Re: New protocol: XIOP (Corba GIOP using XML1.0 +HTTP1.1)

On Thu, May 11, 2000 at 11:16:27PM +0200, Anders W. Tell wrote:
> Hi everyone,
> 
> Im working on a protocol wich is not directly new, OMG GIOP. However
> this
> version uses XML 1.0 as encoding format and HTTP/1.1 as primary
> protocol.
> It is also possible to send raw CDR encoded IIOP messages..
> 
> One notable feature is that all encodings (incl. CDR) are based on a
> Grove-like
> information model called "Gren" (Swedish for 'tree branch')
> 
> <!-- Template -->
>  <tr>
> <td><a name="compare.XIOP"
> href="http://xiop.sourceforge.net">XIOP</a></td>
> <td> Financial Toolsmiths AB </td>
> <td><a href="mailto:anderst@toolsmiths.se">Anders W. Tell</a>
> (Financial Toolsmiths AB) </td>
> <td>Alpha 2Q 2000</td>
> <td>"http://xiop.org/"</td>
> <td>0 due in 1.0</td>
> <td>0 due in 1.0</td><td><ul>
> <li><a href="#layers_serialization">serialization</a><ul>
>     <li><a href="#grammerType_custom">custom</a></li>
>     <li><a href="#grammerType_automatic">automatic</a></li>
>     <li><a href="#grammerType_fixed">fixed</a></li>
>     <li><a href="#dataTyping_implicit">implicit (contextual)</a></li>
>     <li><a href="#dataTyping_explicit">explicit</a></li>
>     </ul></li>
> <li><a href="#layers_extensibility">extensibility</a> encoding
> stereotypes, encoding forms, OMG IDL constructs, XML document and
> fragments</li>
> <li><a href="#layers_skinnyness">skinnyness</a> raw IIOP (binary CDR
> encoding) instead of XML stream</li>
> <li><a href="#layers_protocol">protocol</a> <a
> href="http://www.omg.org"> GIOP</a></li>
> <li><a href="#layers_Idiscovery">interface discovery</a> <a
> href="http://www.omg.org"> Interface Repository</a></li>
> <li><a href="#layers_transactions">transactions (ACIDity)</a> <a
> href="http://www.omg.org"> Transaction service</a></li>
> <li><a href="#layers_procedure">remote procedure</a> <a
> href="http://www.omg.org"> OMG Corba</a></li>
> <li><a href="#layers_Bprocess">business process</a><a
> href="http://www.omg.org"> UML, Business Objects Domain Task
> Force</a></li>
> <li><a href="#layers_security">security</a><a href="http://www.omg.org">
> Security service</a></li>
> <li><a href="#layers_routing">routing</a><a href="http://www.omg.org">
> Event- and Notification services, Object Reference framework with
> bridges, routing etc</a></li>
> 
> <li><a href="#layers_oo">Object Oriented framework</a> OMG's OMA</li>
> <li><a href="#layers_lang_client">Language bindings: Client</a> C, C++,
> Java, ADA, Cobol, Smalltalk, (Python ,Perl,...)</li>
> <li><a href="#layers_lang_server">Language bindings: Server</a> C, C++,
> Java, ADA,  Cobol, Smalltalk, (Python,Perl,...)</li>
> <li><a href="#layers_instances">Instance concept</a></li>
> <li><a href="#layers_instance_discovery">Instance discovery</a> Naming
> Service, Trader Service</li>
> <li><a href="#layers_realtime">Realtime</a></li>
> <li><a href="#layers_service">Services</a> OMG Corba services</li>
> </ul>
> </td><td><pre>[comment]</pre><pre class="clsCode"></pre></td></tr>

I've been waiting for a peer review of IOTP. Anyways, I put it in
(see http://www.w3.org/2000/03/29-XML-protocol-matrix#compare.XIOP)
but commented out the facets as they were out of scale with the
others protocols and referenced no-exsitent facets. Compare the
rendered XIOP facets:
    serialization
          custom 
          automatic 
          fixed 
          implicit
          (contextual) 
          explicit 
    extensibility
    encoding
    stereotypes,
    encoding forms,
    OMG IDL
    constructs, XML
    document and
    fragments 
    skinnyness raw
    IIOP (binary CDR
    encoding) instead
    of XML stream 
    protocol GIOP 
    interface discovery
    Interface
    Repository 
    transactions
    (ACIDity)
    Transaction
    service 
    remote procedure
    OMG Corba 
    business process
    UML, Business
    Objects Domain
    Task Force 
    security Security
    service 
    routing Event- and
    Notification
    services, Object
    Reference
    framework with
    bridges, routing etc 
    Object Oriented
    framework OMG's
    OMA 
    Language bindings:
    Client C, C++,
    Java, ADA, Cobol,
    Smalltalk, (Python
    ,Perl,...) 
    Language bindings:
    Server C, C++,
    Java, ADA, Cobol,
    Smalltalk,
    (Python,Perl,...) 
    Instance concept 
    Instance discovery
    Naming Service,
    Trader Service 
    Realtime 
    Services OMG
    Corba services 
to IOTP which had the second longest facets:
    serialization
          custom 
          implicit 
    protocol 
    extensibility 
    transactions
    (ACIDity) ??? 
    business process 
    routing 

The other problem is that the detail links following the facets all
pointed to "http://www.omg.org" which didn't help to clarify or
explain the facet entry.  (For instance, it would be usefull if the
Interface Repository link pointed to some text about an interface
repository on the OMG site.)

The goal of the facets was to provide a checklist of features. Too
much extra text runs the risk of making that matrix entry look like
an advertisement.

A quick scan of the forms encoding spec
[http://xiop.sourceforge.net/html/forms.html] gave me the impression
that the serialization entry should look like:
    serialization
          fixed 
          explicit 
but I thought I'd wait to hear from Anders and others on the list who
were more acquainted than I with the spec.

-- 
-eric

(eric@w3.org)

Received on Wednesday, 31 May 2000 11:25:58 UTC