Hi,
I found LimeSurvey in june 2020 and my colleges installed Version 4 back those days. I stumbled into many problems, which to me seemed to be browser related during editing questions. (I usually use Firefox, with Chome there seem to be less question)3.25.4
Now we want to publish a survey, and I am not sure, if I should use 4.3.17 (or newer) or 3.25.4 (or newer) The given Versions are the installed versions.

What I need to do in the survey: Give people a list with about 60 small pictures, where they can choose the preferred ones and put them in a ranking. I assume both might work with both version; but I do not know how "stable". In LS 4 I have three different approaches, where I want to put some CSS on to make it better looking.

The thoughts I have:
a.) Version 4 seems to be really early stage. I can do what I need to do, but sometimes the interface for the questions is english only (I am germany located; I tried to get in touch with the translation team, but without any luck), questions are not saved for "unknown reasons", big red info boxes appear with _no_ text. I am willing to help, but cannot program, I only can try to give detailed description of what I am doing. I also see the problem, that I might open duplicate issues in the bugtracker, as I noticed in my searching that I use different wordings.
So I personally stumbled in about 20+ bugs all from "nice to fix" to "bummer, show stopper", in a combination with "Firefox a problem, chrome less a problem"

b.) Use Version 3. My preferred solution at the moment, but I do not know which features (I am not missing now) I might miss compared to Version 4. And if "4" is the future, I want to help making 4 better, than investing time into version 3.

I am stuck, and have no real idea what to do.

Any suggestion?

all the best
Juergen

4.x is still not listed as stable in the downloads page. If you need a production survey, use 3.x.

- community.limesurvey.org/downloads/

Hi,
we noticed, that 4 ist not for download. But if not tested in production, stabelizing will take longer. So my hope was, that V4 is usable for production in the frontend (if a survey is there, and can be clicked thru, it will work always, and no answers are lost), and the "only" problems appear for the people that use LimeSurvey for generating questions.
To me it seems to be "buggy"; but with some knowledge (which we have in the company for CSS, Javascript, etc. pp.) We could deliver some input to get V4 stabelized faster.
Or is this a very bad idea?
many thanks
Juergen

Hi, Jürgen,
why do you think the two 4.x. versions are not to download?




So, please, download, install and test. Therefore we have the new Release Candidate,

Joffm

Where ist the future of LimeSurvey

Easy: the future of Limesurvey will be version 10 and more one day.

Joke aside, of course the future of LS will be LS 4.4.x and then LS 5 thereafter.

However, at the moment LS 3.x is the one to use for production, LS 4.x is currently still too unstable for that. You probably won't miss any features from LS 4.x in LS 3.x.

However, you are free to download it and play around, test and give feedback, write bug reports. It will help to get it (back) to "stable" quicker. A couple of days ago they released LS 4.4.0 RC1, which is supposed to be solving a lot of the problems in LS 4.x (haven't tested yet).

The translation problems are a known problem with LS 4.x, it has mostly nothing to do with the translation itself, but rather with LS 4.x not picking up some of the translated strings. So even if the language is 100% translated, there will be parts in English still. Not sure if this is solved in LS 4.4.0 already.

So I would recommend to go with LS 3.x if you want to do surveys now. I don't think that LS 4.4.x will be stable with the next 2-3 months. At least not with my definition of stable. In the past the developers have shown with the release of "stable" LS 4.x that their definition of stable is different to mine, but as they changed back to not stable I have hopes that after two big fails (one with LS 2.5x and then again with LS 4.x), that they might have learned from it - fingers crossed.

Hi,
ist there a way to know where LS4 is not stable. I noticed it beeing unstable in the Backend, where the user generates the questions.
I can live with that (somehow), as I managed to get the survey done. But it would be a desaster, if answers in a running survey are lost, or answers are added to wrong questions or similar things.
Does someone have any experience on that?

many thanks for any answers.
Juergen

ist there a way to know where LS4 is not stable.

You would have to go through the bug reports in the Mantis to see which bugs are still open and which are already fixed. But then there might be Bugs that have never been detected.

But it would be a desaster, if answers in a running survey are lost, or answers are added to wrong questions or similar things.

Go with LS 3.x! It is not a big deal to upgrade later, especially with Comfort Update. But even if you don't want to pay for Comfort Update, it is pretty easy to upgrade to a higher LS-Version later.

I mean you can always risk it, but then that is up to you... In my personal opinion, for serious work, you should go with LS 3.x. With LS 4.x no one really knows if there are any major but hidden bugs.

As @holch suggested above, go for LS 3.x (LTS) since you want something for production.

LS4.4.0 will be released in the next few weeks - but we rely on community and users to report everything that passed QA & automated testing.

It would be great if you can run the two versions in parallel and see what issues exist in LS4 and have them reported in the bugs tracker.

HI. Many thanks for the answers; but beeing a newbie, it is really hard to distinguish between a bug and my personal stupidity
I go back to 3.X, learn, and than I will spent time in using and testing Version 4

lsurvey20 wrote: Many thanks for the answers; but beeing a newbie, it is really hard to distinguish between a bug and my personal stupidity

No need to be a newbie for this

Well. In another thread, I though I got a bug, but in fact, it was a screwed up CSS, that made the problem. So as a newbie I did not catch this at the first glance. After digging around for 4 hours, I asked here in the forum, abd got the tip within 10 minutes. And testing a system only with "fantasy" data, is not what really tests it, IMHO. I strongly believe, that real life problems will face real life bug, as often combinations of dependencies make a bug. Anyway: I wioll try to get used to it, and do some testing with V4

I strongly believe, that real life problems will face real life bug, as often combinations of dependencies make a bug.

I agree that not all bugs will be found by testing. But I belieave that testing can probably can weed out 95-99% of the most obvious bugs which is already a great step forward. Then the rest will appear during the use in production, but these should be minor bugs and bugs that only occur in very specific occasions, which makes them less problematic in my point of view.