EMBOSS: Project Meeting (Apr 19th 1999)
Peter demonstrated EMBOSS to the Bioinformatics course in Beijing on April 10th, the first live demonstration of the project to an outside audience.
Alan has increased the default buffer size in ajfile to avoid warning messages. The warnings can stay for now to watch for cases where the buffer is being increased, but can later be commented out.
Alan asked for functions that make a new string, rather than a reference counted copy so that it can be safely edited. Currently there is only ajStrMod which makes a string single-copy. ajStrCopy was proposed as the name for a new function that does this automatically.
Sequence names written to output files need an automatically generated USA so that they can be read again if the file is used as input. There are some USAs that are "read once" which need a default, perhaps a dummy database name of "UNKNOWN:". Examples are reading from applications or from stdin.
HGMP now have purify installed for testing EMBOSS memory leaks.
Gary asked for a new field in the ACD application line, "groups:", to hold a list of functional groups. Peter will add this.
Some ACD files have fields defined that now have good defaults provided automatically. Peter will document the current recommendations in a mail to the emboss-dev list.
Peter will review the course book and list the new applications functions needed.
For applications that use a sequence range there is a need for a function to trim the sequence and return an offset so that numbering is correct. Peter will write one.