RE: Submitting incomplete forms

A model can have multiple submission elements --
though what they typically do is to submit possibly different portions
to possibly different locations.
It's not for doing multiple levels of validation --though you could
achieve that 
by having the different locations where the data goes to do different
levels of validation.




    Mark> Thinking out loud ...

    Mark> So, again, unless we are missing something, it might be
    Mark> useful if either an XForms model could have more than one
    Mark> submission element, each with a validation level; or maybe
    Mark> multiple submission elements, but have the validation level
    Mark> set at submit time (be nicer to have in the model maybe?).
    Mark> Something along those lines though.

    Mark> I suppose it would be quite nice if you could just bind form
    Mark> elements to multiple models, each model specifying different
    Mark> validity constraints and different submission conditions.
    Mark> Each model could constrain as little or as much as required.
    Mark> You would even be able to get a user to "fill in" two (or
    Mark> more) XML instances for the price of one. It might be
    Mark> horribly messy to manage though ... you would have to make
    Mark> sure you only used one at a time, yuk!

    Mark> With things as they stand, I am wondering whether I will
    Mark> just have to resort to lax structural validation in the
    Mark> XForms form, and simply pass back a flag element/attribute
    Mark> to indicate whether the user is "saving" or "submitting".

    Mark> All the best

    Mark> Mark


    Mark> -----Original Message----- From: Karandikar, Shailesh
    Mark> [mailto:Shailesh.Karandikar@dendrite.com] Sent: 04 November
    Mark> 2002 15:55 To: 'XForms@yahoogroups.com'; 'www-forms@w3.org'
    Mark> Subject: RE: Submitting incomplete forms



    Mark> Hi,

    Mark> XForms requirements mention save and continue feature.

    Mark> I would like to distinguish between incomplete-forms from
    Mark> invalid-forms. Typically, 'Required' data would result in
    Mark> incomplete forms. It would be up to the author to decide
    Mark> whether to allow 'saving' of such forms. Incomplete forms
    Mark> are not 'submitted' but 'saved' to a local or remote
    Mark> persistence layer to be retrieved later; The exact mechanism
    Mark> depends on the underlying framework.

    Mark> Unless I've missed something obvious, I do not find this
    Mark> behavior clearly explained in the processing model. It would
    Mark> be useful to be able to 'declaratively' specify this
    Mark> behavior. XForms author should be able to specify and
    Mark> control following behaviors:

    Mark> 1. Validation modes supported by the form: delayed,
    Mark> incremental, relaxed, or validate-on-submit, etc. Although
    Mark> this would complicate the validation semantics and may
    Mark> result in inconsistent user interactions/interface, it would
    Mark> be the responsibility of the XForms author to ensure overall
    Mark> consistency.

    Mark> 2. Whether an xform instance is allowed save-n-continue
    Mark> feature. This would be a 'computed' property, based on
    Mark> certain conditions: E.g. User must enter name and birth-date
    Mark> before suspending a credit-card application. I.e., ability
    Mark> to impose validations on fragments when using
    Mark> save-n-continue feature. This expression could form a new
    Mark> model item property, to be applied depending on the
    Mark> validation modes supported by the forms.


    Mark> Regards, Shailesh Karandikar, Dendrite Inc.

    Mark> -----Original Message----- From: Mark Seaborne
    Mark> [mailto:MSeaborne@origoservices.com] Sent: Monday, November
    Mark> 04, 2002 3:32 AM To: XForms@yahoogroups.com;
    Mark> www-forms@w3.org Subject: Submitting incomplete forms



    Mark> Hi

    Mark> I assume it is a fairly common requirement for someone
    Mark> filling in a form to be able to make regular saves, and
    Mark> maybe to be able to start editing on one machine, and finish
    Mark> on another machine at a later time. This suggests to me that
    Mark> the ability to submit incomplete instances, for appropriate
    Mark> processing, would be useful. Has anyone done any work like
    Mark> this with XForms?

    Mark> I haven't read the latest working draft as carefully as I
    Mark> would like, but it looks to me as though saving an
    Mark> incomplete form that is not valid against a referenced W3C
    Mark> XML Schema is not possible, unless the schema has been
    Mark> designed for this use (in which case the instance would be
    Mark> valid, of course). Given that many of us will have no
    Mark> control over the schemas used for validation, that may not
    Mark> be possible (I assume).

    Mark> I cannot see that I will fare much better using model items
    Mark> to validate. Iis there some way to make properties, such as
    Mark> required, conditional. If I have missed something, I would
    Mark> be grateful if someone could point me in the right
    Mark> direction.

    Mark> It seems to me that it would be useful for forms based
    Mark> applications if the author could define multiple states of
    Mark> validity, each with their appropriate actions, such as
    Mark> submit to X for processing, submit to Y for storage and
    Mark> later completion, submit to Z for some external validation
    Mark> processing, etc. If a user has incorrectly completed a form
    Mark> for some reason, one still might be interested in retaining
    Mark> the XML instance in order to see what they have done wrong,
    Mark> and to provide assistance maybe. Kind of thing that
    Mark> Schematron lets you do.

    Mark> I expect I am just missing something really obvious. If so I
    Mark> would be really grateful if someone could point me in the
    Mark> right direction. Otherwise, I guess it will be a matter of
    Mark> minimal validation within the XForm, and do most of the work
    Mark> at the server. Perhaps that is the safest route to go
    Mark> anyway.

    Mark> All the best

    Mark> Mark Seaborne

    Mark> Origo Services Ltd








    Mark> **********************************************************************
    Mark> This email and any files transmitted with it are
    Mark> confidential and intended solely for the use of the
    Mark> individual or entity to whom they are addressed. If you have
    Mark> received this email in error please notify the system
    Mark> manager.
    Mark> **********************************************************************

-- 
Best Regards,
--raman
------------------------------------------------------------
T. V. Raman:  PhD (Cornell University)
IBM Research: Human Language Technologies
Architect:    Conversational And Multimodal WWW Standards
Phone:        1 (408) 927 2608   T-Line 457-2608
Fax:        1 (408) 927 3012     Cell: 1 650 799 5724
Email:        tvraman@us.ibm.com
WWW:      http://www.cs.cornell.edu/home/raman
AIM:      TVRaman
PGP:          http://emacspeak.sf.net/raman.asc
Snail:        IBM Almaden Research Center,
              650 Harry Road
              San Jose 95120

Received on Tuesday, 5 November 2002 12:49:01 UTC