EMBOSS: Project Meeting (Mon 1st Aug 11)

Attendees

EBI: Peter Rice, Jon Ison, Michael Schuster
Visitors:
Apologies: Alan Bleasby, Mahmut Uludag,

1. Minutes of the last meeting

Minutes of the meeting of 25th July 2011 are here.

2. Maintenance etc.

2.1 Applications

Peter is testing the impact of restricting qualifiers to unique names that are not a short form of another qualifier. It appears that the code only compares to earlier qualifiers in the ACD file. The tests will be run again with a comparison against all qualifiers.

Peter reported that a user has offered to help update the isoelectric point calculations by requesting data values from ExPASy and will try updating the code.

Jon is updating the one-line descriptions in ACD files to make the grammar consistent.

Alan has reported fixing the tcode scoring bug which was caused by a missing line the the data file.

2.2 Libraries

Peter fixed an issue with using PCRE regular expressions to extract very long translations in GFF format. The regular expression hung (or took a very long time) when checking for a long quoted string with possible escaped quotes. The revised code first replaces escaped quotes with another character, then extracts the string and if necessary reinserts the escaped quotes.

Peter and Jon have revised the generated library documentation. Redundant lines have been removed, as have lines which only repeat a standard text. The lists of datatypes, sections and functions have been made into bulleted lists which are easier to read. For each function, the initial release number has been added by inspecting past releases.

Peter will review the book sections on handling datatypes to see how much of them could be automatically generated.

Michael noted that the functions for memory checking are called ajMemCheck and ajMemProbe but the configuration test is for HAVE_MCHECK. Peter will check on this and consider renaming the option, but as the GNU library function is called mcheck it may remain unchanged.

Michael noted that the symbols and enumerations in libraries are not documented and some globally available enumerations could be renamed for consistency and to avoid possible clashed with non-EMBOSS names.

2.3 Configuration

Michael is trying to clean up the main configure file.

The Makefile for the AJAX ensembl library has been cleaned up to use only direct dependencies. Alan is checking the modified files. The aim is to have library links only where they are needed, for example PostGres and MySQL are currently only used in the ajax/core library and are not needed in the CFLAGS for other libraries.

The X11 library location test for MacOSX is probably no longer needed.

2.4 Other

Peter has split QA tests so that each EMBASSY package has its own test/qatest.dat file. The qatest.pl script searches for all qatest.dat files. In EMBASSY packages where there are files generated by another package for use in a test, the other tests are imported under new names with new "AX" and "AY" line types for the application and package name.

Peter has been replying to discussions about licensing of data files on the EMBOSS and Debian mailing lists. DebianMed developers are concerned that swissprot example files have a license that forbids changing Swissprot without permission. Peter pointed out that while this is not a serious issue, DebianMed need to consider how they handle the licensing of GO and other data files. Other Open-Bio projects have pointed out similar issues.

Michael noted that the licensing information in the source files still refers to the old name of "GNU Library GPL". All files should be updated to "Lesser" which is now the official name.

Header files should also have the license information as this is often the first place developers will look.

We should also consider an @source documentation block in header files.

3. New developments

3.1 Reference sequences

3.2 Support for interfaces and GUIs

3.3 Data access

4. Administration

5. Documentation and Training

5.1 Web server

Peter would like to replace the home page with a new style that better indicated the documentation, news, mailing lists and other information. BioPython's home page is a good example, but is wiki based.

5.2 Books

Jon has updated the user manual with corrections from the original proofs.

Peter has updated the version numbers and some examples in all 3 books so that the web pages and original XML match release 6.4.0.

For the book text Peter suggested we should use names for the pages rather than chapter and section so that we can link permanently to, for example, functions handling translations.

6. User queries and answers

All done.

Peter is working through the remaining items on the bug tracker list. Most were already resolved. None were serious bugs. Less than 20 items remain.

7. AOB

None.

8. Date Of Next Meeting

The next EMBOSS meeting will be on Monday 8th August.