Re: Feature Request, Bug Data Export

Let's see.. First a general comment on the state of the different branches 
of mod_survey. Today there are two available code branches.

The stable branch ("3.2.4") is what people use. I view this as mature and 
stable, apart from some minor bugs (some of which you point out in your 
mail). If i would fix a bug report, this is the branch I would fix it in.

The development branch (no version number) contains large code updates and 
feature additions. While it works technically, it is by no means ready for 
production use. Further, it is in some places incompatible with the stable 
branch.

The stable branch has some fundamental code structure issus which makes it 
difficult to extend. In some parts, it builds on perl code as old as end 
of the 90s. The development branch was hoped to address this, but 
unfortunately there were not enough personel resources to make the major 
intervention that was needed. Instead, changes were made in parts of the 
code. This fixed some detail problems with the stable branch, but in sum 
total, the fundamental code structure problems remain, making the code 
very difficult to overview and maintain.

At this point in time I have a good grasp of the stable branch and know 
its strengths and weaknesses. For the development branch, I do not feel 
secure enough to do large interventions, since I cannot fully overview the 
consequences.

All in all, the development has stalled on both branches since they suffer 
from the same basic problem: Being very difficult to overview, maintain 
and extend. Solving this problem would require more effort than what is 
currently possible to put into an open source project maintained on a 
hobby basis.

Now, what all this means is that when I say that basic code exists in the 
development branch, then that's not really a comfort. The development 
branch isn't likely to hit production anytime soon, and backporting the 
changes to the stable branch will prove very difficult.

To address your points in more detail:

* Progress bar: Yes, I'm aware of this deficiency. Problem is as stated 
above: difficult to backport working code from the development branch. It 
is unlikely this will be fixed anytime soon.

* Mandatory fields. Same. This actually works in the development branch, 
but only through the cost of having made really intrusive code changes in 
the basic management of routing and branching. I am unwilling to make 
these changes in the stable branch at the moment, since the consequences 
are difficult to overview.

* Delimited fields: This was a new one for me. I don't have any 
explanation for it at the moment, but I doubt it should be very tricky to 
fix. I have been playing around with doing a maintainance release of the 
stable branch for some time, and this could likely by fixed there (see 
also list of other minor fixes below).

* Text below/under submit buttons. This is a feature which should not be 
hard to add in the stable branch. I'll take a look at it. At the moment 
you'd have to edit the perl code to add text there.

* SPSS export: Yes, this is a limitation in older (versions 7.5 and 
similar) SPSS. I've let the code remain since I don't really see any point 
in exporting large text fields to SPSS and since it is optional anyway.

So a summary of things that could end up in a maintainance release:

* Fix for delimited fields
* Feature addition for layout around submit buttons.
* Uniqueness checking on IP and/or cookie (stub exists, but not 
implemented)

No promises whatsoever on when/if it will be released, but at least I'm 
thinking about it. :-)

  // Joel


On Fri, 3 Aug 2007, Clemens Gruber wrote:

> Joel Palmius wrote:
> > And, oh.. If you have ideas and bug reports I'll be interested in
> > hearing about them no matter what happens with MATC. :-)
>
> Hi Joel,
>
> thanx for this encouragement. ;-)
>
> First I'd like to reminde the already reported bug concerning the
> progress bar
> http://article.gmane.org/gmane.comp.apache.mod-survey.general/1224
>
> You wrote:
> > But it seems something is missing in the code, it must have fallen off at
> > some point.
>
> A really improvement (this is a feature request :-) for mod_survey would
> be a revision of the mandatory fields handling. I wrote this some time ago:
>
> > The implementation of <CHOICE .. MUSTANSWER="yes"> is a problem. You
> > get the message that an input is missing on the next page (...).
> > You have to go back (by browser
> > function) and you have to search "by hand" the missing fields. Would it
> > be possible to present the "normal" page with an additional text on top
> > "mandatory fields are empty" and a message "pleas fill out" next the
> > empty but mandatory fields? It would improve the usability for the
> > mandatory fields a lot.
>
> You wrote
> (http://article.gmane.org/gmane.comp.apache.mod-survey.general/1243)
> "initial code for it exists in the development branch"
>
> It would be great to have this function in a stable version. The lack of
> this function is the reason for using OSP instead of ModSurvey in some
> surveys where we need mandatory fields.
>
> By the last data export I noticed a minor bugs:
>
> - in the "Abgegrenzte Felder" export (I hope you know the english
> aquivalent, we have here a german installation) the option
> > Benutze [ ] als Trennzeichen.<
> with the pipe | as custom specified delimiter makes trouble, a "normal"
> delimiter like @ generates this output
>    20070621_082601_9828@-1@-1@- bitte auswählen -@
>
> but the pipe provides this:
>    2 0 0 7 0 6 2 1 _ 0 8 2 6 0 1 _ 9 8 2 8 | - 1 | - 1 | - b i t t e a
> u s w ä h l e n - |
>
> Btw the SPSS option >String-Werte anfuehren (" enthaltene Werte werden
> ersetzt durch \').< would by good by the "Abgegrenzte Felder" export too.
>
> Another feature request: It would be nice to add some text next the send
> button. This does not work
>
> >      <ROUTE CONTINUE="question-2-5.survey"/>
> >   <CUSTOM>{¤footer.inc.html¤}</CUSTOM>
>
> Is the limitation to 250 characters >Sofern nicht ausgeschloessen, sind
> MEMO-Felder enthalten, werden aber nach 250 Zeichen abgeschnitten.< on
> the SPPS export a mod_survey limitation? I guess longer stings were a
> problem of old SPSS versions, perhaps it is no problem with newer one.
> So the message could be deleted or adjusted.
>
> We use the latest stable releases 3.2.4.
>
> Kind regads.
> Clemens
>
> --
> Clemens Gruber (Dipl.-Psychologe)
>  Zentrum fuer Informationsmanagement und
>  virtuelle Lehre der Universitaet Osnabrueck (virtUOS)
>
> Schloss-Strasse 9, Raum E07, D-49074 Osnabrueck
>
> E-Mail:  cgruber@xxxxxxxxxxxxxxxxx
> URL:     http://www.virtuos.uni-osnabrueck.de
>
> Telefon: (05 41) 969-6505 oder
>         (05 41) 969-6501 (Sekretariat)
> Fax:     (05 41) 969-16505
>
> Skickat av Clemens Gruber <cgruber@xxxxxxxxxxxxxxxxx>
> till survey-discussion