Cze�� Micha�, Andrzej,

On Tue, 21 May 2002, Michal Frackowiak wrote:

Too fast have I said it works... Boud: the libraries line looks fine in makefile as:

LIBS=-lpgplot -lX11 -L"/usr/X11R6/lib"

I think this is based on what *you* have put into makemake.

I've added a few lines to makemake:

# HOW TO MODIFY: # Modify the file makemake and run it to recreate makefile. #

which will get written near the beginning of makefile.

BUT: variable LIBS is not used anymore in the makefile except this

This is why my own (created) makefile has LIBS= i.e. nothing.

moment. It should be present in all the places where ARGOPTIONS is used (or we shall add LIBS to ARGOPTIONS).

I think ARGOPTIONS is unused - it's left over from when I copied the file from the pgplot makemake file...

OK, I've added some clearer suggestions of values of PGPLOT and LIBS that could be uncommented in the file makemake.

Here's the new full section on PGPLOT from the makemake file: ############################################################## ## pgplot ##

# If you do NOT have PGPLOT, then the following two lines must # be UNcommented, and the alternative option below commented (with #).

# 2 lines to UNcomment if you do NOT have PGPLOT: # PGOPTIONS="$OBIN/pgnull.o" # PGDEPEND="$OBIN/pgnull.o"

# If you DO have PGPLOT, then the two lines below must # be UNcommented, and the alternative option above commented (with #). # # You will also need to have LIBS and PGPLOT defined # as environment variables, e.g., setenv LIBS "-lpgplot -lX11" # and setenv PGPLOT -L$PGPLOT_DIR should be in your .cshrc . # Alternatively, you could define them as shell variables # here, # e.g. # ( # PGPLOT='-L$PGPLOT_DIR' # or # PGPLOT='-L/opt/lib -L/usr/openwin/lib' # or # PGPLOT='-L/usr/X11R6/lib' # ) and # LIBS='-lpgplot -lX11' # (Warning: do NOT use spaces around the '=' sign).

# 2 lines to UNcomment if you DO have PGPLOT: PGOPTIONS="$PGPLOT $LIBS" PGDEPEND=" "

#echo "PGOPTIONS= $PGOPTIONS"

### Lines below unlikely to need modification. ### ##############################################################

http://www.astro.uni.torun.pl/sympa/shape-univ/2002-05/msg00004.html

In some free time I can dig some automake docs and see if it is worth converting to gnu configuration.

A GNU style of automake probably would be better than the PGPLOT style, as long as it works for fortran. (And after all, I ended up creating DEconfig.f which is probably not the most standard way of implementing system dependent options...)

Pozdrawiam Boud

PS: I forgot to tell everyone that Micha� proposed adding his SNe data analysis routines to DE later on when he has time! Sounds like a good idea to me :), though it will require some work...

On Wed, 22 May 2002, Michal Frackowiak wrote:

OK, that should do the trick I think.

...

PS: I forgot to tell everyone that Micha� proposed adding his SNe data analysis routines to DE later on when he has time! Sounds like a good idea to me :), though it will require some work...

... I donot think 'some' is a proper word for matching 2 different packages ;)

;)

but pros are:

• ready statistical tools for analysing data (based on maximum likehood)

which would complement the direct null hypothesis tests which will hopefully be obvious by version DE-V0.03 in DEplotcorrnall...

• snia data analyzer module

yep, nice

• cmb anisotropies data analyzis

well, to be more precise, data analysis of *some* statistics of cmb anisotropy data (e.g. the C_l spectrum), I don't imagine you're analysing cmb data (maps) itself...?

these packages combined can serve as a full cosmological package...

Well, maybe we could use a more modest term, e.g. "a more balanced observational cosmology research startup kit" ?

meantime I have an idea about wrting trmendously extra fast package generating cmb power spectrum somewhat similar to DASh but much more convinient and simpler. I estimate getting a cmb spectrum in parts of a second based on smart-compressed grid of models (seem not to have be dense). I am fed up with CMBFAST. but I will have more time only after finishing my mgr...

OK, in my way of thinking it would be good to think of the documentation including links to good introductory URLs on the subject, like Wayne Hu's pages, so that the whole package is more usable by beginners and that the whole learning/research process can become more open and accessible to students.

Boud: I think I got the grant! That's unofficial...

Great!

Pozdrawiam Boud

On Wed, 22 May 2002, Michal Frackowiak wrote:

...

...

• cmb anisotropies data analyzis

well, to be more precise, data analysis of *some* statistics of cmb anisotropy data (e.g. the C_l spectrum), I don't imagine you're analysing cmb data (maps) itself...?

sure - just from cl spectra, but my soft is now integrated with CMBFAST

OK, which AFAIK does not analyse CMB maps (but I only looked at it briefly once several years ago)...

...

OK, in my way of thinking it would be good to think of the documentation including links to good introductory URLs on the subject, like Wayne Hu's pages, so that the whole package is more usable by beginners and that the whole learning/research process can become more open and accessible to students.

I don't think newbies use f77... but thinking about gui now is a bit to early I think

For those who don't know:

gui = graphical user interface

If you know something about programming gui's, preferably in a GNU/Linux type environment, then I'd prefer that you can teach the rest of us sooner than later. (Of course, after your mgr is finished.)

I think that gui's could accelerate learning efficiency and speed by a huge amount, and this way we could hopefully decrease the stress of exams for the students who have to do exams and let them concentrate on the fun of learning itself...

It also means that *I* could finally have time to learn some of the physics that I've always wanted to learn but never had time to, thanks to using GUI's to better organise the information and the explanations...

Cheers boud