Difference between revisions of "SS P03 20JUL2015"

From PHUSE Wiki
Jump to: navigation, search
Line 34: Line 34:
 
At bottom is an ASCII version of this draft.
 
At bottom is an ASCII version of this draft.
  
 +
Is there a way to track qualification stage of a program/script?
  
 +
Will end users use Github directly?
  
 +
Mike talked about linking off phuse.eu, linking to the script index page - to download/click through.
 +
 +
Will SAS and R programmers get familiar with Github to navigate through themselves?
 +
 +
Would it be an option to do a pilot where we take a certain - one paper, and create the folders, etc. - to create a
 +
 +
Need to make a decision? Either what Gustav/Rebeka have proposed - or completely based on topics?
 +
 +
distallation of the proposal:
 +
 +
*Whitepaper topic
 +
**Adverse events
 +
***Contributed
 +
***Documentation
 +
***Qualified Programs
 +
****Output
 +
***Whitepaper
  
  

Revision as of 10:24, 20 July 2015

Attendees

Mike

Hanming

Dirk from Germany

Grace

Gustav

Jeno

Peter

Rebeka

Mike and Chris did not send an agenda prior to the meeting.

Chris will not attend; advised:

Just let Jeno roll with it and maybe get an update from Rebecca and Gustav. I will be happy to implement whatever they come up with for GitHub...

Jeno: has talked w/Chris over the last few weeks about how to store test data, and it sounds like if you do updates to a csv file in Github, it will be editable and track changes (commit, etc.) - but it might start to get complex, hard to refer back easily to earlier versions.

We think we want to store CSV files in Github - put another version there using different or different folder - notes in Github or README file.

(Discussion about how to store/version CSV files)

Gustav sent a draft he had worked on w/Rebeka.

At bottom is an ASCII version of this draft.

Is there a way to track qualification stage of a program/script?

Will end users use Github directly?

Mike talked about linking off phuse.eu, linking to the script index page - to download/click through.

Will SAS and R programmers get familiar with Github to navigate through themselves?

Would it be an option to do a pilot where we take a certain - one paper, and create the folders, etc. - to create a

Need to make a decision? Either what Gustav/Rebeka have proposed - or completely based on topics?

distallation of the proposal:

  • Whitepaper topic
    • Adverse events
      • Contributed
      • Documentation
      • Qualified Programs
        • Output
      • Whitepaper


• Whitepaper topic o Adverse events (whitepaper subtopic)  Contributed (These have not gone through the qualification process. The address the same topic as the whitepapers, but the output is not used directly in the whitepapers.) • Severity (put the programs at this folder level) o Output o Templates (We’re not sure what these are used for. We put them in with their corresponding program folder) • Toxicity o Output o Templates • MedDRA o Output o Templates  Documentation (contains YML, user guides or any other documentation for the scripts)  Qualified Programs (SAS and R programs. These have gone through the qualification process for whitepapers. We could call this qualified programs since most people are used to the term program instead of scripts) • Output (Give an example program output. We can see the display in the whitepaper, but having a snapshot of the actual programs’ outputs are helpful)  Whitepaper (Have copy of actual white papers or a link) o Demographics (whitepaper subtopic)  There are 2 folders for the contributed (DM and Demographics). Are there differences between them?  This would follow a similar structure as the adverse events.

For Contributed output, how much will our group do to the scripts? Will we actually have output? Will they go through the qualification process?