Re: Last call comment about readonly property with calculate (PR#45)

Hi David,

Although it was appropriate for your response to go to www-forms-editor, 
note that I have moved the discussion to public-forms as that is 
appropriate for further discussion of the issues.

I agree that I do not know of the specific optimization you have in mind, 
but I don't recall that you have ever explained the optimization so that 
it can be assessed by the group.  This makes it difficult to make an 
exception to the previously defined behavior. 

We have a very useful use case articulated and even added it to the spec 
to resolve the non-clarity aspect of your feedback.  To adopt the specific 
technical solution you wanted would require eliminating that use case, but 
in favor of an unspecified optimization.

Despite not knowing the specifics of your optimization, I do not believe 
that my statement that "I don't think you can do that kind of isolation" 
is borne of ignorance. 

You cannot see it from the machine-formatted minutes, but if you look at 
the underlying IRC minutes (http://www.w3.org/2007/06/13-forms-irc), you 
will see that we spent 1.5 hours on this one issue, which was a very large 
amount of time on this one issue, especially considering that we were 
facing on the order of 150 last call issues to discuss.  Considering that 
we made it through about half during the three days, it should be clear 
that your issue received many times the average amount of consideration. 
The minutes do not reflect the amount of effort that the working group 
devoted to your issue because it was a lengthy discussion about many 
complex aspects of the existing specification.

More specifically, my own claim that "I don't think you can do that kind 
of isolation" is based on the fact that I think can create legal XForms 
that will break your optimization without setting a calculated node to 
readonly false (if setting a calculated node to readonly false breaks your 
optimization).  This is based on the fact that the setvalue action can 
change a node marked as readonly.  Readonly is a property that is 
communicated to the UI via *notification* events.  Just as non-relevance 
does not make data nodes unavailable to XForms actions, readonly does not 
make them immutable via XForms actions.  Hence, I can use setvalue at will 
to manipulate a readonly calculated node just as if it were not readonly 
and manipulated by a UI binding.  Put another way, the readonly is 
information that affects the state of form controls only, which is in 
principle unrelated to the formation of computational dependency graphs 
and subgraphs and in practice unrelated due to the existence of the XForms 
setvalue action.

In conclusion, it would help to know more about your specific optimization 
since there *may* be some other way to accommodate it or something like 
it.  Either way, I hope you will understand from this email that the whole 
working group does take your feedback very seriously as you are an 
esteemed member of this working group even if this particular issue did 
not turn out according to your expressed preference.

Sincerely yours,
John M. Boyer, Ph.D.
STSM: Lotus Forms Architect and Researcher
Chair, W3C Forms Working Group
Workplace, Portal and Collaboration Software
IBM Victoria Software Lab
E-Mail: boyerj@ca.ibm.com 

Blog: http://www.ibm.com/developerworks/blogs/page/JohnBoyer





David Landwehr <david.landwehr@picoforms.com> 
Sent by: www-forms-editor-request@w3.org
06/13/2007 11:31 PM

To
John Boyer <xforms-issues@mn.aptest.com>
cc
www-forms-editor@w3.org
Subject
Re: Last call comment about readonly property with calculate (PR#45)







I have read through the minutes and it seems you don't even consider my 
request. The statement by John "I don't think you can do that kind of 
isolation." clearly displays the ignorance from which the decissions are 
made in the working group.

I accept the resolution simply because I give up.
/David

John Boyer skrev:
> We agree it was unclear, but we find that calculate merely defaults 
readonly to
> true, and that it can be set to false, and that there are use cases, 
namely
> default value. We tested the use case and found it works. We changed the 
note in
> 4.3.6 [ 
http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#evt-recalculate]
> and put an example in MIP for readonly [
> 
http://www.w3.org/MarkUp/Forms/specs/XForms1.1/index-diff.html#model-prop-readOnly]
>
> Please let us know if this resolution is acceptable.
>
> Thank you,
> The Forms Working Group
>
> 
>> Having the following:
>> <xf:model>
>>   <xf:instance>
>>     <data xmlns="">value</data>
>>   </xf:instance>
>>   <xf:bind nodeset="." readonly="false()" calculate="1"/>
>> </xf:model>
>>
>> It is not spelled out in the specification that it is possible to
>> override the default state when it has a calculate on it. The default
>> value is true() when the node has a calculate on it. On the other side
>> it is not specified that it is not allowed. I think it should not be
>> allowed since it is not clear when the value will be recalculated
>> because a node cannot take itself as dependent. E.g. an insert or 
delete
>> will recalculate the value even if the user has updated the value (this
>> must also happen if an insert happens in another instance). This could
>> be a problem for implementation which isolates the creating of
>> dependencies between instances.
>>
>> Best regards,
>> David
>>
>>
>>