RE: [wsbpel] Issue 11 - Proposal For Vote

["Satish Thatte" <satisht@microsoft.com>]

This is a far simpler and much more attractive proposal, in essence
because it does quite a lot to expand scenarios covered with very little
machinery.  In fact the machinery seems to consist of eliminating a
standard error and providing behavior in place of it.  I haven't spent
the effort to analyze it thoroughly, but to me it is by far preferable
to the original proposal.

Satish

-----Original Message-----
From: Chris Keller [mailto:chris.keller@active-endpoints.com] 
Sent: Thursday, April 28, 2005 2:10 PM
To: wsbpel@lists.oasis-open.org
Cc: 'Ugo Corda'
Subject: [wsbpel] Issue 11 - Proposal For Vote

TC Members,

As promised during yesterday's conference call here is my alternate
(amendment) to the Issue 11 Proposal.

Regards,
Chris Keller

--------------------------------------------------------------------

Proposal: The most requested form of data manipulation not currently
covered
by <copy> is the ability to incrementally create new complex messages in
order to invoke services or reply from a process.  This proposal
suggests we
add a create attribute to the existing query element.  When this option
is
set to "yes" and the query does not exist then the path will be created
(note if the path exists there is no need to create it and it is always
used
regardless of the state of this attribute).

<query queryLanguage="..."? create="yes|no"?>...<query>

When the create option is enabled and the target query is using the
BPEL/XPath binding the query will have the following restrictions:

1 - The path must use the abbreviated form.
2 - The path must not use '//' as that can lead to a non-deterministic
path.
3 - Predicates will be limited to integer values, which in XPath
indicates a
child number within the context. In addition the EII indicated by the
integer value must either exist or be at most one higher than the last
child
EII of that name within the parent.

If these conditions are not met a compliant implementation MUST throw a
bpws:createFailure fault.


Examples
--------

<copy>
   <from><literal>Widget1</literal></from>
   <to create="yes" variable="PurchaseOrder" part="PODetails">
      <query create="yes">
         /PODetails/po:items/po:item[$counter]
      </query>
   </to>
</copy>

When used against the following document with a counter value of 2 (note
XPath child selection counters start at 1 not zero) will produce a
bpws:createFailure.

<PODetails/>


However if used against a starting document of the form:

<PODetails>
   <po:items xmlns:po="...">
      <po:item>Widget0</po:item>
   </po:items>
</PODetails>

Will produce:

<PODetails>
   <po:items xmlns:po="...">
      <po:item>Widget0</po:item>
      <po:item>Widget1</po:item>
    </po:items>
</PODetails>


Newly created nodes are always appended as the next child of a parent.
Since the incremental build-up is the most important case let's not try
to
do a general purpose transformation language, if someone needs a new
ordering or needs to remove an element they can create a new variable by
combining assigns in the right order from an existing variable.

Before:

<PODetails>
  <po:shippingInfo>
     Some Address
  </po:shippingInfo>
</PODetails>

After execution of the sample <copy> with a counter value of 1:

<PODetails>
   <po:shippingInfo>
     Some Address
   </po:shippingInfo>
   <po:items xmlns:po="...">
      <po:item>Widget1</po:item>
   </po:items>
</PODetails>


Attributes will also be created through this mechanism, the following is
an
example for attribute creation:

<copy>
   <from><literal>10</literal></from>
   <to create="yes" variable="PurchaseOrder" part="PODetails">
      <query create="yes">
         /PODetails/po:items/po:item[$counter]/@amount
      </query>
   </to>
</copy>

Before:

<PODetails/>

After execution of the sample <copy> with a counter value of 1:

<PODetails>
   <po:items xmlns:po="...">
      <po:item amount="10"/>
   </po:items>
</PODetails>


-------------------------------------------------------

What this proposal does:

1 - Handles the ability to assign to new locations, which is what the
original issue 11 description states.
2 - Doesn't go into xupdate or generalized transformations (insert,
append,
remove, update type operations), which has been the black hole of issue
11.
3 - Easy to describe within the context of the current copy operation
text.
4 - Assuming resolution of issue 147 this would enable for each to
collect
result sets utilizing the index variable part of the 147 proposal.




---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  You may a link to this group and all your TCs in
OASIS
at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php