Thanks and suggestions for improvement

Discuss the jamovi platform, possible improvements, etc.
Post Reply
deepankardatta
Posts: 21
Joined: Mon Jan 29, 2018 4:30 am

Thanks and suggestions for improvement

Post by deepankardatta »

Dear all

I would like to thank the Jamovi developers for such a useful tool. I've been able to (relatively) easily convert a CRAN library to include a Jamovi module. Hopefully this will make statistical discovery a bit easier to non-experts. And hopefully Jamovi can be promoted a bit more to this target group.

I had some suggestions for improvement:
  • (1) My current version (0.8.2.2 Mac OS) fails to load the Jamovi store for a "SSL: CERTIFICATE_VERIFY_FAILED" error. I hope this was just a build oversight, and will be an easy fix.
  • (2) Could .rda files be an import option?
  • (3) Documentation - could this be a wiki instead so people could contribute? I think GitHub has a wiki function for projects
  • (4) Documentation - could there be a page with specifications to be aware of? My example is that I had to specify in my library it was built for R 3.3 (not 3.4) as the Mac version is kept at 3.3. The reason why is here: https://github.com/jamovi/jamovi/issues ... -361148721 and this is quite useful information to know
  • (5) Documentation - some documentation to describe adding jamovi functions to a pre-existing module
  • (6) As (5) but having a create() function, with a pre-existing switch
  • (7) Documentation - making it clearer how the jamovi "columns" actually are internally handled. The examples seem to indicate how Jamovi expects a function to work is function( x~y , data ) so that the formula becomes e.g t.test(formula, self$data). My function handles data as function( list1 , list2 ) so that the data is not specified. I did get around the issue by doing this:

    Code: Select all

    # read the option values into shorter variable names
    method1 <- self$options$method1
    method2 <- self$options$method2
    
    # get the data
    data <- self$data
    
    # convert to appropriate type
    data[[method1]] <- jmvcore::toNumeric(data[[method1]])
    data[[method2]] <- jmvcore::toNumeric(data[[method2]])
    
    # calculate the results
    results <- blandr.statistics( data[[method1]] , data[[method2]] )
    
    so that I called the columns by data[[column_name]].
Thanks to all involved again!
User avatar
jonathon
Posts: 2609
Joined: Fri Jan 27, 2017 10:04 am

Re: Thanks and suggestions for improvement

Post by jonathon »

hi, starting with 1)

what version of macOS are you using?

jonathon
deepankardatta
Posts: 21
Joined: Mon Jan 29, 2018 4:30 am

Re: Thanks and suggestions for improvement

Post by deepankardatta »

jonathon wrote:hi, starting with 1)

what version of macOS are you using?

jonathon
10.13.3 macOS High Sierra
Running on a 2013 MacBook Pro
User avatar
jonathon
Posts: 2609
Joined: Fri Jan 27, 2017 10:04 am

Re: Thanks and suggestions for improvement

Post by jonathon »

hmm, i've been looking into this, but it all works fine here.

still looking into it.

with thanks
User avatar
jonathon
Posts: 2609
Joined: Fri Jan 27, 2017 10:04 am

Re: Thanks and suggestions for improvement

Post by jonathon »

hi, could you try this version, and see if you can access the jamovi library with it?

https://www.jamovi.org/downloads/jamovi ... -macos.dmg

with thanks

jonathon
deepankardatta
Posts: 21
Joined: Mon Jan 29, 2018 4:30 am

Re: Thanks and suggestions for improvement

Post by deepankardatta »

jonathon wrote:hi, could you try this version, and see if you can access the jamovi library with it?

https://www.jamovi.org/downloads/jamovi ... -macos.dmg

with thanks

jonathon
0.8.2.3 is working without issue now
User avatar
jonathon
Posts: 2609
Joined: Fri Jan 27, 2017 10:04 am

Re: Thanks and suggestions for improvement

Post by jonathon »

ooh, i should revisit these other issues you raised.
Post Reply