[Requests] OpenGIS Web Processing Service (RFPC 28) Response

Jim Greenwood JGreenwood at seicorp.com
Wed Feb 1 08:27:15 EST 2006


Part A is to be completed once per evaluator per comment submission.
Please iterate over Part B as needed. 

 PART A 
 ------ 
 1. Evaluator: James Greenwood, SeiCorp, Inc, jgreenwood at seicorp.com

  2. Submission: OGC Request 28: Web Processing Service (WPS): Request
for Public Comments  

 

 PART B 
 ------ 
 1. Specification Section number: [GENERAL, #] 

 2. Criticality: [MAJOR, MINOR, EDITORIAL, ETC.] 

 3. Comments/justifications for changes: [COMMENTS, SUGGESTIONS] 

Open Geospatial Consortium, Inc. OpenGIS(r) Web Processing Service.
document : OGC 05-007r4, 2005-09-16, Version: 0.4.1 review.

 

Comments:

 

1.      General: 

a.        Request OGC verify all OGC 05-007r4 Table and paragraph
references to OGC 05-008; dated 2005-11-22; Version: 1.0.0; Open GIS (r)
Web Services Common Specification.  Numerous errors were observed. For
example, "Note 1, subclause 9.2.1 references Table 21 when the correct
reference is Table 20.

b.      For all "Note" items, clarify what Table or subclause the Note
is applicable.

c.       Use of 'Unordered Lists' in Input and Output Processes, while
laudatory, will place excessive overhead on server and client portal
interface software to be robust in interpreting the incoming data
stream, transforming that content to internal required processing
structures, and loading to those internal operations.

2.	Specific:

	a.	Table 9, subclause 9.2.1 change to read

Identifier

Process identifier

Character String type, not empty

Value is process Identifier defined in ProcessOfferings section of
service metadata (Capabilities) document

One or more (mandatory)

One for each desired Process, unordered list

	b.	Subclause 9.3.1.a).  Incomplete statement.  Red
annotations by reviewer. "ComplexData" (such as XML or imagery), in one
of the following allowable combinations of format, encoding, and schema.
The value of this complex data structure can be (either) directly
encoded in the Execute operation request or ?? 
	c.	Table 12, subclause 9.3.1.  Recommend removal of "Zero
or one (optional)" from "Process Inputs: Multiplicity and Use" column as
it is in conflict with 'One (Mandatory)' based on Table 13 requirement.
	d.	Table 12, clarification 'd'. Recommend "The
"statusSupported" parameter value "true" is recommended when a process
takes a long time to execute."  be changed to The "statusSupported"
parameter value "true" is recommended when a process is not
instantaneous."  Elsewhere [05-007r4] clarifies 'long time' as more than
a minute.  The reviewer noted that normal web processing today provides
the client with estimated time to execute a download and should be
implemented in this application.
	e.	Table 12, clarification 'c'.  Acceptance of  the
recommendation for Table 12 clarification 'd' eliminates existing
clarification 'c' and recasts clarification 'd' as clarification 'c'.
	f.	Table 18, 'ComplexOutput', clarification 'c': Request
for clarification.  Why does the standard allow for an alternative
output form/format that complicates both client and server software
instead of the 'normal XML structure' even if the output is a 'single
output'?   Even that 'single output' can be voluminous and benefit from
direct client interpretation of the response.
	g.	Table 21 - Parts of SupportedCRSs data structure.
Request incorporation of reference to OGC standard for SupportedCRSs to
insure singular set of CRSs.
	h.	Table 27, Note 2.  Clarify wording.  Numerous use of the
word 'not' confuses interpretation.   "NOTE 2            The data type
of some parameters is specified as "Character String type, not empty".
In the XML Schema Documents specified herein, these parameters are
encoded with the xsd:string type, which does NOT require that these
strings not be empty. " 
	i.	Table 36, clarification 'b'.  Change 'should' to
'shall'.
	j.	Page 46, XML.  Correct example that contains an
incomplete sentence.  "<documentation>Identifies the unit of measure of
this literal input or output. This unit of measure should be referenced
for any numerical value that has units (e.g., "meters", but not a more
complete reference system). Shall be a UOM be identified in the Process
description for this input or output."   
	k.	Subclause 10.3.1.  Request definition of size limitation
for a GML document that the Web Processing Service would support in a
single transaction.  "For example, if a WPS process creates one GML
document (with no size/volume limitation?) as its output, that GML
document will be returned to the client as a direct response to the
Execute request."
	l.	Table 38.  Strongly recommend that clarifications 'a'
and 'b' be changed to 'shall' requirements.  The correct implementation
of these two functions provides a literal and legal history of data
processing necessary in current and future server/client geospatial
processing.
	m.	Table 41, clarification 'b'. "b     Recommended for use
when processes take more than a minute to complete."   Is this wall or
processor time?  Recommend this be changed to mandatory if the process
is not instantaneous.  Rationale: It is a normal occurrence in current
web download operations and a client expected result.
	n.	Subclause 10.3.4, Table 45.  Recommend inclusion of the
VersionNegotiationFailed exception report from Subclause 7.4 of [OGC
05-008].  Rationale: Geospatial data sets and associated processing are
not static and will have unique, and changing versions based on the date
of the process invoked..

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.opengeospatial.org/mailman/private/requests/attachments/20060201/5439492c/attachment.htm


More information about the Requests mailing list