stamps to do the non-parametric analysis and run GALFIT multimode on them. Also some interesting discussion with my office-mate Joerg Dietrich, this will lead to nice things!! Days were interrupted by dinner with friends and soccer (Portugal X Germany), on thursday and trip HOME on friday.
Nice talk by Leon Koopmans where they use a joint analysis of lensing and IFU to study galaxy structure and dynamics, also being "able" to find dark substructures predicted by LCDM simulations! Pretty cool.
Another important activity that took some time those days and my trip back home was the paper proof. The main questions where almost none, but the large amount of small changes they do on the text, unmarked and that you MUST check is amazing. Most of them were great, but some are inconsistent. Like putting dash separated words in part of the text and letting them "undashed" in other parts of the text, or changing all the angular units to "full writing" (e.g. " by arcsec ...) and changing "degree" by "^o" ... anyway. Done.
Monday was a shorter day, solving a couple of stuff back home and being "father on duty". Also trying to solve a problem on a small code I'm writing for our spectral analysis.
Trying to generate a fits table as output. The output can only be properly read again on IDL ... after checking and re-checking I found out there's nothing wrong with mwrfits, but with the way my IDL structures are being generated.
The structures I create are showing on "help" as struct ARRAY[1] and "help, /str" is showing then ARRAY[real number of lines] for each tag of my structure. That, when saved on a table and read somewhere else shows all the columns (one for each tag) but only the very first element of it.
Structures from any fits table I read, show on "help" as struct ARRAY[real number of lines] and "help, /str" shows the first element of each tag, but tables can be read anywhere ...
Any hint?????
Images and Proofs
The small problem in the images became a big problem since my processing skipped one filter name and that did a whole cascade of wrong associations NAME - FILTER.
Re-started from scratch and things seem to fall into a much better place now, amazing, isn't it?? hauahuaauhauh
This processing is gonna last until tomorrow.
Nice informal discussion from Enrichetta Iodice about polar disk/ring galaxies. Basically we learned that the polar rings are mostly polar disks and that cold accretion seems to be a plausible option. The other ones would be mergers or galaxy accretion.
Also received the proofs for the paper. The major points (marked with Q) were small things like "publisher" of a reference or more general things like "check the figures" ... Now I'm in the process of reading to see if they didn't change something that modifies the meaning of a sentence. Bodo was commenting that MNRAS once changed all the arcminutes into arcseconds ... no big deal just a factor of 60 of difference ... :-)
But that will go on tomorrow also, because I'm gonna sleep!
Thousand small things
Got the postage stamps with pixel size 0.05 for the one cluster we have pixels 0.03. Now a comparison of parameters will be made to address how this changes the values.
Nice lunch talk by Don Hall about the new generation of IR detectors, cool things ahead! Can't comment much, since solid state physics was never attractive to me, but the idea is finding a way to minimize read noise, since seems that CMOS have an intrinsic read noise and that could not be improved in the last 20 years, so new materials and approaches are necessary.
Some discussion about the FORS2 imaging and I realize something was missing, now I'm checking what was wrong. Also discussing some IDL coding with Bodo to prepare tables for sky subtraction.
Trying to figure out the best way of writing from IDL into a SDSDAS table. Anyone???
Finishing data
Train to Garching. Got the not so fast that allows you to take bikes, so that I could take one of my city-bikes to Garching. The trip was boring, no plugs for laptop (so I worked on battery) and no table, so no mouse, so no image inspection ... some coding, some object selection and long reading.
Getting to Munich, after a connection in Ausburg, the train was late and it was announced that it would stop in Pasing, not in the main station ... great, a longer subway :-/ Specially because the weather was dull and I was not willing to take the chance of rain (like yesterday while mountain biking).
By chance, got the S8, that goes to the airport and passes in Ismaning, which is about 5 km away from Garching, but at the other side of the river. And I never saw anyone with bike at the U-Bahn (I know that in the S-Bahn it is allowed), so instead of taking chances of a failed connection in Marienplatz, I took the chance on cycling from Ismaning to Garching in the rain. Fortunately it didn't rain, just a few drops, so that you can remember that Murphy exists ...
Got in the middle of the afternoon at ESO with my bike. Gave the ESO bike back!!!
Finished some points on the cluster preparations and the last multi-component modeling is running. Also retrieving some data from the archive to have it in the same pixel size as the others, so that I can check the pixel size effect in the results. Tried to resample the postage stamps, but never got convinced.
Photometry
Working out some important things in the morning. Some stuff on the street and I finally and officially gave back my office and my access card at the Institut für Astrophysic in Goettingen. Now whenever I stay in Goettingen I'll be officially working at home.
In the afternoon photometry of the last cluster. Couldn't finish it all yet. Tomorrow. And some other small things so that our collaborator has everything he needs for the parametric analysis.
More on calibrations
It's amazing how you get 3 images from the same cluster (in that case, 2 of them taken with the same instrument, but with different CCDs - old and new setup), calibrate each independently, compare the magnitudes and you get any sort of result, but the right one ... of course the night were "almost" photometric, but not really ...
Now the thing is "lighting a candle" to one of them and go ahead ...
Failed calibration intents
A day of failed calibration intents. You try the values on the website and they do not reproduce the standard stars in a simple zero point. Then you try the standard stars, and you can't get an equation that makes any sense ... great, isn't it?? Specially when you're running against time ...
More and still on clusters?
FDF and WHDF are being modeled in multi-component and were already sent to the next step of the project also.
Today was a images day. Fixed the gap problem of the THELI reduction and combined minimizing the gaps and corrected the WCS. Doesn't seem much but takes time.
On the "computer side". Trying to convert the SDSS backup from EXT2 to NTFS so that it can be properly read by the MS-SQL (still can't believe the mirrors of SDSS are meant to be used with an MS soft ... free software, please!!). That took time, and will take time. First the computer meant for that, my computer from the institute in Goettingen, reboots after trying to copy one of the 40 GB files (they are 60 ... and I need to take things out to another disk, format and bring them back).
Express install on another computer and now is copying, just hope it doesn't collapse also ... trying to put on yet another computer, let's see if it works now.
Also some reading about non-parametric morphological classification (CAS, Gini, ...) that will happen for some time also (not much).
Tomorrow: photometric calibration of the FORS2 new images, run sextractor and produce a matched catalogue. Select postage stamps, run multi-component modeling and finish this part to get to the analysis (not all today, just until the let the modeling running).
And ACTION
FDF and WHDF modeled.
Postage stamps sent with the other data to one step of the analysis. Other step running for FDF and waiting CPUs for WHDF.
Tomorrow the last clusters.
I'm really tired today ... being the "Parent in charge" is tiresome! Fernanda had the whole day of activities in her institute's review by the Max Planck Society and I was taking care of Matheus ... I'm broke! :-))))
It goes
Today finished the WCS correction of the WHDF images, catalogue, PSF, new catalogue excluding the masked objects, running the brighter 5%, should be done my tomorrow, since there was a problem and 6 hours were lost because of this issue between IDL_IDLBRIDGES and X11 connections ... always a mess. Why do we use IDL??? It's expensive and everything good they create, they find a way to prevent you from using properly.
The FDF will be ready soon. Then I can match it with the ground based photometry and start calculating parameters.
Photometry for the last clusters ... I did the pre-reduction of it with THELI, until I get a mosaic of the two FORS2 CCDs ... not the official way I suppose, but there was so many errors or things that didn't happen when I was trying to use it, that getting this mosaic was one of the bypasses I've used. Gotta write Mischa Schirmer about it. I'll compile all the problems and send him.
Anyway, in the train back home, I combined the few frames we have, trying to minimize the effect of the gap between the CCDs ... then recovered the WCS information, since the header was missing in this bypassed mosaic, adjusted it and decided to calibrate using my ACS photometry ... what a surprise ... the gap was not correctly filled, so that when the WCS is write in the upper CCD, it's wrong in the lower.
I'll try to fix easily and quickly this weekend, or monday, since I'm the "parent in charge" for the whole weekend! :-)))
Besides that, things will have to run pretty quickly since there's an important deadline related to this project approaching.
Yet another problem ...
Of course, there was a problem and GALAPAGOS didn't run during the night. The positive point on that was finding a but on the code that creates the bad objects masks. Now it's running. It may run for a while yet.
WHDF being prepared, but the WCS is a bit more complex than the FDF ... maybe the published catalogue has some wrong or imprecise values ...
Also the processing of our new FORS observations is running, now on combining and calibrating stage.
Unfortunately with all going "wrong" I couldn't go to Dimitri Gadotti's talk in the morning and Simon White's one in the afternoon ... bad.
Running
So! Today I fixed the WCS of the images, according to the FDF published data. Then you run things, prepare the bad objects masks, psf, postage stamps cut and now GALAPAGOS is running. Tomorrow will be WHDF day ...
Also today an informal discussion about infrared spectroscopy of early-type galaxies by Mariya Lyubenova. In the afternoon this months wine and cheese.
Tomorrow Dimitri Gadotti will speak about modeling 1000 galaxies with Budda (I wanna check if something useful to our 10000 galaxies with GALFIT) and in the afternoon Simon White will give the joint Munich Seminar. Another busy day ... fortunately it's raining so I don't feel bad for being to busy to bike-training.
WCS, the nightmare is back!
Long time ... I headed home on thursday because Fernanda had a strong cold. She's better now.
Friday was a day of solving bureaucratic things, like entering the request for governmental fund for Matheus day care and so on ...
Weekend tried to be as much of a dad as I could and monday was the first day of Matheus at the day care, so we were there, both of us, very worried :-)
Today back in Garching. Trying to analyze the ACS images for the FDF, and of course there are WCS problems. Matched the tiles with each other, matched them with USNO stars and the FDF published objects do not match the image ... of course it matches the FORS original image, which also matches the USNO stars ... after some trial back and forth, I'll have to start all over again ... that means one day almost lost. Also matched the WHDF tiles, of course there will be some sort of problem ... not as simples as the clusters were (of course, the clusters I'm reducing everything from scratch ...).
