SAB 2000-01-30: Turn off all debugging options with ./configure flag. (Do not give the -g flag to the compiler, set NDEBUG.) SAB 2000-03-27: Add an author field to MdRecord? SAB 2000-06-17: Change interpolation in absorption part to interpp. SAB 2000-08-16: - Implement abs_tags and wf_tags, also abs_tagsDefine and wf_tagsDefine. - Change absComp to return absorption matrices only for wf_tags SAB 2000-08-21: Some ideas for absorption: - Method lines_per_tgAddMirrorLines Adds mirror lines for those transitions for which the calculation frequency would be inside the cutoff. Requires the lineshapes and the frequency grid as input. This would remove the need for a special implementation of the VVW line shape, only Lorentz-type lineshape would be needed. - Method lines_per_tgRemoveOutsideCutoff Remove all lines which would not contribute anything due to lineshape frequency cutoff. Saves unnecessary calls to lineshape function, which would only return 0. Needs lineshapes and frequency grid as input. - Should it be enforced that the frequency grid is sorted? If not, this must be done by the above methods. - Make a method to set the lineshape functions. (All the rest is already there!) SAB 2000-08-21: Give a nice error message if control file does not exist! (At the moment it just says `abgebrochen') AvE 2000-08-22: Check that partition funciton coefficients are valid, some species are not defined, nevertheless a dummy partition function coefficient is given. check that the returned value of the partition function (NOT the ratio) is larger than zero after calling. AvE 04.09.2000: Agnes Perrin mentioned that the simple scheme to calculate partition functions at different temperature is accurate enough, so one should actually perfom an accuracy check. simple scheme: Q_rot(T')=Q_rot(T)*(T'/T)**1.5 for non linear molecule and Q_rot(T')=Q_rot)*(T'/T) for a linear molecule This way might be more efficient timewise. SAB 2000-09-13: Remove `1D' from WSV names. SAB 2000-09-13: Change names of automatically generated files to _auto... AvE 2000-09-27: There is optimization potential in the lineshape.cc file, read FIXME comments over there. SAB 2000-10-05: Implement a mechanism for `meta-methods', i.e., methods that can execute parts of the controlfile with changed parameters in a private copy of the workspace. SAB 2000-10-05: Implement sparse matrices and multiplication of sparse matrix with a sparse matrix, matrix, or vector. SAB 2000-10-05: Change matrix/vector package from TNT to MTL. A new version of MTL with a much better user interface is expected. However we cannot be sure exactly when it will be available. SAB 2000-10-05: Implement sensor model by sparse matrix. (Needs sparse matrix first) SAB 2000-10-05: Implement optimization algorithms. (Needs meta-methods first.) SAB 2000-10-05: Make a first release. SAB 2000-10-05: Set up ARTS mailing list. (Oliver) SAB 2000-10-05: Set up ARTS web page. (Oliver) SAB 2000-10-05: Create IDL version of AMI. (Wolfram) SAB 2000-10-05: Fix pdf version of Uguide. (Oliver) PE 2000-11-02: Change some data types as follows: Numeric -> NUMERIC size_t -> INDEX string -> STRING ARRAYofsizet -> ARRAYofINDEX ARRAYofstring -> ARRAYofSTRING Los -> LOS OneTag -> TAG TagGroups -> TAGS (What about the last suggestions. If a remember correctly, we don't have tag groups, only tags. Correct?) Note that int shall be removed as workspace variable group and be replaced with INDEX. The basic idea is that all ARTS data types are typed with upper case letters (beside of in ARRAYofXXX). We should try to keep the number of such types low, partly beacause there should be IO functions for each type. The absorption part has several special data types (Line, LineRecord, SpeciesRecord etc.). Can some of you (Axel?) look into this and change name according to the rule above and fix IO functions (only binary). Are all those data types really needed? PE 2000-11-11: Write the workspace variables in alphabetical order to wsv.txt. SAB 2000-11-08: Send information about ARTS to Gerald Nedolua for destribution to NDSC. SAB 2000-11-10: Adapt sample control file in user guide to new method and WSV names. AvE 2000-11-30: Check that assert does not have a negative impact on the calculation time if compilation flags avoid assert calls, this can be done in the lineshape.cc file, where assert is called for each lineshape function and lineshape normalization factor. SAB 2000-12-04: Make WSV groups (groups.cc) and Token types (token.h, token.cc) consistent. SAB 2000-12-04: Implement this method: md_data.push_back ( MdRecord ( NAME("raw_vmrs_1dReadFromFiles"), DESCRIPTION( "Read the individual VMR profile for each TAGS from the list of\n" "files given as keyword parameters. One file name must be specified for\n" "each TAGS. The name may include a path." ), OUTPUT( raw_vmrs_1d_ ), INPUT( tag_groups_ ), GOUTPUT( ), GINPUT( ), KEYWORDS( "filenames" ), TYPES( ARRAY_string_t ))); SAB 2000-12-09: Remove all resize calls when MTL is completely in place. SAB 2000-12-09: Include a good example file for RT. Patrick? SAB 2000-12-13: Find a solution how to remove WSVs (as if they had not been created). Best would be a *general* method (which is more than generic, since it should work for *any* WSV type). Second best would be any other way to get rid of WSVs. Add the moment this is not possible at all. :-( PE 2000-12-14: Spectroscopic data should be converted to SI units when doing the reading. No conversion should be needed in the absorption routines. PE 2000-12-14: Interpolation in m_abs is not done using the standard function (interpp). PE 2000-12-14: Do what is possible to make the absorption as fast as possible.