From P.J.Briggs@dl.ac.uk Wed Jul 6 16:00:08 2005 Date: Wed, 6 Jul 2005 15:58:44 +0100 From: P.J.Briggs To: Tadeusz Skarzynski Cc: Airlie McCoy , Alexei Vagin , C.C.Ballard@dl.ac.uk, Charlie Bond , Dr Daniel J Rolfe , Paul Emsley , e.dodson@ysbl.york.ac.uk, fei@ysbl.york.ac.uk, Garib Murshudov , Graeme Winter , Keith Wilson , Maria Turkenburg , Martyn Winn , naismith@st-and.ac.uk, Norman Stein , Randy Read , Ronan Keegan , Wanjuan Yang , Phil Evans , Peter Briggs , Maeri Howard-Eales , Martin Noble Subject: Response to Minutes from the CCP4 automation meeting Dear Tadeusz and others Thanks for including me on this list and making the summary available, it makes for interesting reading and I'm sorry to have missed the meeting itself. Prompted by the notes, I think it would be constructive for me to comment on two areas: 1. CCP4(i) Database a) It seems to me that the notes reasonably reflect the current status of the work that is being done by Wendy as part of the BIOXHIT project. For those who would like more detail, there is a password-protected area for the database work which may be of interest: http://www.ccp4.ac.uk/projects/bioxhit/ (username bioxhit, password is the same but reversed). There are a number of documents in that area, however the most relevant is the "Project Handler Description" document: http://www.ccp4.ac.uk/projects/bioxhit/documents/ProjectHandlerDesc.html Amongst other things this outlines the 3 broad "classes" of data (working data, history, knowledge base). b) We are not committed to any one type of database backend implementation. We plan to support different types (relational db, CCP4i def files, others to be decided). An API layer will hide the implementation details from the applications. The choice of backend may be influenced by issues such as ease of installation. c) I have talked to Geoff Battye about integration/compatibility with the new Mosflm GUI with the CCP4i db after the March meeting. He is interested in doing this, however at the time was more focused on working on the interface itself, so the issue still needs to be resolved. 2. Future plans for CCP4i a) CCP4i needs to change from its current "batch mode" model to accommodate more interactivity, in order to cope with the (anticipated) demands of automated applications. (Already Steven Ness is bending the limits of what is possible within the current CCP4i framework in his most recent work on CRANK.) It also needs to present the output from jobs in a smarter way. At the same time it needs to continue supporting the existing way of running jobs. b) It is hoped that this expansion can be achieved through migration of the existing code base rather than requiring a full rewrite, as the amount of work required in a rewrite is not trivial and would most likely take several years. I would envisage a reimplementation of some of the core parts of CCP4i into a more modular system - and this has already begun (a substantial aim of the ongoing BIOXHIT work is to separate the project history database handling from the main part of CCP4i in order to make it accessible from non-CCP4i applications). I would also envisage an extension of the current CCP4i "toolkit" to write interfaces more appropriate for automated applications. c) There seems to be confusion over the choice of language that a CCP4 GUI should be written in. It shouldn't be necessary to write the interface in the same language as your applications provided that there is clear separation between the two parts, and I imagine that Liz Potterton could provide useful input here from the MG experiences. d) I would tentatively suggest that we consider setting up a "GUI" working group in addition to those listed in the minutes. This could explore the issues and better determine what is needed from the next generation of CCP4 GUI. In my opinion CCP4 should also consider making dedicated effort available to work full time on CCP4i and its successor/reincarnation. I could add more however I feel that this not the best place for me to do this. I hope that you find these comments useful, Best wishes Peter On Mon, 20 Jun 2005 tadeusz.j.skarzynski@gsk.com wrote: > Dear All, > Please find enclosed minutes from the Automation Meeting in York (7-8 > June). I hope all participants agree it was a useful and productive > meeting, which, through its actions, will help us to transform CCP4 into > an even better suite over next months and years. > Could you please let me and Garib know about any mistakes and corrections? > We will put the final version on the (restricted) CCP4 developers site, > with available talks - I have talks by Garib (vision, and MR), Alexei > (sfcheck), Graeme (DNA etc), Paul (PyChart) and myself (intro). Could the > remaining speakers send me theirs? Thanks. > With best regards, > Tadeusz > PS. Thanks to Eleanor for taking detailed notes! I have edited them a > little, so any mistakes are most likely mine... > > > > > -- _____________________________________________________ Peter J Briggs, pjx@ccp4.ac.uk Tel: +44 1925 603826 CCP4, ccp4@ccp4.ac.uk Fax: +44 1925 603825 http://www.ccp4.ac.uk/ Daresbury Laboratory, Daresbury, Warrington WA4 4AD