Written by jelo the 9 Dec 09 at 09:38.

Can somebody with french language skills take a look at the attached script and translate the comments into english? It would be a big step towards a broader adaption of maxdiff inside of limesurvey.

Thread:
http://www.limesurvey.org/en/support/forums/5-can-i-do-this-with-limesurvey/259 71-maxdiff-question-style-available-or-workaround

Post with attached workaround script:
limesurvey/25971-maxdiff-question-style-available-or-workaround?start=10&lang=e n#33094

Can somebody with french language skills take a look at the attached script and translate the comments into english? It would be a big step towards a broader adaption of maxdiff inside of limesurvey. Thread: http://www.limesurvey.org/en/support/forums/5-can-i-do-this-with-limesurvey/25971-maxdiff-question-style-available-or-workaround Post with attached workaround script: limesurvey/25971-maxdiff-question-style-available-or-workaround?start=10&lang=en#33094

Written by lemeur the 3 Jan 10 at 13:59.

MaxDiff question is in fact a specific version of Array (flexible labels) by column:
* answers are "Makes me Most want to reenlist" and "Makes me Least want to reenlist"
* Labels are "Time off each week allowed for on-line college coursework", "Receive in-person personal counseling services", ...

Now for the new things (switchable by a question attribute):
* a new layout: display answers between the two columns
* a small javascript that prevents using the "Next/Commit" button if the same "Label" is used for the two "Answers"

MaxDiff question is in fact a specific version of Array (flexible labels) by column: * answers are "Makes me Most want to reenlist" and "Makes me Least want to reenlist" * Labels are "Time off each week allowed for on-line college coursework", "Receive in-person personal counseling services", ... Now for the new things (switchable by a question attribute): * a new layout: display answers between the two columns * a small javascript that prevents using the "Next/Commit" button if the same "Label" is used for the two "Answers"

Written by lemeur the 3 Jan 10 at 14:33.

* First of all, email messages sent by LS must be tagged in some way with an message-dest-id
This message destination id is made up of: the SurveyID, the target email address, the token code, and a specific integrity code.
This integrity code is used to make sure the bounce email is not a fake so that we don't disable some valid tokens.
The integrity code would be an ASCII verison of a HMAC of the previous elmeents encrypted with a key containing the token ID (internal element inside LS, and a global value specific to this LS installation and set in config.php).

* the probe will decode the bounced email to read the error message and get the original message-dest-id, then it would send the two data to LS-remote-command (via SOAP)

* The LSRC API will provide a service that would check the integrity and authentication of the message-dest-id and then update the token status accordingly.

Some references:
* Tagging with VERP solution: http://en.wikipedia.org/wiki/Variable_envelope_return_path
* Tagging in SMTP header (my preferred way): http://www.php.net/manual/fr/function.mail.php#78934

* First of all, email messages sent by LS must be tagged in some way with an message-dest-id This message destination id is made up of: the SurveyID, the target email address, the token code, and a specific integrity code. This integrity code is used to make sure the bounce email is not a fake so that we don't disable some valid tokens. The integrity code would be an ASCII verison of a HMAC of the previous elmeents encrypted with a key containing the token ID (internal element inside LS, and a global value specific to this LS installation and set in config.php). * the probe will decode the bounced email to read the error message and get the original message-dest-id, then it would send the two data to LS-remote-command (via SOAP) * The LSRC API will provide a service that would check the integrity and authentication of the message-dest-id and then update the token status accordingly. Some references: * Tagging with VERP solution: http://en.wikipedia.org/wiki/Variable_envelope_return_path * Tagging in SMTP header (my preferred way): http://www.php.net/manual/fr/function.mail.php#78934

Written by miamiman the 5 Jan 10 at 00:39.

FEATURE REQUEST: That stats/graphs take into consideration the sub-sets created by one or "n" branched questions.

FEATURE REQUEST: That stats/graphs take into consideration the sub-sets created by one or "n" branched questions.

Written by lemeur the 8 Jan 10 at 15:39.

Since 1.87, questions that have been hidden because will have answers field set to NULL in DB, which makes it possible to count only the total number of participants that has answer a given question.

Thus, we could add an option just before running the stats survey so that computations of percents are not computed against the total number of respondants for the survey, but the total number of respondant for this question.

Note however that this NULL differentiation of conditionnal questions makes sense only for completed responses (when the final submit button has been clicked), and thus the option would be available only when the Completed response filter is set.

Since 1.87, questions that have been hidden because will have answers field set to NULL in DB, which makes it possible to count only the total number of participants that has answer a given question. Thus, we could add an option just before running the stats survey so that computations of percents are not computed against the total number of respondants for the survey, but the total number of respondant for this question. Note however that this NULL differentiation of conditionnal questions makes sense only for completed responses (when the final submit button has been clicked), and thus the option would be available only when the Completed response filter is set.