Can XSD validate inter-field dependencies? - xsd

I'm looking for a way for my XSD file to validate one field based on the content of another.
For example, it may need to validate that my "start date" field is before my "end date" field.
Is this possible?

Short answer: No, you can't do it with what you refer to "XSD" and probably means W3C XML Schema 1.0
Long answer:
You CAN validate such things with RelaxNG + Schematron or if it has to be XSD, use W3C schema 1.1 with it's assertion mechanism. Version 1.1 is currently in a Candidate Release phase so not a full recommendation, so you tools probably don't support that yet.
On a more philosophical level, you might want to consider whether weaving such fine-grained business rules is a right approach. I am currently involved in several standards and in those standards, I make sure it's the end user application that validates such business rules - it's too broad for a standard schema to define things in such detail, the schema would be too big, verbose, have a too steep learning curve and be much harder to maintain.

Related

Use-case specification of included use-case

Does anyone know how to describe (specify) the use-case that includes another usecase in the use-case specification table?
For example, I have a use-case "change password" that includes use-case "login" (this is just an example, i know that login should just be in the pre-condition of the use-case)
May i specify it like this?
Use-case identity: Change password
Actor: User
Pre-condition: ...
Normal flow:
Do use-case "Login"
....
....
....
What is the proper way to specify the use-case that includes another use-case?
It depends on the further rules and conventions used together with the "use-case specification table" document. Consult your company documents for the rules (and some examples).
Overall yes, your approach is valid.
The "Login" use case might be also invoked implicitly if you'd move it to the "Pre-condition: User is logged-in"
Even for free text without any Software Requirements Management tool consider using the "Name and unique-number" referencing best practice
To wrap your head around it I'd recommend (among other Google-able resources) to read the Alistair Cockburn's book "Writing Effective Use Cases", draft version is available online on his site - http://alistair.cockburn.us/get/2465
EDIT: after #BobRodes's challenge to share some alternative methodology
For more serious work consider using a Requirements Management tool, or at least evaluate some tools to adopt the best practices they support/recommend.
For one example see the link below explaining how Sparx Systems Enterprise Architect formalizes work with Use Cases and how Use Case traceability works and how it can generate documentation, UML Activity Diagram etc.
Sparx Systems: Model-Driven Use Case Analysis with Structured Scenarios, Webinar August 2013, 18min video
For another example of managing Use Case traceability and what usually happens after the Use Cases are written see
Jama Software: Jama Recorded Demo, 30min video
You might want to look into "use case narrative". In such a narrative, one of the things you will do is list included use cases and extension points. So, add a section called "Includes:" to your narrative, before the normal flow section.
Edit: Here's a bit more from Tom Pender's "The UML Bible":
The features of a use case narrative aren't standardized, but these
are common elements in wide use:
Use case initiation, or trigger, describes how to start a use case.
Assumptions define conditions that must be true before the use case may execute, but are not tested by the use case.
Preconditions define conditions that must be true before the use case may execute, and are tested by the use case.
The use case dialog explains how the user (whether an actor or another use case) interacts with the system during the execution of
the use case.
Use case terminations define the different mechanisms that can cause the use case to stop execution.
Post conditions define the state of the system that must be true when the use case ends. This helps prevent use cases from leaving
the system in an unstable condition for other use cases that follow.
Minimum guarantees describe what the actors can expect from a use case, no matter what happens during the execution of the use case.
Successful guarantees describe what the actors can expect from a use case when it completes successfully.
Although these elements are valuable, they are by no means exclusive.
Definitely look into other books and online resources on use cases,
and augment the narrative to support your own method of development.
So, perhaps I shouldn't have said that you "will" list included use cases and extension points, instead saying that you "may" do so. I like to put them in, so I don't have to keep referring back to the diagrams to look them up. Of course, you can just mention them in the dialog, too.
In UML there is standard relationship named include. Start relationship from including UseCase and end on included UseCase. See UseCase section of UML Superstructure document. UML website

Is there a way to check the validity of a model in a programmatic manner such as Validation > Validate Model does?

Is there a way to check the validity of a model in a programmatic manner such as Validation > Validate Model does?
If you mean a model is valid, iff your self-defined well-formedness rules (WFR) are valid then you can use OCL. Define your constraints in OCL and use something like Dresden OCL to validate those WFRs.
You can use Dresden OCL either on its own integrated into your Eclipse with its views or you use the API of Dresden OCL in your program to verify the validity programmatically.
Speaking about the model validation, you must consider two aspects - model syntax and semantic.
Syntax validation is a trivial thing, partially available in some tools. UML has a precisely defined syntax and it can be easily checked automatically.
If the model represents the subject correctly is another problem and can be eventually validated via model execution. There are very few tools that provide very partial implementation of such execution.
In summary - model validation is comparable with source code validation - syntactic part is easy, while semantic is hard to prove.

What is the canonical document schema for representing UML?

I want to represent UML models in a serialized document using the most standard conventions available. I am looking for a canonical schema document that explicitly dictates the best way to do this. I suspect that, if such a document exists, then it would be written in either XSD, RDF and OWL. So far, I have only been able to find OMG specification documents and academic papers on the subject. I have found some XSD schemas, but they only apply to a particular aspect of the implementation.
Is there a canonical document schema? Where can I find it? If not, what are the most common representation formats and where can I find their schemas?
Are there any libraries that facilitate the generation of mappings from well-formed documents of this kind to graphical representations (e.g. SVG)? The library would ideally be .NET compatible, but this is not a requirement.
Are you looking for XMI (XML Metadata Interchange)? It's an OMG spec for XML representations of UML models; the current version appears to be 2.4.1, and the spec page links to two XSD schema documents.
In theory, XMI should answer part of your first question. XMI is not, however, without perils. From my own experience trying to achieve exactly the same, I would think that the first thing to do is to understand what others did about it, and how these standard documents would be used. These days I would suggest a system setup that involves IBM RSA and Sparx Enterprise Architect (eval versions available, registration may be required, and I am not affiliated with any of these companies). Build at least one comprehensive UML model which would give you the coverage you need, in one tool, then use the XMI export/import function to move the model between them. See what happens. Take a look at a company such as this (the makers of TOOLBUS, not affiliated with them) and think how XMI should affect their business model, and yet it seems that it doesn't...
I would think that success is achievable, but may prove feasible for a (very) narrow scope - scoping is key here.
As for the second part, I don't think you're going to find it, particularly for .NET.
XMI is the keyword you need. But XMI specifies only the document format. It does specify the document structure or contents. The UML metamodel does this. A library which does both for you is org.eclipse.uml2 - it implements the UML 2.4 metamodel and allows to store and restore it in XMI. IBM RSA uses this library as well. As several other Java Tools do.
I'm not aware of a UML2 implementation in a .NET language.

MDD: How dynamic is MDD at runtime?

Over the years, I've investigated a lot of ways to use code generators and MDD. I've always felt that something is lacking: Patching and changes to the model at runtime.
Patching: If you have a code generator, all your classes should look the same. Now you have a single exception. All code generators so far would require that I modify the template or the template engine to make this work.
Wouldn't it be better if I could apply patches to the result of the code generation step to fix the exceptions?
Well, it depends on how you build your model. If fact, it depends on what code generator you are using, its approach, and what it lets you do.
Creating an exception to a rule (model) is more or less against the nature of MDD, unless the applied modeling approach allows you to add exceptions as modeling entities.
I think ABSE is the only modeling approach that accepts "custom code" as a first-class entity, just like a text or an integer. If you create a template that contains a "CustomCode" parameter, you can later add your exception code only when necessary, without breaking your model rules. This can be used to add or replace code. You just need to specify it in your template.
AtomWeaver is a free implementation of the ABSE modeling methodology.
MDD doesn't work because it is based on a view of the domain and not the entire domain. I mean that usually MDD take an XMI in entry coming from an UML diagram. The problem is that this diagram is only a view of the domain and therefore you have many alternative and the real world is a mot more complex specially at deployment stage.
The only company which has provided me real value in my project was Omondo with EclipseUML. EclipseUML doesn't try to do MDD but create UML at diagram level live synchronized with code. Deployment is made using stereotypes which are added in the java annotation in the code. I can therefore model and if I add deployment stereotypes then my application can be deployed immediately/ If I manually change my code, then my model is refactored and all my views updated. If I want to add a documentation then I just add notes in the metamodel. These notes are live available when I click on each element. No more printed documentation needed because live navigation, dynamic views creations etc...
My EclipseUML model is always up to date, and I can deploy it immediately because Java annotations are lived synchronized between the model, the metamodel, the diagrams and the code. Really cool :-) :-)

Formatting XSD scheme for peer review

I designed a data model which is represented by an XSD scheme.
The data model also provides the types that are being used as web service parameters in a WSDL descriptor.
I would like to send the XSD scheme around and ask the people involved to peer review the data model.
What tool or presentation method would you suggest to be used as a basis for peer reviews? The data model should be readable for non-skilled people, at least when it comes to the semantic meanings of the parameters
Edit:
To be more specific: Of course, syntactically, the scheme validates. Actually I'm already working on code which is based on JAXB generated classes. My goal is
to freeze the data model and thus
the input parameters
to make sure
nothing got lost or forgotten from a
semantic (in the meaning of
business-relevant) point of view.
Edit 2
I've been thinking about how it probably would be best to spread a datamodel around. I'm thinking of something like a JavaDoc for XSD schemas. Anyone knows if something like that exists? Basically it would be done with a set of XSLTs, right?
I know the following tools that generate documentation from XML Schema files (XSD):
xs3p
XSLT stylesheet that generates single XHTML from XSD
xsddoc
free / LGPL
mainly XSLT based
JavaDoc like output
see xsddoc examples
xnsdoc
improved commercial version of xsddoc
free for personal/educational use
JavaDoc like output
XSDdoc 2.0
commercial
JavaDoc like output
For small a XML schema, I would probably suggest using the xs3p XSLT stylesheet. For more a complex schema, I suggest using xsddoc.
I recommend using the XSD for something. Specifically, show some actual applications, with examples as real code.
Actual applications are what make a schema interesting. The examples don't have to be big, sophisticated or completely realistic. They just have to compile. Other people will want to copy and paste the code samples.
These examples are the "hello world" of the schema. And they act as a kind of unit test for the schema.
The closest thing to Javadoc for an XML schema that I've seen is running the Javadoc tool on source generated from the schema. This requires two things: 1) That your schema has internal annotation elements documenting it, and that 2) your source generator uses those annotations as Javadoc elements.
The very useful Oxygen XML developer also supports generating documentation, see
http://www.oxygenxml.com/xml_schema_documentation.html
(commercial, but there's a fully functional 30 day trial available)
I'll try it out now, need a simple way to generate a document with all types and available xsd:documentation description as a simple interface description...
** Disclosure : I work for Innovasys, the producer of the documentation tool mentioned below *
You could take a look at Innovasys Document! X. As well as automatically generating a structured and linked page for every element, simple type, complex type, group and attribute group it will also generate linked XSD diagrams (including sequences/choice etc.) and structure tables that include the annotations from your XSDs and make sense of the relationships between the elements in your schemas. The output is template based so you can adapt it to your preferred style and structure. It will build output to web ready html or compiled help files.
Uniquely it also includes a WYSIWYG editor that allows you author additional content to supplement the stuff that's automatically generated and the annotations from the XSD source - so you can provide additional contextual information for your peer review. There is also a Community Extensions feature that allows people viewing the generated output to record comments and feedback and that can be viewed and actioned directly from within Document! X.

Resources