EMBOSS: Project Meeting (Mon 7th August 2006)
For XML work Alan has looked at Microsoft's VDW Virtual Desktop for HTML Authoring which is free and has a good XML editor.
Peter will check XMLspy pricing for machines at EBI and other options for home use.
Peter has a fix for a reported bug in reversed sequence positions in dreg and fuzznuc. He also has an alternative way to resolve Henrikki Almusa's changes to make sure reports are not empty, by writing a header and tail when the report is closed. The existing 4.0.0 code has the problem that a search through an entire database currently reports no hits found for each sequence. 3.0.0 ignored sequences with no hits.
Shaun is looking into Axis and JAX-WS for EBI webservices, starting with a simple service for water.
Mahmut has built SoapLab for EMBOSS 4.0.0 on a test server (emboss16). The current server is still running 2.8.0 - frozen to prevent service interfaces changing until EMBOSS command line qualifier names stabilised. However the underlying dispatcher was pointing to a release 2.9.0 installation. There is a memory leak in SoapLab, avoided for now by increasing the memory allocation for tomcat.
Tim reported extended EMBL features are working with Artemis in the modified splitter and union applications. Peter will aim to make include files for feature format definition files (Efeatures and Etags) so that Sanger can maintain a modified file with their extra source feature tags and other Artemis extensions.
Tim listed a few additions needed to make GFF version 3 feature tables fully compatible with the standard.
Alan plans to standardize the use of size_t and unsigned integers throughout the code to avoid compilation warnings.
The next meeting is on Monday 21st August.