v0.10.4

no message

OSF API integrated + survey module reworked

v0.9.0.

improvements to testing

• when the test button is clicked in the edit run view, test emails are sent to your admin email address, not mailinator
• there are new buttons when testing surveys: a light bulb shows items that were hidden on the page, a magnifying glass shows debugging info (item names, requests sent to openCPU)
• our openCPU debugger (the accordion with info on the openCPU request) is now shown whenever there is an r problem (or you click the magnifying glass). it has the option to download the r code or the rmarkdown. That way you can edit it in Rstudio, which has much better code highlighting for debugging.

improvements to the survey module

• now hidden items work consistently: every showif is only called once unless the return value in R is NA.
• only items on the currently visible page are evaluated (before this calculate items at the end of the survey were also evaluated)
• we've minimised costly openCPU requests

css improvements

• there are no more flashes of unstyled content (where the mc_buttons only appear after a moment)

Open Science Framework API connection

• barring unforeseen changes on the side of osf.io it is now possible to connect runs to an OSF project. Currently, you can only export the Run-JSON there (in essence, saving all of the run structure and survey items). You can use this for pre-registration or version tracking.

0.8.1

bug fix release, testing how zenodo mints DOIs for releases

v0.8.0

Version 0.8.0

Issues closed by this release here.

Important additions:

• Items can be shuffled (in and between blocks) now.
• Timed submit buttons allow you to specify a minimum time in which to submit.
• Item tables can be imported straight from Google Docs, reducing the edit-download-upload-cycle previously necessary. Makes collaboration easier.
• A proper test mode
  • you can now create test codes for runs. They're special in that they enable a bar with specific features (auto-filling forms, skipping to the next step in the run/ending a pause).
  • you can now share test codes for runs before releasing the run to the public
  • test codes have names like zanyElephant, which allow you to find them again if you're looking for a specific test case
• the first semblance of an OAuth2.0 API (manual activation needed)
• users can turn off emails from your study for a week or forever if you allow them to.

v0.7.2

Release v0.7.2

v0.7.1

Release v0.7.1

v0.7.0

There is only one change that could change behaviour in some special cases:
It's now possible to present submit buttons by themselves on a page.
Sometimes this might happen unintendedly if all items on a page are skipped via a showif, but the submit button doesn't have the same showif.
In this case, previously, the submit button would never be displayed. We decided it would be more consistent to show it once. This shouldn't break anything, but might lead to an extraneous step in some very complex studies
(most of which we've pre-tested though).

Here's a list of other changes to expect:

• calls to openCPU are now grouped and hence much faster, this will be especially noticeable for very complex studies
• dynamic item and choice labels now have the survey context (ie. you can simply write r item1 instead of r survey1$item1, but the latter still works). This makes surveys more portable (you can simply rename them).
• debugging R mistakes that you made should now be generally easier and more consistent (if you test your studies as an admin, you'll get to see the familiar debugging interface)
• notes should no longer be sticky (in some cases users reported problems with notes being repeated)
• instant validation has been turned on. Previously if you made a mistake, this was only alerted before submission, now it happens after you go to the next question
• sending text messages has been simplified (see the Import button in your run) or here https://github.com/rubenarslan/formr.org/wiki/How-to-send-text-messages-(SMS)
• it's now possible to set an expiry time (in minutes) for your survey. If a participant is inactive for longer than the specified time, the survey will wrap up and the run will move on. This is especially useful for diary studies, where participants might sometimes start filling out and then close their browser, getting stuck.
• a few other minor bug fixes, see https://github.com/rubenarslan/formr.org/issues?q=is%3Aissue+milestone%3Av0.7.0+is%3Aclosed
• it's now possible to scaffold a complete diary design by simply clicking the import button and selecting basic diary. We'll be looking for other "standard" designs, that can be imported this way, to make formr easier to use.

new run modules

reorganised run modules

branches are now SkipForward and SkipBackward. TimeBranches have been eliminated in favour of better-documented examples for SkipForward, that fill the void.
Shuffle has been added as an easy way to randomise participants.
Generally, tried to add a music player controls theme to it to make it more intuitive.

rethought some of the logic behind the scenes

e.g. skipifs now are showifs

improved documentation

all modules now have documentation with examples, so do item types, the survey. Internal documentation has also been proved.

mobile layout improved

mainly by fixing some minor annoyances and switching the viewport to device-width

formr now deserves the R in its name

Now all user-supplied run logic is written in R. This makes formr much safer and hopefully easier to use. Also, emails now support embedded figures, something which worked only on iPhone Mail until now.

last release using MySQL for branches, skipifs etc.

0.2.0

last release with MySQL based Branches, skipifs etc.