Who copies whom? Is one better than the other - why?

Discuss the jamovi platform, possible improvements, etc.

by stats2019 » Wed Jan 23, 2019 8:34 am

Hi,

after I discovered JASP yesterday and today JAMOVI, I am just wondering who is first: The henn or the egg - or in better words:
Is there an original and a copy or how do these two fit together?

JASP ( https://jasp-stats.org ) or JAMOVI ( https://www.jamovi.org )

Is there a way to tell which is better (the first or the second ...)? Or does one actually need to install both GUIs?

If you look at the features, they are at least similar, GUI ... very similar...

It seems that jamovi can export the R code that has been used. This is a GREAT GREAT plus - if this is possible.
Also I can copy output from jamovi directly into LibreOffice. This did not work with jasp 0.9.2 - I had to copy to Word first and then paste again to Libre. Another plus.
I also like the possibility in jamovi to ass plugins. Nice (but gets the ribbon crowded).

Thanks already
Stats
stats2019
 
Posts: 7
Joined: Wed Jan 23, 2019 8:02 am

by jonathon » Wed Jan 23, 2019 9:12 am

hi,

here's a bit of a discussion of the history of the projects:

http://blog.efpsa.org/2017/03/23/introd ... ower-of-r/

but both were designed by the same people.

cheers
User avatar
jonathon
 
Posts: 724
Joined: Fri Jan 27, 2017 10:04 am

by stats2019 » Wed Jan 23, 2019 9:23 am

Hi Jonathon,

thanks for that very fast reply. I will look into the discussion.

Stats
stats2019
 
Posts: 7
Joined: Wed Jan 23, 2019 8:02 am

by stats2019 » Mon Jan 28, 2019 5:58 pm

Hi Jonathon

I will also post this to the jasp forum as I hope it will be seen as a request of a fellow programmer and let's say interested scientist.

I have had a look on both, JASP and jamovi in the last couple of days. They look quite similar, they have many features in common, but there are definitly differences. One (jasp) can easily be used to utilize a different data source by just copying file a.sav to the name used in jasp, the other has the feature to reorder your variables.
One could probably built a long list of commons and differences, and actually one might find himself still lost which one to use. Or, as in my case: use both. I liked some features better in jasp, others in jamovi.

So why do I bother to write.
I have been developing software for quite some time now myself. I know it's work, a lot of it. I also know, that working in a (programming) team brings different set of interest into development. One wants to put preference on this, the other on that.
Unfortunately you, jamovi and jasp programmers have split efforts instead of working somehow modular on the same project.

I understand, that each of the development groups (or people) can say, they take the other project as a challenge to be better. What I do not understand though, is why you do not join (again) and try to support a single project where each group can make the best of it's input to.

The GUI is great, really. But to attract the scientific community, we should learn from each other and support progress. This does not only mean, one unified group could be even more productive, but would also mean that some things would be intergrated faster (and better). GUI is not all, getting back to attracting not only students, I think of R script export. The R syntax mode in jamovi is a good start, but why bother and use either jamovi or jasp if one can not reproduce results with a new set of data, or publish script code for everyone to read.

In one single sentence: For the good of all, please consider working together with the other development team (again).

I am looking very forward to the progress we will see and hope my request will not fade unheard, especially because you both jasp and jamovi have your code on GitHub.

Please, reach out to the other party and join forces. Probably everyone will be better off.

Stats
stats2019
 
Posts: 7
Joined: Wed Jan 23, 2019 8:02 am

by jonathon » Tue Jan 29, 2019 3:20 am

hi stats,

The R syntax mode in jamovi is a good start, but why bother and use either jamovi or jasp if one can not reproduce results with a new set of data, or publish script code for everyone to read.


you would need to ask some of our many users :) but sure, these features are on the way.

wrt the separate projects; what you say is true, and it does grieve me that the projects aren't able to work together more.

however, there probably are more technical differences than you realise. for example, a goal which is important to us is that we provide a (native and integratable) web application. we're only a couple of months away from this goal, where as jasp would require more-or-less a rewrite to achieve the same thing.

however, this is less important than the philosophical differences.

jasp was written to promote bayesian analyses, where as my (and a few others) goal for the project was for it to be, additionally, an open platform for anyone to promote new statistical approaches. we found the goal to promote bayesian methods often overrode other goals, meaning we couldn't get the resources to develop other analyses, and in some cases we were outright blocked.

for us, it's very important that jamovi (and statistical software in general) is neutral, and doesn't have a strong agenda to promote one school of thought above another. i call this the separation of church and state :) it's important there's an open playing field where ideas can battle it out, and history decides.

you can see this reflected in the jamovi library, where we host a diverse range of modules written by people of wildly different statistical persuasions. this, for me, is the essence of what we're on about, and is crucial for science.

before we left the jasp project, we did try and negotiate this 'separation of church and state'. i proposed we split the project into the platform and the analyses, and we did negotiate for quite a while, but were ultimately unsuccessful.

similarly, before we launched jamovi, we approached the jasp folks about joining us (we had a lot of features at that stage that they lacked ... i guess we still do). but again, we were unsuccessful.

so, i agree it's a shame, but sometimes this is just the way things pan out.

so to sum up:

a) the differences between the projects are non-trivial
b) we've tried hard to bridge the two projects, but so far haven't succeeded
c) we're hopeful we'll find a way that our two projects can work together in the future
d) but if we don't, that's ok.

jonathon
User avatar
jonathon
 
Posts: 724
Joined: Fri Jan 27, 2017 10:04 am

by stats2019 » Wed Jan 30, 2019 10:46 am

Hi Jonathon,

thanks for your reply. I guess that's the way it is for right now. EJ from JASP answered as well (here: http://forum.cogsci.nl/discussion/4761/ ... her-why/p1 ).
I will keep an eye open on both projects and hope, as you are open source, there will be a time that you are able to cooperate again.

Stats
stats2019
 
Posts: 7
Joined: Wed Jan 23, 2019 8:02 am


Return to General