EMBOSS: Project Meeting (Feb 23rd 1999)


Sanger Centre: Peter Rice, Ian Longden, Richard Bruskiewicz
HGMP: Gary Williams, Alan Bleasby, Sinead O'Leary, Thon de Boer, Mark Faller
Other: Tim Littlejohn, Encompass Bioinformatics and ANGIS (the Australian EMBnet node)
(Apologies: Rodrigo Lopez)

1. Matters Arising

Tim Littlejohn was visiting Hinxton. The meeting was rescheduled so that he could join us.

2. Progress on Release 0.0.4

We would like to produce a stable release for developers as soon as possible. The main requirements are corrections to the make procedures and additional documentation.

CVS access for HGMP is set up. Secure shell 2 is required.

Alan will merge the HGMP updates with the master CVS copy at Sanger.

SGI systems have given some problems with compiler versions. These are probably resolved, and will be tested after the meeting.

Alan has added a set of system functions as "ajsys" in AJAX.

A data file directory and search path is needed. Alan proposed searching:

Peter proposed changing the use of EMBOSS variables. These are environment variables (in upper case) with the prefix "EMBOSS_" and local definitions in lower case with the prefix "emboss_". To make code simpler, and to make AJAX more general, code will only specify the remainder of the name, and the prefix will be added automatically. Prefix "emboss_" is already provided by the "embInit" call in all applications. This will break existing code, so a check for "emboss_" as a prefix, and a warning message, will be put in place when this is first implemented.

3. ACD extensions

Peter has implemented simple functions in ACD files. They all have the syntax "@(...)" where the "..." can be a simple arithmetic expression, a C style boolean test with two results, and a switch statement with a list of tokens and results.

Mark requested greater than and less than comparisons (and equality) returning boolean results.

Peter will try to implement "@()" functions recursively. The parsing mechanism could make this trivial.

Gary proposed a required "doc:" field for the application. This should produce a warning if not provided.

A "-options" qualifier is proving useful in many applications. Peter will make this a standard for all applications. Some qualifiers could then become "required" if this is set, probably by an alternative to "required". These would only generate prompts to the user if "-options" was on the command line.

A new option "-filter" is added for all applications. This forces "-auto" so user prompts are turned off, and the first input and output files default to "stdin" and "stdout" respectively.

Output filenames are now generated from the "name:" and "extension:" fields in the ACD definitions, with defaults of "outfile" for the name and the sequence format (for sequences) or program name (for text files) for the extension, or "out2", "out2", etc. for further output files. Peter recommends always defining a value for "extension:" for text files.

4. Documentation

All functions and data types in the source code have formatted headers. Peter has a script to convert this into SRS databases EFUNC and EDATA. This will be added to a new "scripts" directory in the EMBOSS distribution.

Peter will write an introduction to the libraries for new developers.

Thon's ACD guide has been updated. It will be converted to HTML and added to the Web pages. Everyone is encouraged to contribute documentation.

5. New applications

The applications proposed at last week's meeting are listed on the Applications web pages.

These need a draft document for each application. Potential authors should then generate ACD files and samples of input and output plus any other information needed to make a detailed program specification.

After testing the ACD files, development can begin with the highest priority applications.

6. Web Pages

Peter will add the Applications, Documentation and Minutes pages to the main EMBOSS web page.

7. Any other Business

8. Next meeting

Next meeting Monday 1st March, usual time and place.
Peter Rice, Informatics Division, The Sanger Centre, Wellcome Trust Genome Campus, Hinxton Hall, Cambridge, CB10 1SA, UK.