Hi all,

Is there a way we can actually determine what devices were used to answer the EXPIRED surveys? I am familiar of the useragent string and I have implemented it to some of my surveys. But of course, it only works when the current survey is or will be going active. Do I perhaps need an external analytics tool?

Hmm, I am not sure if I understand.

You want to know for a finished survey (respondents have given their response already) what device they used? This won't be possible, because you didn't save the necessary information.

If this is not what you are trying to do, please explain more in detail.

Expired surveys cannot be answered anymore. There is currently no hidden information captured inside LS to reveal the browser infos afterwards. The only way might be the webserver logfiles to match certain entries. But that may not work nicely and can be problematic in terms of compliance.

#
The workarounds with questions and javascripts to capture useragent infos are not good enough to capture cases like this.

I would like to see a user agent capturing on php code base which can capture information before the first question is displayed.
That way a debugging for cases where the limesurvey isn't displaying a question at all could be easier handled.
The javascript part is still needed for infos like screen resolutions and plugins. But a out of the box a user agent would be possible with plain php.

This could be implemented into paused or expired surveys too:
www.limesurvey.org/forum/future-features...tallation-per-survey

jelo wrote: I would like to see a user agent capturing on php code base which can capture information before the first question is displayed.
....

Existing plugin : git.framasoft.org/SondagePro-LimeSurvey-...in/findUserAgentInfo

You mentioned that plugin already but stated that it isn't stable or production ready.

And since you're that productive and develop plugins I wonder why the team couldn't ship them with the installation disabled.
That would bring more traction to the development. You want more plugins to ease maintenance of the core.

Hi,

Actually : all my plugin are distributed without any guarantee (it's GPL).
Stable enough to be used in some production environnement.

For this one : i don't do the documentation, zip distribution and "home page". But can be used , need test on different LS version.?

And since you're that productive and develop plugins I wonder why the team couldn't ship them with the installation disabled.

1 : If i distribute plugin non my own : i'm responsible of foixing bug for this plugin. Actually, i don't have time to test all plugin for 2.5 compatibility. If it's include in LS default distribution, i'm responsible of the compatibility when 2.5 is out.
I need to work for my client sometimes ....

2 : Take the exemple of worpress : wordpress have a lot of success , this success is due to a lot of plugin and template. A empty wordpress have nothing and it just can be used for "blog". Have something without extrenal plugin seems better ro have a lot of plugin.

3 : ANy one can take my plugin and make own adaptation : is more easy to adapat an external and separated plugin than a plugin aloready existing in core.


Denis

Thanks for elaborating theses points out. Because if you post just a link people think a feature request is obsolete and it is already all done.
Which is not.

jelo wrote: ....
Because if you post just a link people think a feature request is obsolete and it is already all done.
....

This feature request is done ?

ANd remind : i'm not the lead dev. I don't make any choice. If Carsten want to include a plugin in core, no problem .

The, : we such system
- User put a feature request
- External dev make a plugin and put the link in FR
- Decision is made to include this plugin in core or not (in ./core/plugins or ./plugins/ or ./plugins/demo/ )

I remind that (for a running survey) you can always use Google Analytics to monitor the kind of devices used.
You just need to add the google code in the notification & data management panel.
Cheers,
Elo

But you shouldn't be able to connect the Google Analytics data with specific responses. Otherwise this would be a big problem. So if you need to connect the user agent with the responses, the Google Analytics approach won't work.

If you just want to know in general some statistics about your users, then of course the Google Analytics approach is valid.

Correct me if I am wrong about not being able to connect Google Analytics results with specific answers.

With GA you cannot link the device to a specific respondent.
If you want capture the data of the device for a respondent (and other, such as the browser) you have to add a javascript frame in your survey.
You can find how here (tanks to Tony Partner):
www.limesurvey.org/forum/can-i-do-this-w...espondent-s-paradata
Cheers,
Elo

Holch wrote: But you shouldn't be able to connect the Google Analytics data with specific responses. Otherwise this would be a big problem. So if you need to connect the user agent with the responses, the Google Analytics approach won't work.

Not that sure...In theory, yes you can since you can perfectly use LS {keywords} in GG analytics event tracking code

For example you can set a next/submit or any link with an onclick like: