New features of XML for Microsoft Windows 2000
Author: Charlie Heinemann
Update Date: January 10, 2000 Published: December 15, 1999
The Microsoft XML Analysis program (MSXML.DLL) released by Microsoft® Windows® 2000 is actually a service pack version. It contains not many new features, but a large number of errors are fixed. This version of MSXml.dll can also be obtained from the Web version of Internet Explorer 5.01, so it can also be used for Windows NT 4.0. However, compared with its earlier versions, the most prominent improvement of this version of the analysis program is greatly improved reliability and scaling. The new version of the analyst has a better performance in the face of large workloads and has more powerful zoom capabilities in multiprocessor environments, thereby better meeting the server requirements.
The most prominent improvement is greatly improved reliability and zoom.
This paper summarizes the new features of this analysis program and describes changes in behavior that affect its function. For more details, see XML Reference Documents for XML Developer Center (English).
New features for Microsoft XML analysis programs
This article will detail the following new features of the Microsoft XML analysis program:
For the data type of the attribute supports the "ID" element as an element name of the element name to create a COM object in the XSL script block.
Data type support for attributes
The data type support provided by the Internet Explorer 5 version of the Microsoft XML analysis program has been developed to provide simple data type support for properties (such as "IT", "float", and "date"). For a complete list of supported data types, see the "XML Data Architecture" reference documentation.
Analyze the extended data type support features within the program, implement data with attribute-based core, which can still be specified in the data type. The data type can be declared on the attribute node of the entire architecture, as declarations on the element node. DT: Type properties using AttributeType elements:
"Id" type element
Now, the element can be "id" type, as shown in the following code:
The BAR element has a "ID" type, which means that the text value of the BAR element can be used to reference the foo element. The architecture can also be used to specify the element as the "ID" type except that the element on the instance is specified as the "ID" type.
When the element has the "ID" type, the parent element of the element can be referenced by the value of the ID element. For example, if you want to pass the string "fooID" to the NodeFromid method, the method returns the foo element.
XMLDoc.nodefromid ("fooid")
Data type as element name
Currently, the data type can be declared using DT: DT attributes on an instance, or in the architecture with a DT: Type property or a DataType element to declare data types. For Microsoft XML analysis programs released with Windows 2000, they can also specify their types through the name of the element. The following XML element is an integer type because its name has "URN: Schemas-Microsoft-COM: Dattypes" namespace, and its local name is a valid data type - "int":
This new method for this designated element type provides a convenient way to create a type of element.
Example of COM object in XSL script block
In the Microsoft XML analysis program released with Internet Explorer 5, there is an example of the COM object in the XSL: Script block for security reasons. For Microsoft XML analysis programs released with Windows 2000, this has been corrected, and now you can safely illustrate COM objects in XSL: Script blocks.
Microsoft XML analysis program behavior changes
Some behaviors of the Windows 2000 version of the Microsoft XML analysis program have changed. These changes are generated by correcting errors and user feedback. The following table lists errors that cause accidental changes in behavior.
Error correction
Functional Area Error Description Note The XSL Mode For an element defined by the entity, the query with // will return the result of the repetition. The contents of the DocumentType node are now not allowed to make XSL mode navigation. The data type even if the data type declaration is deleted (ie "DT: DT = 'int'"), the element still retains its data type. After deleting a data type declaration, the element will no longer retain its data type. The namespace allows the reserved namespace to limit the element. Regardless of the "XML" case, the user cannot set it to prefix. It is a reserved namespace prefix. Namespaces cannot keep the namespace declaration on empty elements. It is now able to keep the namespace declaration on empty elements. Data Type Data Type Validity Judgment is performed before the XSL transformation. This will cause the following XSL elements to have an error:
The Microsoft XML analysis program no longer takes effect before the XSL can transform. XML Data Source Objects The onCellChange event will inspire twice. Now, for each instance of the data, the onCellChange event is only excited once. The previoussibling feature of the first child element of the object model element will return the last attribute in the Attributes collection on the parent element (if the Attribute collection is not empty). The previoussibling feature of the first child element of the element will now return NULL (whether it is empty in the parent element). The document with the object model is replicated to the original document ID mapping (ie, id navigation will return you back to the original document). The ID in the copied document is mapped to the node inside the copied document, not the node in the original document. The object model replication document does not contain replication of the internal subset of the original document. The copied documentation now contains replication of the internal subset of the original document (if such a subset) is present. XML DSO does not cause the recovery set in the form of a re-constructor by setting the DocumentElement property. By setting the DocumentElement property, replace the root element, it will now lead to the form of a reconstruction set. An invalid XML is allowed in the effectiveness DTD (the value of the entity) is allowed. There is no longer allowed to have invalid XML in DTD (when the entity is decomposed). The object model does not provide an error message in Visual Basic®. Error information is now available in Visual Basic. The object model sets the NodeTypedValue feature of the Boolean element, which generates an invalid XML after the hold operation. Now when the Boolean value is set true, the "1" value is maintained, not "-1". "-1" value is not a valid value for the Boolean node. IE4 object model attribute values are not compatible with Internet Explorer 4.0. Now, / t and wraps in the attribute value are converted to '' (spaces). This is compatible with the implementation of Internet Explorer 4.0. Analysis Program Microsoft XML 2.0 Analysis Procedure does not support "US-ASCII" encoding. Microsoft XML 2.0 analysts can now support "US-ASCII" encoding. Object models are in terms of HTML-based XML applications, unregistered wraps are serious problems. Now the carriage return and the wrap are converted into a wrap. The object model uses the loadXML method to load an XML document, which will force ResolveExternals to false. If there is no security environment (such as in C, C or HTML behavior), the ResolveExternals feature will be false. If there is a security environment (such as in the HTML page), it should be set to TRUE. Object model If the Issuccessful parameter is null, the loadXML method will return S_FALSE. When the analysis fails, the Microsoft XML Analysis program only returns S_FALSE. Analysis Program Microsoft XML 2.0 Analysis Procedure Unable to process parameter entities in the element declaration, as shown in the following example: The data type is incorrect to bin.base64. Microsoft XML 2.0 Analysis Program The "/" in this character set is "*".
The base64 character set is: "Abcdefghijklmnopqrstuvwxyz abcdefghijklmnopqrstuvwxyz0123456789 /". The Microsoft XML analysis program can now be identified correctly. The script in XSL XSL changes the DOM in Internet Explorer, but does not change the DOM in other containers. Now, regardless of the container, the Microsoft XML analysis program changes the DOM. The value of the XML DSO $ TEXT domain is different from the node text values mapped to the domain. The $ text domain and node value are now formalized. XSL XSL can expose the "DT" attribute of the data type defined in the architecture. After the data type is defined in the architecture, it will not be revealed in the DOM tree. DT: DT attribute. Analyze the ID of the program registration entity declaration. Microsoft XML Analysis Program no longer registers the ID under the entity declaration. Other changes in behavior
The following changes are made according to the feedback of users, which will result in changes in behavior of your application:
Microsoft URLs are no longer case sensitive. If the ID () is called twice, the same node is referenced, the second reference is active. When the Namespaceuri feature of the XMLns: Foo and XML: Space property is obtained, "http://www.w3.org/xml/1998/namespace" and "" are obtained separately. Now this change is an error that is behaving abnormal behavior in the Microsoft XML analysis program released with Internet Explorer 5. This error will be corrected in the future release.
Improvement in workload and performance
One of the extent of automation has been resolved, so that the analysis program can be substantially zoomed in "Active Server Page (ASP)". This means that in multiprocessor computers, the throughput of the .asp file using XML will be significantly improved. The following chart shows these improvements, where the X-axis represents the number of processors, the y-axis represents the number of ASP requests per second:
This is measured based on the simple XML.ASP page that performs some DOM tree operations.
In addition, some errors caused by some workloads are also corrected, and thus greatly increases the stability of the analysis program at workload. We deploy this analysis program to some users of Microsoft.com to access the most frequent sites, which proves that its stability under workload has been greatly improved.
Charlie Heinemann is a planning manager of Microsoft's XML team. He is from Texas, good at high hopes to think about problems.
