Open Standard Voting Localization with CAM - Localization Mechanisms - Publishing Localizations - Leveraging Open Standards - XSD ingesting David RR Webber Chair OASIS CAM TC (Content Assembly Mechanism) E-mail: drrwebber@acm.org http://wiki.oasis-open.org/cam
Overview Goal is to promote the use of open public standards for voting systems Develop in-country localizations from standard international base set Provide common open public lexicon for elections management and verification Leverage XML and open standards approach, including NIEM* Provide developers with tools to aid delivery, documenting and testing of solutions *NIEM National Information Exchange Model slide 2
Approach Voting Localizations built from OASIS EML EML has 29 pre-built components (schema) for voting processes Use desktop visual editor to pick components Trim and select only what you need from each standard schema Desktop toolkit makes templates for the XML required Generate localization details Automatically build XML components to use Build your own dictionary, compare to EML standard Create readable documentation of details Test, Package and Publish The CAM desktop toolkit environment supports NIEM IEPD approach Available as Open Source on SourceForge.net Implementation of the public OASIS CAM standard for schema templates (camprocessor) slide 3
Localization Mechanisms Creating a Country Voting Localization Package
Localization Key Components 1 XML Schema EML Schema XSD aligns voting information model, processes and domain dictionary 2 3 Documentation Readable information use details (HTML), EML cross-reference Spreadsheet XML artifacts The localized schema and picklist.xml Key components (aka IEPD*) 4 Examples & Test Cases Realistic test XML instances for conformance and interoperability testing *IEPD Information Exchange Package Documentation slide 5
Localization Approach Steps EML Voting Exchange Manage 11 develop XML XSD 22 ingest Base CAM Template 33 localize Revised CAM Template 44 Local Generate XML XSD 55 Compare EML Dictionary Steps: 1. OASIS develops standard international XSD schema Implements the voting process, management and information exchange needs 2. Ingest the standard XSD into base CAM template 3. Apply localizations to CAM template Exclude unwanted optional components Align local exchange needs to standard Extend for local needs 4. Generate new localized XSD schema that conform to EML standard 5. Build dictionary cross-reference of localization usage 6. Publish documentation and examples 66 Publish Documentation + Examples slide 6
Creating a Package for Localization Step 1 - Ingest EML XSD rules, structure, annotations Step 2 - Select & mark out your use model in visual editor - Save your pick list selections - Compress template to desired structure Step 3 - Generate your localized use XSD schema subset Step 4 - Generate and test rich concrete data examples (complete with content hints / pass / fail / random options) - Run rules engine - Verify operation and outcomes Step 5 - Build localization documentation of structure elements, usage, rules and cross-reference map Step 6 Package, Publish and Share (IEPD) slide 7
Challenge: XSD schema has limitations! The schema contains the superset of voting exchange component variations for every country XSD syntax does not have direct context mechanisms Hence people make everything in schema optional Dependencies are not clear It is difficult to understand the constructs and to document the rules clearly for election authorities to verify It is hard to create test cases and instances (the pick list tough to visualize) These are all limitations inherent in W3C Schema language itself slide 8
Completing Localization Picture: Base EML XSD schema structure model CAM template + rules / refinements Documentation of use patterns (aka pick list + subset XSD) Test cases and examples Content hinting (localization use data) Validation engine for unit and regression testing Open standard, open platform and open source allows consistent agreements between participants slide 9
Localization Package Use Scenario Publish XML Structure Rules Context Examples Documentation 1 Package 2 Report Templates Results Validate Test Cases Test Testing, Agreement / Alignment. Developer Uses & Creates 3 XML Localization Content Hints Verify Result Outcomes html 4 XML Pass / Fail slide 10
Tutorial Base CAM templates (ingesting XSD schema ) step by step example Documenting the Localization Patterns Creating pick list selections Compressing template structure Subset XSD generation Publishing Localization package Cross-reference spreadsheet Documentation reporting options Generate use dictionary Testing and Conformance Creating Test Case examples Content Hinting Running Test Cases Advanced Techniques slide 11
Base CAM Templates (ingesting XSD Schema) Using jcam editor Wizard (http://www.jcam.org.uk)
EML V6.0 provides base CAM templates Templates come pre-packaged in the EML V6 download However, for completeness we show how to generate one slide 13
Ingesting XSD to CAM template format OASIS EML XSD Collection XML Schema XSLT XSLT tools tools process process entire entire EML EML XSD XSD collection collection Structure Common xsd Components CIQ xsd Rules Namespace INGEST Localization EML xsd Documentation EML xml CAM = Content Assembly Mechanism slide 14
Step 1 & 2 Pick the XSD schema to ingest Choose File / New Template 1 Specify XSD Location root directory 2 Pick XSD First location is the root folder for the XSD collection. Typically this is the same as the location for the XSD schema you want to ingest. (Complex XSD can have this in separate folder; hence need for option) slide 15
Step 3 Choose the XSD parent element Confirm and start the XSD ingesting 3 Choose XSD parent element From the dropdown list pick the correct root element you wish to use (EML). XSD s may have more than one collection in them (as shown here) or may not have the parent node element cleanly defined. The Wizard shows you the list of all possible ones it finds so you can select and confirm the right one. slide 16
Step 4 Ingesting complete Save Results Completed template is loaded and save dialogue appears 4 Processing usually takes a few seconds, some larger EML XSD can take a few minutes to process. 4 Tip: jcam runs the ingesting as a background task so you can continue to use the computer while such long ingesting is proceeding. slide 17
Step 5 Review ingested structure Completed template is shown in the structure navigator panel for review and editing 5 Template contains all the default content model and structure rules ingested from the XSD. All annotations and documentation from XSD also ingested (show as paperclip symbol). Code lists and typical content values inserted for easy visual reference. slide 18
Documenting the Localization Patterns Pick lists, documentation and XSD subset generation
Building a pick list Structure EML structure is all inclusive!!! Make pick list Rules Documentation MARK MARK WHAT WHAT IS IS NOT NOT NEEDED excluded items DESIRED RESULTS slide 20
Marking Items for exclude - pick list Can exclude at all levels within the structure excludetree() excludeelement() excludeattribute() Use XPath operators to control scope: Specific node Group of nodes Anywhere occurs in structure Contextually based on condition Can add new domain elements with own namespace to EML template slide 21
Using Editor to mark Exclude items Structure Editor Tools Select focus on item Invoke action menu (right mouse click) Pick action Tip: exclude rule display is context sensitive and only available on optional items Tip: use Add New Rule mode to specify different XPath for exclude (quick mode assumes current path ) slide 22
Export and Save completed pick list File Menu Option Select Export / Pick List and then specify filename of destination. Excluded items are designated with red dot in structure editor and italics font with no bold highlight slide 23
Pick List Details Pick lists provide a handy way to catalogue the localization model and can be re-used later by importing into other templates EXCLUDE FLAG VALUE Exported Example slide 24
Compress Operation File Menu Option Select option and specify filename for new copy of your template. Compress process removes all rules and structure items marked with an exclude statement. Note: ignores excludes that have a conditional context expression (advanced section technique). Note: You can retain both original and new compressed templates slide 25
Extending the base EML schema In addition to excluding existing optional components you may need to add new pieces New pieces should be assigned their own extension namespace prefix to identify them from the base EML components Typically the OASIS EML committee will review new pieces as candidates for adding to future versions of the standard Send comments and suggestions to election-comment@lists.oasis-open.org slide 26
Adding New Domain Elements 1 Select root element node in structure; right mouse click for context menu; select Add Namespace ; then enter prefix and URI 2 Select element node in structure; right mouse click for context menu; select Add Child Attribute / Element ; then enter prefix and name slide 27
Generating your EML use schema Once your localization details are completed you are ready to complete the remainder of the process This starts with generating your own new subset schema Then once schema details are OK creating test cases, cross-reference spreadsheet and documentation slide 28
Generating subset use schema File Menu Option Select Export CAM as XSD menu option 1 Confirm the location and filename, and namespace mode. 2 Select default for Mode to write a regular import set of XSD schema for each namespace slide 29
Schema subset generated Set of XSD files with filename and namespace suffix Each namespace file is import for those specific type definitions Reviewing XSD results in a schema visual editor tool slide 30
Information Restriction Considerations The CAM template allows full use of XPath conditional expressions and a rich set of over 30 functions including: setnumberrange(), setlength(), setlimit(), setdatemask(), setnumbermask(), restrictvalues(), lookup() Those that are compatible with XSD constraints will cause constraint schema assertions to be written out when exporting to schema In the advanced topics section we will look at cross field validations using XPath conditional rules slide 31
Generating Testing and Conformance Examples Generating base examples Customizing content with Hints Selecting valid and invalid modes Running rules validation check
Test Case Generation Quick Start File Menu Option Default directory to write examples into Name to be used for the examples How many examples to create Repeating elements count for Quick Test just click OK to use default settings slide 33
Test Case Results Active links to view the generated examples slide 34
Advanced Generation Options Use content type or item name (name is useful for checking backend transform processing) Use namespaces or not; if false is selected then XML instances are created with minimized namespace usage. Optional content hints (explained next ) Optional schema file validation link; use this to have example validate with EML schema or local subset schema How to handle optional items: all random none If you want deliberate errors for fail testing; (will give variety of data and structure errors) Use slider to pick a specific seed value or leave blank for random seed slide 35
Test Case Generator Feature Summary Make both Pass / Fail testing examples Content hinting so examples use real not fake data Test optional item logic with: all / random / none Uses exclude() assertions so does not include those items makes realistic examples of your use pattern Can pass in seed value use when adding and testing hints (each test case is labelled with its seed value) Make hundreds of test cases without manual editing Can link test case to XSD schema for structure tests You can modify XSLT tool for own testing needs slide 36
Run CAM Rules Check on Examples Run Menu Option 1 Pick Test Case Example to VALIDATE; click Finish to run validation rules Review validation results 2 3 slide 37
Content Hinting Mechanisms Designed to create realistic data examples Hints can be provided in two ways Firstly - using Value notes in annotations on specific items in the structure editor Second create your own Hints XML file and add matching rules to globally apply across your template(s) e.g. FirstName, LastName, Address, BirthDate, etc. Can export from one template, import into another slide 38
First Approach: annotation Value Hints Select focus on structure item Invoke action menu (right mouse click) Choose Edit Annotations slide 39
Then add Value annotation item 3 Select Value tab 2 4 1 Enter values terminated with character Click on Add New, then enter Value as Type and confirm OK 5 Re-run Example export to see results slide 40
Second: Hints File Mechanism (XML file) 1 like / with partial name matching use for component match on items e.g. first with name matches <nxnl:first_name> 2 key matching on tag name use for exact match of items 3 key / parent path matching use when same name occurs within different parents e.g. Country and Person / NameElement with different content and context Note: matching is case sensitive but ignores namespaces TIP: can use Export Hints to create initial XML file for editing slide 41
A- Using Examples Generator with Hints Select XML hints file to be used here slide 42
B- Import Hints into Annotations (merge) set and select as needed Option to autoreload new anno file into current template slide 43
Documentation Cross-Reference Spreadsheet Default reporting options
Generating Cross-Reference Spreadsheet Use Compare to Dictionary tool Compare to default EML-V6-dictionary Drag and Drop XML file into Microsoft Excel Review and save spreadsheet workbook slide 45
Dictionary Compare Tool Select compare tool menu option Browse and choose EML dictionary slide 46
Drag and Drop to Excel Spreadsheet Click OK slide 47
Documentation Layouts Five options Source XML Component details (XML) Tabular format (HTML) Interactive web page (wiki) Code list Tabular format page layout for data analyst use and designed to make rules and use patterns clear Each documentation layout XSLT script can be customized as desired slide 48
Tabular Documentation Documentation Menu Option Select Format, resulting HTML opens in browser viewer slide 49
Tabular HTML Content Details Clear use pattern Extended Code list handling Annotations XPath references and functions Enhanced Data type Logic slide 50
Summary Ingesting XSD Creating use pattern (aka pick list) Generating test examples Hints system Generate XSD schema subset Running tests
Localization Checklist Review 1 XML Schema EML Schema XSD aligns voting information model, processes and domain dictionary 2 Documentation Readable information use details (HTML), EML cross-reference Spreadsheet 3 4 XML artifacts The localized schema and picklist.xml Examples & Test Cases Realistic test XML instances for conformance and interoperability testing slide 52
Summary Capabilities covered Ingesting XSD to Template Creating use pattern (aka pick list) Generating test examples Hints system Generate local use XSD schema subset Running tests Generating documentation and spreadsheets Allows rapid localization development Enables consistent implementations Documentation provides clear details on structures, content, rules and definitions Cross-reference to EML standard slide 53
Advanced Techniques Extending rules for actual use cases Using XPath selector wizard Handling ingestion recursion issues SourceForge XSLT repository
Importing pick list operation File Menu Option Select Import and specify filename of your existing pick list xml. Import process matches the path expressions in your pick list to the XPath expressions in the template. Exclude statements generated for matching items. Makes it easy to re-apply a pick list on new versions of schemas, or on similar schemas with same blocks of content address, company, person, etc. slide 55
Illustrative EML requirements When Ballot type > 0 - AND Language code > 1 Require Ballot State = CA, NV, NM Resource Matching When Registered Voters Total < 1000 - AND Ballots > 1000 Audit alerting Region Restriction slide 56
Template rules syntax <as:businessusecontext> <as:rules> <as:default> <as:context> <as:constraint condition="//castballots /Total > 1000 and( // RegisteredVoters /Total < 1000 )" action="restrictvalues(// CastBallots/Total, Alert possible audit required ) "/> <as:constraint action="restrictvalues(//eml:administrativearea /eml:nameelement, 'CA NV NM')"/> </as:context> </as:default> </as:rules> </as:businessusecontext> slide 57
XPath selector wizard XPath syntax is automatically generated for you Check boxes allow user to select scope of rule action Conditional context rules can be optionally used Documentation notes can be added here slide 58
CAM Kit of XSLT tools used XSD 2 CAM extracts structure and rules and builds template from schema XML 2 pick list builds pick list from any XML instance (uses Level Depth setting) CAM 2 Examples Generates a collection of XML instance test cases Import/Export Hints Manage and apply content hinting across CAM templates CAM 2 XSD export Creates a subset XSD from CAM template slide 59
SourceForge XSLT svn repository Using any svn client the XSLT scripts can be retrieved from: https://camprocessor.svn.sourceforge.net/svnroot/camprocessor/ camed/uk.org.jcam.camed/trunk/xsl/ slide 60
Resources Selection of useful links and additional technical details
Simple EML 440 Example Example using the EML 440 cast vote schema Simplified to minimum requirements Example can be downloaded here: http://markmail.org/message/qyrpfq75mtctblyq slide 62
Resources: www.oasis-open.org/committees/election docs.oasis-open.org/election www.oasis-open.org/committees/cam wiki.oasis-open.org/cam docs.oasis-open.org/cam www.jcam.org.uk slide 63