ebXML Registry Reference Specification: Business Domain
ebXML Working Draft Issue
Date May 12, 2000
First Comment Review
Disposition
|
May 15, 2000 |
Close of 1st comment period |
May 29, 2000 |
Submit revision and disposition for 2nd commend
period |
July 29 , 2000 |
Close of 2nd comment period |
Aug 7, 2000 |
Submit final revision with comments and disposition |
Aug 11, 2000 |
Comment number |
Submitter name |
Date submitter |
Start line |
End Line |
Description of comment or cut and paste section |
Resolution of comment |
1 |
Jarry Eric |
05/17/2000 |
|
|
General comments about word format. |
Changed figure numbering to be automatic |
2.1 |
Mary Kay Blantz |
5/18/2000 |
43 |
51 |
Indent for readability |
|
2.2 |
Mary Kay Blantz |
5/18/2000 |
53 |
58 |
Indent for readability |
Changed |
2.3 |
Mary Kay Blantz |
5/18/2000 |
60 |
65 |
Indent for readability |
Changed |
2.4 |
Mary Kay Blantz |
5/18/2000 |
66 |
68 |
Indent even further for readability |
Changed |
2.5 |
Mary Kay Blantz |
5/18/2000 |
69 |
69 |
Indent for readability |
Changed |
2.6 |
Mary Kay Blantz |
5/18/2000 |
70 |
70 |
Indent even further for readability |
Changed |
2.7 |
Mary Kay Blantz |
5/18/2000 |
105 |
105 |
Change "interfaces" to "interface" |
|
2.8 |
Mary Kay Blantz |
5/18/2000 |
123 |
125 |
Include the survey or statistical results and which users,
associations, consortia, and standards organizations that prove the mentioned
distrust. |
Removed the comment about survey etc… |
2.9 |
Mary Kay Blantz |
5/18/2000 |
145 |
145 |
Start the sentence with "The ebXML…" |
Changed |
2.10 |
Mary Kay Blantz |
5/18/2000 |
152 |
152 |
Change "as according to" to " in accordance
with" |
Changed |
2.11 |
Mary Kay Blantz |
5/18/2000 |
183 |
183 |
Remove the first "other" |
Changed |
2.12 |
Mary Kay Blantz |
5/18/2000 |
219 |
219 |
Put "the" between "in" and "past" |
Changed |
2.13 |
Mary Kay Blantz |
5/18/2000 |
231 |
231 |
Unless the word "purchase" has some special
meaning in this context, I'm not sure why we are including this. Perhaps change the word
"purchase" to "obtain" which may or may not imply some
financial transaction. |
Changed |
2.14 |
Mary Kay Blantz |
5/18/2000 |
237 |
237 |
Change "Common Business Object" to "Core
Component" |
The group doesn’t agree with this change and will revisit
this suggestion. |
2.15 |
Mary Kay Blantz |
5/18/2000 |
243 |
243 |
Change "each semantic unit" to "one or more
semantic units" |
Changed |
2.16 |
Mary Kay Blantz |
5/18/2000 |
266 |
266 |
Is "knows" valid in this context? Maybe "is programmed to" instead? |
Changed the word to “can” |
2.17 |
Mary Kay Blantz |
5/18/2000 |
268 |
268 |
Must the developer be a "team", or is an
individual possible? |
Didn’t change since a team can be a team of one |
2.18 |
Mary Kay Blantz |
5/18/2000 |
290 |
293 |
Domain Use Case provides an overall, high-level context of
the business problem, a boundary to the scope of the functionality, and
traceability to the actors involved as defined in the Actor Relationships
hierarchical diagram. It does not
delve into how the system performs the tasks on a functional basis. |
Changed to the sentence suggested by Mary Kay |
2.19 |
Mary Kay Blantz |
5/18/2000 |
316 |
316 |
Add "schemas" to the list. |
Changed |
2.20 |
Mary Kay Blantz |
5/18/2000 |
324 |
325 |
Must the RA communicate with the SO if changing the
submission? I say yes.
And the rules to do so must be documented somewhere. |
The rules will be defined in Part 2 Document |
2.21 |
Mary Kay Blantz |
5/18/2000 |
348 |
351 |
The sentence needs work, but I can't fix it. |
Changed
to the following: The
ISV is a special type of Corporation that can, through its association to the SO, register and submit
software components that have been constructed, tested, and certified to be
compliatnt to the retrieved specification(s). The components will then be available for download by the Business
User. |
2.22 |
Mary Kay Blantz |
5/18/2000 |
354 |
354 |
The use of "good" implies that "bad"
specifications may be found in the R & R's. |
Removed the word “good” |
2.23 |
Mary Kay Blantz |
5/18/2000 |
383 |
383 |
An example of a reason for an ebXML-compliant Business
Application to interact with an ebXML Registry. |
Changed |
2.24 |
Mary Kay Blantz |
5/18/2000 |
400 |
400 |
and/or 2) |
Changed |
2.25 |
Mary Kay Blantz |
5/18/2000 |
403 |
404 |
Remove this sentence.
Any actor anywhere might charge for anything. |
Changed |
2.26 |
Mary Kay Blantz |
5/18/2000 |
409 |
409 |
Change "cases" to "case" |
Changed |
2.27 |
Mary Kay Blantz |
5/18/2000 |
417 |
417 |
Line 414 says optional, but 417 says optional or
mandatory. Are these in conflict? |
Changed word from “These” to “The” |
2.28 |
Mary Kay Blantz |
5/18/2000 |
452 |
453 |
Change "common business objects (Core
Components)" to Core Components |
The group doesn’t agree with this change and will revist
this suggestion |
2.29 |
Mary Kay Blantz |
5/18/2000 |
494 |
495 |
The sentence needs something. Maybe just remove "what" in line 495 |
Changed |
2.30 |
Mary Kay Blantz |
5/18/2000 |
501 |
505 |
If TRP in line 505 is Transport, Routing and Packaging,
then line 501-502 needs to have Transport, Routing and Packaging instead of
Transport, Package, and Routing. |
Changed |
3.1 |
Gait Boxman |
5/18/2000 |
145 |
149 |
There is no point in referring to W3C separately. ISO,
IETF and others are equally important. Yes Xlink and Xpointer are key specs.
So are ISO11179, TCP/IP, DSN, HTTP, SMTP and others. Proposed change: either
remove text, or change to something like: In the design and implementation, the ebXML Registry and
Repository project team will consider any existing or appearing standards
from accredited standards bodies that it finds of value to the project. |
Modify sentence to include part of the proposed change |
3.2 |
Gait Boxman |
5/18/2000 |
125 |
153 |
I believe we are highly dependent on ISO11179. It should
be referenced here. Proposed change: Add a
short description of ISO11179 and how we depend on it. Something along the
lines of: ISO specification 11179 defines a general model for
registries and repositories and the roles and responsible parties with
respect to those systems. This ISO specification has been used as a starting
point for the ebXML Registry and Repository specifications. EbXML compliant
registries and repositories are also compliant with ISO 11179 rulings. |
Added the first two sentences proposed by line 125. |
3.3 |
Gait Boxman |
5/18/2000 |
151 |
153 |
This is orthogonal to the project itself. Whether we print
the model on paper, include it in a Word (or PDF) document or PowerPoint
presentation, display it with Rose, or generate an XMI representation has no
significance to the model or the project itself. Proposed change: remove text. |
|
3.4 |
Gait Boxman |
5/18/2000 |
154 |
162 |
Mixing glossaries and terminologies from four different
half-related sources is bound to cause problems with multiple definitions and
uses. We need a way to resolve that problem. I can see two options: we either
set a prevalence order on the sources, or do build our own glossary. Proposed change: Either
augment the text with: When terms are defined in more than one of those
documents, the document that appears first in the above list will prevail. Or change line 162 to: The ebXML Registry and Repository project team will create
it’s own glossary which will be merged into the overall ebXML glossary. |
Changed to add proposed line of text for line 162 |
3.5 |
Gait Boxman |
5/18/2000 |
163 |
404 |
As argued earlier in the Brussels meeting, I have serious
problems with the number of different actors and the roles they play against
the registry and repository. Due to the extensive text, I’m not going to
propose a complete new text. However, I would like to see that the actor
relationship diagram and use case diagram are restricted to actors, which
play a role at the system boundary and use cases within the system. Proposed change: The best guideline here would be the business domain use
case at line 300, with the following modifications: |
|
3.5.1 |
Gait Boxman |
5/18/2000 |
|
|
Remove ‘Create Map between Specifications’ and ‘Develop
Software Components’. These are not use cases/activities within the registry
and repository system, but activities for parties external to the repository.
These parties might use the repository to retrieve and/or submit items
through the appropriate use cases. (Of course, this also removes the related
actors from the diagram, unless you want to link them to Retrieve Items, in
which case they will be removed in the next step.) |
|
3.5.2 |
Gait Boxman |
5/18/2000 |
|
|
Remove ‘Registration Authority’ as an instantiating actor
for ‘Register Company’ and ‘Classify Submissions’, as the ‘Submitting
Organization’ already is the instantiating actor. You could leave the link as
a related actor. |
For debate: Viewpoint/ use case style issue |
3.5.3 |
Gait Boxman |
5/18/2000 |
|
|
Remove all links between actors and ‘Retrieve Items’,
except for ‘Guest User’, as they all play the same role. If needed,
‘Registration Authority’ could be linked as a related actor (e.g. for logging
purposes). |
Maybe; the ebXML Business Application should
remain as it produced/ instantiated by the Independent Software Vendor, and
not part of the inheritance hierarchy. |
3.5.4 |
Gait Boxman |
5/18/2000 |
|
|
The link between “Registration Authority” and “Store
Items’ is also strange. In my view you can only store items that are properly
submitted. Even if the RA does it on it’s own, it plays the role of an SO,
through the ‘Submit Items’ use case. |
This is a viewpoint issue. The RA grants public access to the
submitted items. This may not be an
automatic process to submit an item and let the world be instantly aware of
its availability. |
3.5.5 |
Gait Boxman |
5/18/2000 |
|
|
Next are some proposed renaming exercises: Rename ‘Register Company’ to ‘Register Submitting
Organization’, an SO does not need to be a company. It could be a private
person, non-profit or governmental organization. |
OK |
3.5.6 |
Gait Boxman |
5/18/2000 |
|
|
Rename ‘Guest User’ to ‘Information Consumer’ or ‘Ordinary
User’. It is really the ordinary user who retrieves information for the
repository. |
Disagree, but we understand your concern. This was debated in Brussels and the group
felt the name, as awkward it sounds, is common from an administrative point
of view. |
3.6 |
Gait Boxman |
5/18/2000 |
326 |
338 |
Note that in the ebXML Requirements document, which was
approved in Brussels, it is stated that ebXML compliant registries and
repositories have open and perpetually free access. In my mind, this
precludes the need to authenticate for read access with either system, nor a
trust requirement between a registry and repository. Proposed change: remove lines 335-338 |
Disagree, RA may impose a authentication requirement. |
4.1 |
Bob Miller |
5/23/2000 |
99 |
99 |
Labels inside chart overlap schedule bars, reducing
readability |
This graph is from UN/CEFACT and will be communicated to
UN/CEFACT the comments on readability |
4.2 |
Bob Miller |
5/23/2000 |
168 |
169 |
Text in these diagrams range from unreadable to
acceptable. Ideally, the font size would be consistent across all diagrams,
and would be the font size used in the document proper. |
Okay Font will be increased. |
4.3 |
Bob Miller |
5/23/2000 |
154 |
162 |
All terms and definitions, regardless of their source,
should (and are intended to appear in an ebXML Glossary, which should be
referenced in this section. |
Changed |
4.4 |
Bob Miller |
5/23/2000 |
246 |
247 |
Change ‘and / or’ to ‘and of the’ |
Changed |
4.5 |
Bob Miller |
5/23/2000 |
300 |
325 |
The diagrams and text suggest that items may only be
stored into a Repository via the Registry.
The submitting organization should have the option of storing data in
the Repository prior to Registry, and then instructing the Registry to
reference the stored data in the Repository, rather than creating a new
Repository copy of the data. |
The UML model use cases don’t implied whether it is
physically registered in the registry from the repository or SO. |
4.6 |
Bob Miller |
5/23/2000 |
312 |
312 |
I am concerned here with the use of the word
‘package’. IMO, what is being
submitted is an object (represented as a stream of bits/bytes). This object could contain other objects
which also are to be Registered. 1)
Change ‘Submit Items’ use case to ‘Submit Object’ Use case
wherever it appears. 2)
Change 312-318 to: The Submit Object use case provides the
ability to submit an object to the Registry.
The use case may request storage of the Object in a Repository to
which the Registry has ‘write authorization’, or may request that the Registry
reference the object as previously stored in a Repository. A record of the object is created in the
ebXML Registry (hence the term “register”).
Imbedded objects within the submitted object may also be registered in
the Registry, at the discretion of the SO.
(For example, if the submitted object is an XML DTD, the SO may also
request registration of each XML Element defined within the DTD.) The RA … |
Will be discussed more on the list-server |
4.7 |
Bob Miller |
5/23/2000 |
361 |
361 |
Change ‘the ebXML Registry’ to ‘an ebXML Registry’ Change ‘an ebXML Repository’ to ‘has “create” access to at least one
Repository.’ |
Changed |
4.8 |
Bob Miller |
5/23/2000 |
362 |
363 |
‘a complete background check’? Suggest: The RA may authorize or reject the SO’s
registration request that was initiated in the Register Company use
case. Which action is taken by the RA
is beyond the scope of the ebXML specification. |
Changed |
4.9 |
Bob Miller |
5/23/2000 |
366 |
371 |
Surely the SO should have some say in whether public
access to the submitted object is to be granted. An SO may wish for example to provide public access to a digest
of the submission, but retain control over access to the object itself. |
What is the proposed change? |
4.10 |
Bob Miller |
5/23/2000 |
414 |
416 |
There cannot be additional services within the ‘Registry’
<<Service>> package that may be performed by the Repository. The services of the Repository must be
fully defined by the Repository, though of course some such services may be
optional. Furthermore, I cannot think
of an optional Registry service that requires a Repository service not
already required by the basic Registry service. |
Changed the paragraph to explain better the meaning. |
4.11 |
Bob Miller |
5/23/2000 |
440 |
440 |
I would drop ‘and store them in the Repository’, as
earlier text in this document already attributes the ability to store them in
the Repository to the Registry services.
As currently worded one might read that the SO must perform two
independent operations (and I would suggest in improper order). |
Changed |
4.12 |
Bob Miller |
5/23/2000 |
570 |
571 |
The relationships among ‘Work in Progress package’ (e.g., line 448)
and ‘Workflow Service’ and ‘Library Control System’ are not clear in 405-466. Also, is there a relationship between workflow service and
library control service? |
This is implied by the dependencies between the work in
progress and registry as well as the repository to the registry. |
4.13 |
Bob Miller |
5/23/2000 |
495 |
495 |
Delete ‘what’ |
Changed |
4.14 |
Bob Miller |
5/23/2000 |
503 |
505 |
‘the ebXML Registry and Repository is’ suggests one
inseparable service, whereas the Repository is an independent service. Suggest: Since the ebXML Registry and Repository are also
ebXML-compliant Business Appllications, communication with these services
relies upon the ebXML TRP messaging services. |
Changed |
5.1 |
Melanie McCarthy |
5/23/2000 |
73 |
89 |
After reading the various sections that are being brought
forward, I am concerned that the specific information that I believe that
ebXML has promised to deliver is not being included. Specifically, my expectation was to see
published a recommended architecture model for a repository to be ebXML
compliant. |
We have three more specifications to produce that will
cover this in detail. |
5.2 |
Melanie McCarthy |
5/23/2000 |
166 |
167 |
If this is a web of Registries & Repositories, I would
expect that various security models would be recommended – not to allow for
the security policy to be at the discretion of the registration authority. |
Security will be expanded in detail in part 3 and 4
specifications. |
5.3 |
Melanie McCarthy |
5/23/2000 |
173 |
174 |
It appears that the “guest user” must act on behalf of an
organization. Can he not also
represent an individual? |
The Arrow is a UML inheritance association. |
5.4 |
Melanie McCarthy |
5/23/2000 |
190 |
190 |
I don’t understand the purpose of the Business User that
identifies him as different from the Business Domain Expert. |
Changed and remove Business user |
5.5 |
Melanie McCarthy |
5/23/2000 |
216 |
218 |
The Business Process Analyst should not ‘challenge’ the
Business Domain Expert, rather he could provide constructive recommendations
regarding alternative methodologies.
But ultimately he should not be in a position to override the need of
the Business Domain Expert.
Alternatively, if this concept is pursued, there should be an appeal
process where the Business Domain Expert’s point of view can be appealed. |
Changed paragraph |
5.6 |
Melanie McCarthy |
5/23/2000 |
322 |
324 |
Again there should be an appeal process, if the SO
believes that the Registration Authority has acted erroneously and
re-classified information, there should be a process to appeal the decision. |
This will be defined in Part 2 specifications |
5.7 |
Melanie McCarthy |
5/23/2000 |
353 |
356 |
Information that may be stored in the repository may be
very confidential, potentially providing a competitive solution. Thus it should not be accessible by the
ISV. |
Disagree since the RA and SO can control access rights |
5.8 |
Melanie McCarthy |
5/23/2000 |
363 |
363 |
Is there a time frame in which the SO must complete the
background check? |
Changed |
5.9 |
Melanie McCarthy |
5/23/2000 |
366 |
366 |
Is only information that is ‘public access’ allowed in the
repository? |
No what is the proposed change? |
5.10 |
Melanie McCarthy |
5/23/2000 |
476 |
478 |
As more information is cached in memory, is there a
concern regarding responsiveness of the registry? |
No what is the proposed change? |
6.1 |
Jacques Littre |
5/25/2000 |
168 |
264 |
1 - Figure 2 - line 168 &
section 2.4 line 255 to 264 The figure 2 as well as
the section 4 is not clear about what a "Industry Consortium" is
exactly in regards to the "corporation" actor whilst it seems natural that
there is a clear link between Industry Consortium and Corporation (there is
no Consortium possible without corporation
members). So, we
propose therefore to add in the figure 2 an UML aggregation relationship
between the Industry Consortium actor and the Corporation actor. On the
aggregate side, the multiplicity should be [0..N] whilst on the Corporation
side, the multiplicity should be [2..N]. An second aggregation relationship
should also be reflexive on the Industry Consortium itself (since a consortium
may have consortia members) with a [0..N] multiplicity adornament. This solution also illustrates the
fact that a Corporation may act (as Registration Authority) on behalf of a
consortium (which is very
often the case anyway). |
Changed |
6.2 |
Jacques Littre |
5/25/2000 |
175 |
260 |
This
should also impact figure 3 (line
175) and figure 7 (line 260) which should show those aggregations. |
Changed |
6.3 |
Jacques Littre |
5/25/2000 |
259 |
259 |
Section 2.4 (line 259) should also contain a new paragraph
explaining that "The industry consortium is made of corporations and/or
other industry consortia". And that " Standardization organizations
are considered as an Industry Consortium too". (Otherwise they probably should appear as a new
actor in the figure 2 as subclass of Registration Authority). |
Changed |
6.4 |
Jacques Littre |
5/25/2000 |
168 |
168 |
Still on figure 2 (all actors) Shouldn't we align the
Actors and roles illustrated in this figure with those defined in the Eco
framework ? Why do we
diverge from Eco on this ? |
The group doesn’t agree on this and we should revisit. |
7.1 |
Martin Bryan |
6/06/200 |
234 |
238 |
If the Technical
Assessment actor is supposed to "identify semantic overlap" it is
important that this actor is well versed in the application of the techniques
developed by the Core Components team to remove semantic overlap within core
components. He should also be aware of how overlap will affect core
components, and the procedures used for reviewing core components. |
Correct. No
change. We called the Core Component output common business objects. We also understand there isn’t an
agreement whether a Core Component is a Common Business Object, however, this
is the main intent as you state. |
7.2 |
Martin Bryan |
6/06/200 |
273 |
282 |
The
role of Software Agents in the business process is likely to be difficult
to specify. What is important is that the repositories and registries team define
APIs that can be used by software agents, and means of distinguishing when a
request comes from an agent and when it comes directly from a Guest user.
(Indirected requests should normally be given lower priority as they are
asynchronous.) |
Yes. No change. It
is agreed that the agents requests be tracked versus that of other
requests. That is why we included it
as a separate actor. Others still
feel that these actors could be consolidated and you are validating our
initial thoughts. |
7.3 |
Martin Bryan |
6/06/200 |
307 |
320 |
The process by which
Submitted Items may be approved by Submitting Organizations needs to be
clarified by reference to lines 367/369. |
We added “…by the RA (see.3.4)” on line 307. |
7.4 |
Martin Bryan |
6/06/200 |
336 |
337 |
How can a guest user be
"authenticated to the Registry".? |
Authentication is part of the security service. The Registry service will receive requests
only if the session is granted by the security service. Some systems require simple authentication
to track who has visited a site. The
RA is responsible for defining the rules. Line 336: Through the use of the Security Services and dependent on
the access rights defined by the RA, the Guest User …” |
7.5 |
Martin Bryan |
6/06/200 |
386 |
388 |
These lines (except for
the first word of Line 386) are not appropriate in an ebXML specification. |
Sentence deleted. |
7.6 |
Martin Bryan |
6/06/200 |
430 |
465 |
Technical
Specifications should follow Work in Progress rather than precede it.
Rationale: Work in progress is by definition something you do not want the
public to see. At the point you want to make it visible then you declare it
as a Technical Specification, and make it accessible to people browsing the registry. Only when
a public comment period has passed should it become a Submission that can
then be approved as a registered object. |
No change. The arrows you are referring to are UML
dependency relationships and not flow directions. We understand this could be confusing at first glance. Technical Specifications are dependent on
the results of Work in Progress.
Technical Specifications are for public view and work-in-progress is
restricted by user authorization. |
7.7 |
Martin Bryan |
6/06/200 |
526 |
527 |
The Subscribe to Related
Repository use case mentioned here does not appear to be defined in this
document, but no reference is made to another document. |
Thank you.
Sentence deleted. |
7.8 |
Martin Bryan |
6/06/2000 |
General |
|
There no mechanism by which a Guest User can request information
about technical specifications or other proposed or accepted change to an
object that he has determined Is needed as part of an application. Related to this, there is no mechanism by
which ISVs can identify which Software Agents are related to which repository
objects, or to record which of their software agenets may need to be updated
when a new version of a particular object is proposed / created. |
1) Figure 1 clearly shows there is a use case allowing the
guest user to retrieve items. The
detail behind this Use Case will be elaborate in Part 2 and Part 3 of the
ebXML Registry and Repository specification. |
7.9 |
Martin Bryan |
6/06/2000 |
General |
|
There is no model of the APIs that will be required too provide
interfaces between the identified actors and the repository. |
Correct. That will
be defined in Part 4: Design. |
8.1 |
Mike Rawlins |
6/06/200 |
3 |
3 |
Issue or rationale: Inconsistent versioning between different specifications, yours
starts at 1.0. Inconsistent with
"ebXML Specification Approval Process", May 30, 2000, section 2.1 Suggested change: Set next version to 0.x, with note of
explanation. Set final published
version at 1.0 |
No Change. At the time of the draft of this specification,
1) there was no Approval Process and 2) we felt that the document when
released would be called v1.0, and that the issue date was sufficient to
ensure that people were clear of which version they were looking at. |
8.2 |
Mike Rawlins |
6/06/200 |
126 |
153 |
Issue or rationale: Omission of dependency on "ebXML Requirements Specification" Suggested change: Add reference to dependency on
"ebXML Requirements Specification" |
OK. Sentence add
before Line 127: |
8.3 |
Mike Rawlins |
6/06/200 |
141 |
143 |
Issue or rationale: Inappropriate technical detail for functional specification
document dealing with business domain Suggested change: Strike sentence "Therefore the XML interfaces
... XML document responses." |
OK. Despite the
fact that there were many business people in the meeting when the sentence
was drafted. |
8.4 |
Mike Rawlins |
6/06/200 |
154 |
154 |
Issue or rationale: These are more references than "Conventions and
Terminology". Suggested change: Change title to "References" |
OK. |
8.5 |
Mike Rawlins |
6/06/200 |
157 |
160 |
Issue or rationale: Incomplete references Suggested change: Include URLs to references |
No change. Adding a
URL creates a maintenance nightmare if URL changes over time. |
8.6 |
Mike Rawlins |
6/06/200 |
163 |
404 |
Issue or rationale: Section 2 on Actor Relationships and Section 3 on Domain Use Case
seem to be based on the *proposed* development process and organization
(distinct from TMWG modeling methodology) that TMWG/SITG have been working on
for CEFACT/EWG and X12. That process
and organization have yet to be approved.
Even if they are, there are other organizations that will wish to host
a repository that implements this specification that do not have the same
actors or use cases. These sections
are too detailed around one specific business domain (this proposed
organization and process), and leave out details of other similar domains
which would also use the repository. Suggested changes: Rewrite all actor descriptions to indicate that the actors
detailed are roles or functions and not necessarily specific unique entities.
[NOTE: This may be implied by UML or
the modeling methodology, but if so it needs to be made explicit in this
document.] This is particularly
important for the actors within the registration authority. |
The UN/CEFACT methodology is based on Rational’s Unified
Process (RUP) which is a common framework by which methodologies can be
defined. The UML Profile used within
the context of this methodology was complete and is being used. |
8.7 |
Mike Rawlins |
6/06/200 |
166 |
166 |
Issue or rationale: Grammar Suggested change: Should be "role based" |
OK. |
8.8 |
Mike Rawlins |
6/06/200 |
172 |
174 |
Issue or rationale: Unclear about just what a "guest user" is. Suggested change: Add further clarification about what a guest user is. |
Added to Line 174: “…organization, but has very limited responsibilities and
privileges, e.g., access rights.” |
8.9 |
Mike Rawlins |
6/06/200 |
177 |
179 |
Issue or rationale: Unclear about scope of "as defined by ISO 11179" and
definitions. User is forced to refer
to ISO 11179 for definition(s). For
example, it is unclear how non-profit organizations and government agencies
fit into the Actor Relationship class hierarchy. Are these "corporations"? Suggested change: Include specific ISO 11179 definitions. |
Changed: (for
definitions of SO and RA, see ISO 11179). |
8.10 |
Mike Rawlins |
6/06/200 |
182 |
186 |
Issue or rationale: Same as 177-179, still unclear about what specifically a
"corporation" is in this context. Suggested change: Add more explicit definition either here
or in previous section. |
Replace Line 183 with: A corporation is a collection is
individuals that work together to achieve a common goal through its policies,
processes, and resources. |
8.11 |
Mike Rawlins |
6/06/200 |
197 |
202 |
Issue or rationale: ISV actor description excludes development groups within a
"corporation". Suggested change: Change "ISV" actor to
"developer" actor, and include mention that this could be an internal
corporate function rather than an external entity. This, again, is more along the lines of making these actors'
roles or functions, rather than specific entities. |
Changed to Software Developer |
8.12 |
Mike Rawlins |
6/06/200 |
198 |
198 |
Issue or rationale: ISV produces software that meets the needs of the corporation,
not the business domain expert. Suggested change: Change to "meets the needs of the corporation, as expressed
by the business domain expert". |
OK: Changed to The Software Developer produces software
that meets the needs of the corporation, as expressed by the business domain
expert. |
8.13 |
Mike Rawlins |
6/06/200 |
198 |
202 |
Issue or rationale: Sentence "The business needs of the Business Domain Expert ...
Registration Authority)." Suggest specific entities and processes. Also, sentence is too long. Suggested change: Strike, or qualify with "depending
on the organization and processes, these needs *may* be communicated"
rather than "are communicated". |
The model will be changed to add the
unidirectional associations (which are based on the RUP framework), and add
verbiage to explain this. |
8.14 |
Mike Rawlins |
6/06/200 |
204 |
206 |
Issue or rationale: Same issue as 177-179. Unclear
about just what a "Registration Authority" is. Suggested change: Include more specific definition of "Registration
Authority". |
Per 11179 |
8.15 |
Mike Rawlins |
6/06/200 |
214 |
215 |
Issue or rationale: Several issues with the sentence "These role players ...
Process Modeling". a) Unclear about whether this is the "UN/CEFACT/TMWG UML
Profile and Methodology" mentioned in line 135, or a different document. b) There is an implied dependency on this methodology that is
not mentioned in section 1.2 c) Same issue as general comment on lines 163-404 regarding
sections 3 and 4. Other organizations
and processes need to be supported beyond just this one. Suggested changes: a) Make consistent document references, and include in section
1.3 b) Include dependency in section 1.2 c) Same comment as 163-404. |
a) Changed to UN/CEFACT/TMWG UML Profile and Methodology
and role players to Actors, b) now referenced, c) don’t agree that they
should be specified here versus Part 2. |
8.16 |
Mike Rawlins |
6/06/200 |
204 |
254 |
Issue or rationale: ebXML works with "core components", not "common
business objects", and they are not the same. Suggested change: Change all occurrences of "common business object" to
"core component", and remove parenthetical references to "core
component" |
This is still in debate.
No change. |
8.17 |
Mike Rawlins |
6/06/200 |
204 |
254 |
Issue or rationale: Same issue as general comment lines 163-404 sections 2 and
3. Too specific about one particular
type of organization and process. Suggested change: Make more generic. |
Without specific recommendations, we find this task
request to be inappropriate. In
addition, specify this high level detail about actors is appropriate as it
leads very well into Part 2: e-Business Requirements. |
8.18 |
Mike Rawlins |
6/06/200 |
255 |
259 |
Issue or rationale: Unclear about definition of "Industry Consortium". Suggested change: Add more detailed definition of "Industry Consortium" |
See SWIFT’s comments and changes. |
8.19 |
Mike Rawlins |
6/06/200 |
306 |
307 |
Issue or rationale: Sentence "The SO must be ... use case." Implies a
specific policy of the registration authority. Out of scope for this document. Suggested change: Change to "Depending on RA policies, the SO may be..."
rather than *must* be. |
Changed. |
8.20 |
Mike Rawlins |
6/06/200 |
322 |
324 |
Issue or rationale: It is clear that the functionality needs to supported. However, this implies specific policy and procedure of RAs which might use the
R&R, which might not be accurate. Suggested change: Add qualification to make this dependent
on policies and process of the RA. |
Changed “can” to “may”, as it is important to allow this
classification, since it is very probable that an RA may misclassify an
object. |
8.21 |
Mike Rawlins |
6/06/200 |
336 |
337 |
Issue or rationale: "May" unclear. Suggested change: Add qualification "Depending on
policies of the registry host..." |
Others had the same comment. Changed. |
8.22 |
Mike Rawlins |
6/06/200 |
348 |
351 |
Issue or rationale: Several a) Implies a testing and certification process, that may or may
not exist or be used b) States that software components (presumably executable code)
can be registered and downloaded.
This adds technical and business complexity to the operation of the
R&R. Suggested change: Strike paragraph. |
a) We think this is an open issue as this was debate at
the first San Jose meeting and there seemed to be positive discussion in
favor of certification, b) Why is this more complex? |
8.23 |
Mike Rawlins |
6/06/200 |
354 |
354 |
Issue or rationale: Unclear about what makes specifications "good" Suggested change: Strike "good" |
Already changed. |
8.24 |
Mike Rawlins |
6/06/200 |
358 |
359 |
Issue or rationale: Extraneous business detail not appropriate for the scope of this
document. Suggested change: Strike |
OK. |
8.25 |
Mike Rawlins |
6/06/200 |
363 |
363 |
Issue or rationale: Sentence about background check implies policies of RA that may
or may not be relevant. This is
irrelevant, unrelated detail beyond the scope of this document. Suggested change: Strike, or change to append to the previous
section "The RA authorizes ..., according to criteria established by the
RA". |
See comment from Robert Miller and changes. |
8.26 |
Mike Rawlins |
6/06/200 |
367 |
371 |
Issue or rationale: Too much detail about a process based on assumed or hypothetical
policies of the RA. Suggested change: Strike, or change to mention that some
RAs may have policies that withhold public access to submissions until
reviewed by the RA. Therefore,
functionality to limit access until review is performed must be supported. |
We changed the text to include “may” in several sentences. |
8.27 |
Mike Rawlins |
6/06/200 |
403 |
404 |
Issue or rationale: Extraneous business detail not appropriate for the scope of this
document. Suggested change: Strike |
Others made similar comments. Deleted. |
8.28 |
Mike Rawlins |
6/06/200 |
435 |
435 |
Issue or rationale: reference is to an "ebXML-compliant" specification,
and not to an RA standard or specification that may or may not be
"ebXML-compliant". The
purpose of this document seems to be to enable a network of interoperable
registries and repositories, i.e., R&R that are
"ebXML-compliant", and *NOT* an ebXML R&R. This sentence implies the latter. Suggested change: Change "an ebXML-compliant
specification" to an "approved standard or specification" |
OK. Changed. |
8.29 |
Mike Rawlins |
6/06/200 |
447 |
465 |
Issue or rationale: Same as 435-435. The
R&R is not an "ebXML R&R" designed to support "ebXML
Technical Specifications" or items that meet "the requirements of
the ebXML Business Process Metamodel", it is an R&R that is
ebXML-compliant that may or may not contain any ebXML artifacts, or artifacts
that are compliant with ebXML. Scope
here is way too narrow. The goal
should be a single network of linked R&Rs that contain artifacts that are
both compliant and non-compliant with ebXML, with compliance status noted as
an attribute of the artifact. Suggested change: These lines need major re-write to
address the issues noted. Make them
more generic, and not just applicable to the development of ebXML or ebXML-compliant
artifiacts. Change "technical
specification" to "public item". This makes the package more generic, and makes the submission
of private items not intended to be subject to a standards development
process or compliance criteria more straightforward. |
We don’t see what the problem is with this verbiage as
several scenario examples are represented (i.e., ebXML Specification and
non-ebXML Specification which covers everything else). The point of this section is to show that
User Interface workflows exist, which will be elaborated on in Part2. |
8.30 |
Mike Rawlins |
6/06/200 |
477 |
478 |
Issue or rationale: "especially ... cached in memory." is an
implementation detail inappropriate for the scope of this document. Suggested change: Strike. |
OK. Entire
sentence deleted. |
8.31 |
Mike Rawlins |
6/06/200 |
490 |
497 |
Issue or rationale: "check-out/check-in" implies that developers will use
the repository in the same fashion as a source code library. Unclear about the items that are checked-out
and checked-in. It makes much more
sense to make the developer's access read-only, and any developed artifacts
to be handled through the submission services. Having a quasi-public R&R
serve as a source code library for private entities adds unnecessary
technical and business complexity.
Some R&Rs may wish to provide this facility, but it is clearly
beyond the scope of this specification. Suggested change: Strike references to
check-out/check-in. Strike paragraph
494-497. |
This was debated heavily in Brussels and the team
comprised by making this an optional service under repository, however the
interfaces still should be defined. No change except for the deleted word “what” in line 495. |
8.32 |
Mike Rawlins |
6/06/200 |
516 |
518 |
Issue or rationale: This is an implementation detail inappropriate for the scope of
this document. Suggested change: Strike. |
This debated heavily in Brussels the team felt that the
simplest case should be detailed. No change. |
8.33 |
Mike Rawlins |
6/06/200 |
567 |
569 |
Issue or rationale: Similar to others.
R&R needs to support multiple organizations and processes, not
just ebXML. Suggested change: Change "events; e.g. ...
model" to "events, dependent on policies and procedures of the Registration
Authority." |
Added: …events, dependent on the policies and procedures of the RA; e.g.,
….” |
8.34 |
Mike Rawlins |
6/06/200 |
577 |
583 |
Issue or rationale: Unclear about what this appendix is. These appear to be references with embedded URLs, but they are
not accessible as links in the PDF.
They also appear to be somewhat duplicative of the references in
section 1.3 Suggested change: Merge this list of references with those
in 1.3, and strike this appendix.
Include specific URLs after embedded URLs. |
Links are removed altogether. No URLs will be specified. |