W3C home > Mailing lists > Public > www-forms@w3.org > March 2005

Re: Xforms submissions can rarely be reloaded, can this be fixed?

From: Jason Eacott <jeacott@hardlight.com.au>
Date: Sun, 6 Mar 2005 09:48:00 +1030
To: Stefano Debenedetti <ste@demaledetti.net>
Cc: www-forms@w3.org, chiba-users@lists.sourceforge.net, joern Turner <joern.turner@web.de>
Message-ID: <422AD1D0.10877.28ED0E61@localhost>

thank you.
	yes I do mean that parts definitely arent submitted because relevant is false on them (I am using Chiba here).
this seems to be correct given this entry

http://www.w3.org/TR/2003/REC-xforms-20031014/slice6.html#model-prop-relevant

and means that you could be throwing away good data just because you dont want to display it. just because its not relevant for a user to see 
doesn't mean the data should be dumped, it may become relevant later (much later, after multiple rounds of reloading!), and it may have been data 
that they originally entered.
and because switch is not as dynamic as it should be there is often no choice but to use relevant to control the ui.

but what I read into the references you mention is that I should be able to reload this submission regardless of missing parts during the same form 
session and the  missing parts SHOULD be regenerated using any bind elements as a reference. 

in this case I think Chiba is  broken here. (I have gotten around it given a known schema by reprocessing the instance with xsl & replacing the 
missing nodes & attributes manually on request - but this sux IMHO)
Its submission is correct but its reloading does not seem to try to recreate missing nodes, but rather just throws xpath exceptions.

I like that I can control what is actually being submitted by using a relevant property, but it means that using the UI has side effects in the instance. 
there should be another switch, or separate instance-relevant property to indicate whether you intended the relevant to be ONLY for the ui or that it 
should result in culling data from the instance.

thanks for the response.
Jason.



Date sent:      	Sat, 05 Mar 2005 18:32:03 +0100
From:           	Stefano Debenedetti <ste@demaledetti.net>
To:             	www-forms@w3.org
Subject:        	Re: Xforms submissions can rarely be reloaded, can this be fixed?
Forwarded by:   	www-forms@w3.org
Date forwarded: 	Sat, 05 Mar 2005 17:29:06 +0000

> 
> Hello Jason,
> 
> Jason Eacott ha scritto:
> > Hi All,
> > 	I just noticed this little caveat of the relevant attribute:
> > 
> > 
> >>Instance data nodes with this property evaluating to false are not > serialized for submission.
> > 
> > 
> > this means that If I ever use relevant (and its a dull form that doesn't) then I can never reload that 
> > same instance data if I have binds or refs to ANY of the parts that relevant is set on.
> 
> Not sure I understand what you mean (do you mean the parts that weren't submitted because relevant was false on them?) but it doesn't seem to be the case according to:
> 
> http://www.w3.org/TR/2003/REC-xforms-20031014/slice4.html#evt-modelConstructDone
> 
> > surely this is ridiculous. isnt one of the primary outcomes of x-forms supposed to be that you 
> > can easilly manipulate xml, save it, and then read it back in top modify it sometime later?
> 
> The XForms design goals are stated here, it doesn't seem to me that they cover exactly what you mentioned:
> 
> http://www.w3.org/TR/2003/REC-xforms-20031014/slice2.html
> 
> > This is just about the last straw for me and x-forms I think.
> 
> Do you mean you like h^TML forms better? ;-) Sorry about this one..
> 
> > UNLESS
> > 
> > can anyone tell me whether if I load (ie via <submission replace="instance" />) new instance 
> > data that is missing elements and/or attributes that are bound in the form, whether the x-forms 
> > processor should ignore the missing parts, or whether it should throw an error?
> 
> See reference above for the exact rules and algorithm.
> 
> I may be getting it wrong but it seems to me that XForms 1.1 will make even easier to control this issue as it would allow to create arbitrary instance nodes on the fly in response to events.
> 
> > hoping someone can shed some light.
> 
> HTH, ciao
> ste
>  
> > thanks
> > Jason.
> > 
> > 
> > 
> 
Received on Saturday, 5 March 2005 23:22:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:22:00 GMT