Presentation is loading. Please wait.

Presentation is loading. Please wait.

NEW APPROACH TO STREAMLINING ETRIKS USER REQUIREMENTS DOWN TO EFFECTIVE DELIVERY INTRO OF FDT-LEADS 13-jan-2014 E. Van der Stuyft, P. Rice 1.

Similar presentations


Presentation on theme: "NEW APPROACH TO STREAMLINING ETRIKS USER REQUIREMENTS DOWN TO EFFECTIVE DELIVERY INTRO OF FDT-LEADS 13-jan-2014 E. Van der Stuyft, P. Rice 1."— Presentation transcript:

1 NEW APPROACH TO STREAMLINING ETRIKS USER REQUIREMENTS DOWN TO EFFECTIVE DELIVERY INTRO OF FDT-LEADS 13-jan-2014 E. Van der Stuyft, P. Rice 1

2 Agenda Intro ◦ Current delivery package process assessed ◦ Proposed alternative approach ◦ Key process components and supportive framework Next steps to get on top of ◦ Agree on a FDT working model ◦ Agree on functional domain delineation ◦ Break up consolidated requirements into functional building blocks to distribute across FDTs Getting ready for Annual meeting and further follow up Feedback - Questions 2 Proposed focus of this meeting: Bringing the concept across. Proposed focus of next meeting: feedback+ start working

3 Current delivery package process assessed ◦ Last resource team meeting  Provided some visibility to IMI project status and needs from a user perspective  Resulted in high level user requirements bundled in consolidated features requests listconsolidated features requests list ◦ Sowtware devt-related progress on delivery packages since then  Many delivery package owners  see very little if any progress  Have no clue as to what eTRIKS-sponsored tranSMART development resources worked on  Have no clue on how to get traction on them from a delivery package owner perspective  Even after extensive efforts by some ◦ Assessing the root of the problem  Core questions as to how to effectively funnel URS -> delivery not resolved  How can we ensure overlapping needs across projects don’t lead to fragmentation and redundancy?  How do we articulate URS so developers can work on what users effectively need?  What info is needed before cost/benefit-based prioritization makes any sense?  What ‘URS -> devt -> acceptance -> rollout’- process is adapted in this space?  How can we orient different stakeholders to contribute effectively to an environment that requires delivery-oriented, focused, disciplined development?  How can we deal with a workload that is quite beyond what 6(?) eTRIKS developers can do? 3

4 Proposed alternative approach 4 IMI projects to support in eTRIKS Functional need space How to handle? URS overlapping cross-project? ◦ Project driven - overlapping code? ◦ Driven by distinct ‘Functional Domain Teams’ interacting with ‘Delivery Package Owners’ URS to code streamlining? ◦ Ad hoc betw user reps & dvlprs? ◦ Iterative process driven by FDT-user rep:  Articulation, prioritization, user acceptance FDT –dvlr rep:  Design/architecture, estimation, development, unit testing Focus and critical mass ◦ Practical arrangements on key stakeholder focus and discipline ◦ Encouraged to look for pragmatic collab formulas beyond eTRIKS  Cfr tranSMART-NGS-data working group ◦ IMI-projects with specific priorities are encouraged to integrate resources/budget in team process Supporting framework ◦ Link to initial frame Link to initial frame Data Modelling Functional domains Data curation and Loading Analysis and Visualization Browsing and Querying Core Development Long-Term Development UBIOPRED ONCOTRACK OTHER PROJECTS... ABIRISK Link to functional domain team organization

5 Key process components and supportive framework 5 Deliv Package(s) owned by DPO Functional ref doc (see example)example List of Functional Feature Modules (FFM) for X-FDT (see example)example Breakdown, estimate, prioritize Develop Send for release by CD-FDT Accept URS to FFM map owned by DPO (see example)example User scenario owned by DPO (see example)example DPO+ FDT- user reps break up user scenario in high level functions and associate each with spec FDT DPO- managed Backlog managed by FDT-user rep Managed within FDTs Regular interaction FDT-leads Regular interaction DPOs – FDT-user reps Regular interaction betw FDTs and ‘world’ Governance (Resource Team Meeting) Actively look for resourcing extension (cmmty, IMI project spec priorities,..) All of that supported via the eTRIKS Requirements Streamlining wikieTRIKS Requirements Streamlining wiki How can I work with it?

6 Agenda Intro ◦ Current delivery package process assessed ◦ Proposed alternative approach ◦ Key process components and supportive framework Next steps to get on top of ◦ Agree on a FDT working model ◦ Agree on functional domain delineation ◦ Break up consolidated requirements into functional building blocks to distribute across FDTs Getting ready for Annual meeting and further follow up Feedback - Questions 6

7 Agree on a ‘cross-FDT’ working model ◦ Need a recurring joint meeting platform for FDTs and DPOs  To sync on any relevant topics across stakeholders and across the process chain  Priority setting/resource dedication  High level functional breakup of requirements  Visibility and impact on functional drill down by DPOs  Other touchpoints between FDTs to discuss more in detail  Transparant exposure of ongoing development (eg prototype reviewing,...) ....  Reserve timeslot of 1-1.5 hr per week  With clear agenda items and sub-timeslots to gradually reserve ahead of time  So people can select what’s relevant for them  Aim for closing early if no relevant topics!  Logistics support  Meeting setup and project mgt support  With possibility to tap into off line: Minutes, presentation sharing, recording support where relevant?  (Potential for parallel breakout sessions if needed? )  Link with WP2/WP6 meeting? Also with DP meetings?  eg 1hour FDT; 30 mins support : as long as disciplined and pre-set  Relevance filter on support stuff shared by WP2/WP6 representative to FDT part?  Conclusions – when? 7

8 Agree on a ‘within-FDT’ working model ◦ Teams have considerable freedom to operate as they feel works best  Some areas to possibly think about  Resources and logistics  Agree with stakeholders on expected focus and bandwidth  I will personally commit to min 0.2 FTE and provide transparant evidence of it  For me to remain committed, I will expect at least 1 FTE of steadily dedicated developer effort  I will actively look for pragmatically collaborating with the broader community  Agree on meeting frequency and work format (eg 1 hour per week)  Priority setting  Pragmatic (vs ‘analyse paralyze’) decisionmaking  Picking universally useful base needs from ‘consolidated feature requests’ first  DPOs can create room for specific IMI project needs  By drawing in dedicated resources to operate within framework  By making adequately trained resources available  By providing budget for such resourcing  By drawing in resources from the open source community  Pragmatic modularity  Functionality with touchpoints to other funct. domains will always be discussed in a timely way with the relevant FDT  Every function we build will have a clear definition level (all ‘what’-info which the outside world needs to know), with the ‘how’ modularly wrapped inside  We will focus on interoperability, meaning base query functions will be exposed to both int and ext modules  GUI functions will always build on top of underlying base functions 8

9 Agenda Intro ◦ Current delivery package process assessed ◦ Proposed alternative approach ◦ Key process components and supportive framework Next steps to get on top of ◦ Agree on a FDT working model ◦ Agree on functional domain delineation ◦ Break up consolidated requirements into functional building blocks to distribute across FDTs Getting ready for Annual meeting and further follow up Feedback - Questions 9

10 Proposed ‘querying/browsing FD’ delineation ◦ Querying/browsing Functional Domain (FD)  In charge of bringing right data together intelligently  Data as modelled into tranSMART by data modelling FD  And as uploaded into tranSMART by data curation and loading FD  Output data as fed back by analysis functions for cascaded reuse  Doing so by means of two key resources  Dynamic querying capabilities  Bridging the full landscape, incl subject types, phenotypical (lv1), difft molec profiling data types (lv 2) And drill down to file types (lv3) – wherever TM data model allows  The grid view and associated lists generation and mgt capabilities  In this way providing support to analysis/visualization FD  For analysis/visualization functionality integrated into tranSMART  Pragmatic API-based data input and output-for-reuse support (focus on interoperability)  For analysis/visualization functionality outside of tranSMART  Support for file-based data export  Potentially via dynamic queries, including ‘drill down into file level’ queries (query level 3)  Potentially starting from a selection of what’s available in the grid view  Support for handing down selected files to associated viewing/analysis applications via grid view  In touch with core devt FD for deployment of created capabilities  Referring to long-term devt FD on features requiring fundamental changes 10 Proposed initial focus of iterative development Depending on focus of DAV-FDT, agreements could be made as to where GUI capabilities are handled. This underlines the criticality of a modular approach!

11 Agenda Intro ◦ Current delivery package process assessed ◦ Proposed alternative approach ◦ Key process components and supportive framework Next steps to get on top of ◦ Agree on a FDT working model ◦ Agree on functional domain delineation ◦ Break up consolidated requirements into functional building blocks to distribute across FDTs Getting ready for Annual meeting and further follow up Feedback - Questions 11

12 LTD-FDT? Querying variation patterns in the large diversity of molecular building blocks – a functional ref doc example Defining a limited number of models to capture at a summary level ◦ The identity of a specific molecular building block ◦ Core features of interest about that molecular building block In such a way that  A broad range of data types can be flexibly uploaded into a fitting model  Versatility and performance on querying is adequately facilitated for Providing a querying framework In analogy with cohort select query fcty at phenotypical level ◦ To reach exactly into the molecular building blocks of interest  from the overall picture via a concept tree-like paradigm  Iteratively extending reach into new data models are these are being added by DM-FDT ◦ To apply the appropriate filters on features of these building blocks ◦ To combine the above components with adequate boolean logic ◦ To pass on relevant query results to the cohort selection list ◦ To save query definitions for later reuse 12 DM-FDT BQ-FDT DCL-FDT DM-FDT?

13 Agenda Intro ◦ Current delivery package process assessed ◦ Proposed alternative approach ◦ Key process components and supportive framework Next steps to get on top of ◦ Agree on a FDT working model ◦ Agree on functional domain delineation ◦ Break up consolidated requirements into functional building blocks to distribute across FDTs Getting ready for Annual meeting and further follow up Feedback - Questions 13

14 Proposed resource team mtg activities ◦ Alternative approach explained to stakeholders (DPOs, Developers,...) (1 hour-allowing for interaction)  Adapted presentation  What wiki resources are available to DPOs and other stakeholders?  Who are driving FDTs and how are these FDTs delineated?  Prioritization rationale and how DPOs can affect what effectively gets done ◦ Extensive joint worksession with DPOs and FDT-leads (as much time as we can get Avoid pushback by other item running out; plus actively shift from ‘items that just fill up time’ to ‘pushing forward’?)  To translate as many DPs into Functional Ref doc as possible  Will ask FDT-user reps to prepare for that (volunteering approach) (link to consolidated IMI project user requirements overview )consolidated IMI project user requirements overview  This exercise will also allow people to step into the proposed mindset 14

15 Homework Now: decide on main recurring meeting Next meeting before annual meeting ◦ Key feedback to sync on ◦ Functional Domain delineation and pragmatic priorities Before Annual meeting ◦ Prep functional reference doc facilitation for areas (see list) ◦ Get set and started as a team From there: ◦ Work down into Functional Feature Mapping per FDT ◦ First for priority topics to get started on  get developers started soon 15

16 Prep 4 funct ref doc homework Automated Data Checking post ETL Fabien, Yannis Automated Data Checking post ETL Features identified as foundational to support OncoTrack Emmanuel/Chris Features identified as foundational to support OncoTrack eTRIKS Export Emmanuel, Jay eTRIKS Export Custom annotation files ?ETL? Custom annotation files Multiple Cohorts Mansoor/Chris Multiple Cohorts eTRIKS Security Jay eTRIKS Security Support for longitudinal studies Chris/Anthony Support for longitudinal studies Reproducible Research Datasets (versioning / provenance) Mansoor, Chris Reproducible Research Datasets (versioning / provenance) GUI Mansoor Fail safe data loading / error handling Fabien, Yannis Fail safe data loading / error handling Statistical Test Selection Mansoor, Yannis Statistical Test Selection 16 Project Workspace Mansoor, Anthony Project Workspace Automated Hypothesis Generation Mansoor, Jay Automated Hypothesis Generation Gene Signature/List (Automated Gene mapping and annotation) Emmanuel, Fabien Gene Signature/List (Automated Gene mapping and annotation) Performance (especially ETL) Anthony/Fabien/Chris Performance (especially ETL) Quickly accessing end point (leaves) Emmanuel/Mansoor, Anthony Quickly accessing end point (leaves) Easy box plot creation/manipulation Mansoor, David Peyruc Easy box plot creation/manipulation Date isn't a date Jay, Robin Munro Date isn't a date Robust R interface Mansoor/Emmanuel, Jay Robust R interface Imaging Support Emmanuel, Yannis Imaging Support consolidated environment for TM/eTRIKS Jay, Peter consolidated environment for TM/eTRIKS Zoom in and out on genomic data Mansoor, Peter Zoom in and out on genomic data

17 Agenda Intro ◦ Current delivery package process assessed ◦ Proposed alternative approach ◦ Key process components and supportive framework Next steps to get on top of ◦ Agree on a FDT working model ◦ Agree on functional domain delineation ◦ Break up consolidated requirements into functional building blocks to distribute across FDTs Getting ready for Annual meeting and further follow up Feedback - Questions 17


Download ppt "NEW APPROACH TO STREAMLINING ETRIKS USER REQUIREMENTS DOWN TO EFFECTIVE DELIVERY INTRO OF FDT-LEADS 13-jan-2014 E. Van der Stuyft, P. Rice 1."

Similar presentations


Ads by Google