W3C home > Mailing lists > Public > xmlschema-dev@w3.org > September 2008

RE: XSD to ER Diagram

From: Andrew S Halper <ashalper@usgs.gov>
Date: Thu, 4 Sep 2008 09:22:01 -0700
To: "Tsao, Scott" <scott.tsao@boeing.com>
Cc: "BANBURY David" <david_banbury@rta.nsw.gov.au>, "Jack Lindsey" <tuquenukem@hotmail.com>, xmlschema-dev@w3.org, xmlschema-dev-request@w3.org
Message-ID: <OFC32B4382.B313CA59-ON072574BA.0058B60E-072574BA.0059E9FC@usgs.gov>

UML is more expressive than most ER diagram notations, thus easier to map 
to XSDL. The down-side is that UML is also more ambiguous and, like most 
ER notations, not rigorously defined mathematically. Every UML to XSDL 
mapping I've seen is different.

Some pages to have a look at:


(also, apologies if I've spoken out-of-turn)

Andy Halper
Computer Specialist
National Water Information System

"Tsao, Scott" <scott.tsao@boeing.com> 
Sent by: xmlschema-dev-request@w3.org
09/03/2008 08:51 PM

"Jack Lindsey" <tuquenukem@hotmail.com>, "BANBURY David" 
<david_banbury@rta.nsw.gov.au>, <xmlschema-dev@w3.org>

RE: XSD to ER Diagram

"UML and XSD are a much closer match in terms of their respective feature 
sets, which is one reason why UML tools are more commonly seen when 
modelling for XML implementations..."
I wonder if you could elaborate on this more, as I don't understand 
(theoretically) why UML would be a better notation for mapping into XSD 
than E-R.
As far as I know, there is a standard for E-R (or RDB) mapping to XML 
(i.e., SQL/XML), just as a standard for UML mapping to XML (i.e., XMI). 
So, why would UML be a much closer match than E-R.
Please excuse my ignorance... Am I comparing apple with orange?

Scott Tsao 
Information Architect - Associate Technical Fellow
The Boeing Company 

From: Jack Lindsey [mailto:tuquenukem@hotmail.com] 
Sent: Friday, August 22, 2008 6:20 PM
To: BANBURY David; xmlschema-dev@w3.org
Subject: RE: XSD to ER Diagram

I hope you don?t mind if I say I find this scenario a little perverse. 
If your community of interest considers an E-R model an effective means of 
communication (as pointed out it can only represent a subset of XSD) why 
not publish your data requirements in the form of an E-R data model and 
then generate the XML schema from it?  Some participants might find the 
common E-R model a useful basis for producing a RELAX-NG schema or an SQL 
database design or other physical implementations.
In other words, why implement and then model?  Why not model first and 
then implement, in as many media as participants desire?
Embarcadero probably has the most to offer in this arena.  It allows you 
to specify your own set of standardized XSD data types, for instance.
Umodel is for UML, not E-R, although UML modelling tools can be used to 
represent E-R models if you have the discipline to restrict your use of 
UML features.  UML and XSD are a much closer match in terms of their 
respective feature sets, which is one reason why UML tools are more 
commonly seen when modelling for XML implementations.  But once again, it 
is usually the UML that is used to automatically generate the XML schema, 
not the other way around.
As for ERwin XML, the last time I saw it, it was their own format (not XMI 
or anything) for exporting and importing an ERwin E-R model, complete with 
graphical positioning information.  For instance, we used it to export a 
data model and run it against the names in a data dictionary so we could 
generate a French version from the English version (naturally we had to 
spread the entity boxes out a bit afterwards because the French names were 
longer ? perhaps we should have gone in the reverse direction ;-)
Jack Lindsey

> Date: Fri, 22 Aug 2008 14:36:32 +1000
> From: David_BANBURY@rta.nsw.gov.au
> To: xmlschema-dev@w3.org
> Subject: XSD to ER Diagram
> Please excuse me if this query does not directly concern specific XSD
> issues but I hope it is related closely enough to XSD development to be
> worthy of the group's learned opinion.
> We are using XSD to define an XML file format for data interchange. As
> such the schema is not directly related to a specific system or
> database. The schema relies heavily on key-keyref relationships to
> describe the relational structure of the data model. It has proven very
> useful throughout development to present the schema as an
> Entity-Relationship Diagram but deriving an ER Diagram directly from the
> XSD has been problematic.
> Are there any tools available which will automatically generate an ER
> Diagram from XSD?
> Thanks for any information you can provide.
> Regards,
> David Banbury
> Traffic Systems Branch
> Roads & Traffic Authority of NSW
> Australia
> Tel: +61 2 8396 1417
> Fax: +61 2 8396 1600
> Before printing, please consider the environment.
> IMPORTANT NOTICE: This e-mail and any attachment to it are intended only 
to be read or used by the named addressee. It is confidential and may 
contain legally privileged information. No confidentiality or privilege is 
waived or lost by any mistaken transmission to you. The RTA is not 
responsible for any unauthorised alterations to this e-mail or attachment 
to it. Views expressed in this message are those of the individual sender, 
and are not necessarily the views of the RTA. If you receive this e-mail 
in error, please immediately delete it from your system and notify the 
sender. You must not disclose, copy or use any part of this e-mail if you 
are not the intended recipient.
Received on Thursday, 4 September 2008 16:22:49 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:15:48 UTC