that is exactly why array is not needed, you can also solve it with single choice questions.

I would want to kill anyone who forces me to use single choice questions, if there is an array, when I am on a big screen...

While array questions should be avoided where possible anyway, there are situations where array questions are the best choice (clearer, quicker to respond, etc.).

Now you could argue that we then should not run this survey mobile at all, but I don't think that makes sense either. A template will not force survey creators to create better surveys. They will just switch the template and think that the template is sh*t...

I am not sure if a mobile only template is the best way to go, especially at the moment as we use such a huge mix of different screens.

E.g. I get really annoyed when a website forces me into the mobile version on a 10 inch tablet. A tablet with this screen size is totally capable of showing a normal website and mostly the mobile websites are losing a lot of space and are not very good to use, on bigger screens than a mobile phone (where they are perfect).

So I personally don't think it is the way to go.

Take my word for it, I will never force you to use this template...

PS : i have an old template where the client to make MOBILE survey want

Arrays with 3 lines and 2 Yes/No colored box.

Really clean and really easy to anser on phone (even on my old 4" with my big fingers)

Take my word for it, I will never force you to use this template...

Probably not you personally, but the survey developers that use your template to develop their surveys.

Here is a quick sketch about what we are thinking of for single choice (all answers in one block) or multiple choice (all answers separate).

Feel free to comment, but no death threats please.

Personnaly , i‘m more to move out all question from core to be in external modules, activated or not and loaded or not (or downloaded or not).

Then you can have a complete LimeSurvey with only some simple question, but if yu need more : you download it at «Question type store system»

I am not sure about the words "external modules", but I do see a good business case for "profiles" that can be defined at some level (user? theme?) where question types can be switched on and off.

And yes, let all question types be separate modules, that can be extended by other modules.

tammo wrote: I am not sure about the words "external modules", …

Nopt sure for myself of the exact way to do
But a directory inside LimeSurey where we can found manual.limesurvey.org/Question_object_types

In my opinion : this Question object must have
1. Extend another Question object allowed
2. Completly new Question object
3. Column type+number must be part of the question object (then when someone found times, we don't need all Array different questions workaround and plugin)
Etc …

See

A question type manager, like plugin manager, where it's possible to install, deactivate and manager which questions are available on the system

on wiki for base system.

I think the modular approach proposed is a good idea.

But my fear with this, as with the plugins as well, is that suddenly vital elements of Limesurvey suddenly won't work for a couple of months, when a new version comes out, etc.

So I think the question type modules should come with the core and should be tested before stable release. My fear is, that suddenly we will have a pure barebone system after each new release and I guess this will lead to even less bug testing by users than today, because the platform might not be "usable" for them until a certain time after, when all the modules are up to date.

On the other hand the modular approach, would give us the chance to finally implement many "question types" that currently can only be done via work arounds (nothing against work arounds, but in my opinion they should be kept at a minimum and only for special occasions) as a "real" question type. With the risk that they might not be always up to date, but this can also happen with the work arounds.

I fully agree that all questions that are in LS as it is now, are core modules.

After this, succesful additional modules could be taken up into the core set, but this will mean that the original developer of the custom modules looses "ownership" and the new core module will have to fit in the work flow of the core developers. Wouldn't it?

The same happended to various custom modules of Drupal, that now make part of the core package.

Tammo

Tammo

My exact position about this is
1. LimeSurvey default come with all current question in «question object» and «attributes plugin» (except, maybe some specific part (5 point emoji, timer script (because this one broke a lot and must be improved) etc … Hude text since it's exactly long text with different default attribute).
2. LimeSurvey light can be ditributed alone
3. LimeSurvey «another light»



I already create plugins for core, this plugins still active since creation (remind all exports are plugin currently).

the original developer of the custom modules looses "ownership" and the new core module will have to fit in the work flow of the core developers.

It's FLOSS, original author still their, but other dev can upgragde/improve.

I already create plugins for core, this plugins still active since creation (remind all exports are plugin currently).

Yeah, but you are one of the good ones...