1. Meeting with Paul Ganney (Head of Scientific Computing) and James Moggridge at UCLH
    • As suggested by Gareth Askey, we met with Paul and James to discuss our requirements and identify a computer for installing our uploader software
    • Anna David participated as well
    • The best option seems to have our software running on a virtual server within the Fetal Medicine Unit (FMU). James will liaise with the relevant people to arrange for this.
  2. Initialised GIFT-Proto
    • Currently based purely on NiftyView
    • The aim is to have it building against a pre-build NifTK
  3. GIFT-Cloud extension of XNAT
    • First development iteration complete, with a modified schema allowing for storing subject pseudonyms and query-retrieving them.
    • Unit/integration tests for introduced components in the codebase
    • System tests with Ruby for the relevant REST calls
    • This version to be tagged GC-1.1.0
  4. Requested a virtual machine for GIFT-Cloud, as well as an initial project store of 2TB for storing the patient data
  5. GIFT-Cloud uploader application
    • Started merging the relevant portions of the DicomCleaner and XNAT uploader applet codebases
  6. IGI journal club presentations on software
    • Different aspects of software design and engineering presented, as one aim of GIFT-Surg is to develop high quality, state-of-the-art software for clinical use.
    • Tom: clinical software design
      • ways to think about user experience (UX)
      • how we can improve UX using simple design concepts
      • why and how software design differs between research and industry
      • how good visual design fits well with good software architecture.
    • Dzhoshkun: benchmarks comparing various aspects of different programming languages
      • popular programming languages chosen: C++, Java, Python, Ruby
      • graphs and measurements shown, related to different benchmarks
      • different benchmarks chosen to reflect different aspects of software such as runtime, code size, resource requirements, etc.
  7. Continuous integration (CI) for software development at CMIC
    • CI is important for maintaining large-scale software, developed and used by multiple users
    • It helps detect faults at an early stage
    • It also helps ensure that newly introduced components do not break the existing stable infrastructure
    • Time invested into setting up CI components on the distributed architecture that supports NifTK development
Dzhoshkun Shakir, Tom Doel: Mar 2015

Leave a Reply

Your email address will not be published. Required fields are marked *