U.S. patent application number 10/377275 was filed with the patent office on 2004-09-02 for determining the occurrence of events using decision trees.
Invention is credited to Barthel, Knut.
Application Number | 20040172347 10/377275 |
Document ID | / |
Family ID | 32908106 |
Filed Date | 2004-09-02 |
United States Patent
Application |
20040172347 |
Kind Code |
A1 |
Barthel, Knut |
September 2, 2004 |
Determining the occurrence of events using decision trees
Abstract
The determination of whether an event, such as a taxable event
for a commercial transaction, has occurred uses an executable
decision tree. The decision tree includes a number of test nodes
that each include a comparison field and at least one allowed
value. A data value associated with the comparison field is
accessed and compared to the allowed values. If the comparison
field equals an allowed value, execution proceeds to a child node,
and if not, execution proceeds to a sibling node. The child and
sibling nodes are either another test node, a result node, or an
error node. Execution of the decision tree proceeds until a result
node or an error node is reached.
Inventors: |
Barthel, Knut; (Walldorf,
DE) |
Correspondence
Address: |
FISH & RICHARDSON, P.C.
3300 DAIN RAUSCHER PLAZA
60 SOUTH SIXTH STREET
MINNEAPOLIS
MN
55402
US
|
Family ID: |
32908106 |
Appl. No.: |
10/377275 |
Filed: |
February 28, 2003 |
Current U.S.
Class: |
705/31 |
Current CPC
Class: |
G06N 5/04 20130101; G06Q
10/10 20130101; G06Q 20/207 20130101; G06Q 40/123 20131203 |
Class at
Publication: |
705/031 |
International
Class: |
G06F 017/60 |
Claims
What is claimed is:
1. A computer program product, tangibly embodied in an information
carrier, for determining whether an event has occurred using a
decision tree, the computer program product being operable to cause
data processing apparatus to: proceed to and execute a first of at
least one test node of the decision tree, wherein each of the at
least one test node comprises an identification of a comparison
field and at least one allowed value, and the execution of each of
the at least one test node comprises accessing a data value
associated with the comparison field for the test node being
executed, determining whether the accessed comparison field data
value is equal to any of the at least one allowed value for the
test node being executed, and if the comparison field data value
equals any of the at least one allowed value, proceeding to a child
node of the test node, and if the comparison field data value does
not equal any of the at least one allowed value, proceeding to a
sibling node of the test node, wherein the child and sibling nodes
are either another test node, a result node, or an error node; and
execute any further test node to which execution of the decision
tree proceeds until execution proceeds to a result node or an error
node.
2. The computer program product of claim 1 being further operable
to cause data processing apparatus to: identify for execution a set
of a plurality of decision trees stored in a decision tree
repository, wherein each decision tree corresponds to a different
event; and execute each of the identified set of decision
trees.
3. The computer program product of claim 2 wherein the events that
correspond to the plurality of decision trees are taxable events
arising from a commercial transaction.
4. The computer program product of claim 2 wherein each of the
plurality of decision trees stored in the repository embody a form
comprising a mark-up language.
5. The computer program product of claim 4 wherein the mark-up
language of each of the plurality of decision trees comprises an
extensible mark-up language.
6. The computer program product of claim 1 wherein each of the at
least one test node represents a question that forms a basis for a
determination of whether or not the taxable event has occurred.
7. The computer program product of claim 1 wherein the event is a
taxable event arising from a commercial transaction.
8. The computer program product of claim 7 wherein each of the at
least one test node represents a question that forms a basis for a
determination of whether or not the taxable event has occurred.
9. The computer program product of claim 1 wherein the event is a
fiscal code to be printed on an invoice.
10. The computer program product of claim 1 wherein the event is an
identification of a delivering plant according to certain
optimization rules.
11. The computer program product of claim 1 wherein the event is an
identification of bank and payment method for settlement.
12. In a system that determines which of a plurality of different
taxable events have occurred for a commercial transaction, a
repository of a plurality of executable decision trees, each tree
for determining whether a different taxable event has occurred, and
each tree comprising: at least one executable test node, wherein
each executable test node comprises a question to be answered in a
determination of whether or not the taxable event corresponding to
the tree has occurred, and wherein each executable test node, when
executed using information about a specified commercial
transaction, yields an answer in one of a first form or a second
form, wherein an answer in the first form causes a child node to be
executed and an answer in the second form causes a sibling node to
be executed, and further wherein each child and sibling node is
either another test node, a result node or an error node; and at
least two result nodes corresponding to a determination of whether
the taxable event corresponding to the tree has occurred.
13. The repository of claim 12 wherein each decision tree
corresponds to a transaction tax type.
14. The repository of claim 13 wherein the transaction tax type for
one of the at least one decision trees is a value added tax.
15. The repository of claim 13 wherein the transaction tax type for
one of the at least one decision trees is a use tax.
16. The repository of claim 13 a decision tree corresponding to a
transaction tax of a specified type is able to determine whether or
not a taxable event of the specified type has occurred in a
plurality of taxing locations.
17. A computer-implemented method for generating a user interface
for a decision tree, the method comprising: associating a decision
tree with a series of user interfaces to request decision tree
information from a user; displaying each user interface in the
series of user interfaces associated with a particular decision
tree; receiving decision tree information from a user using at
least one of the user interfaces; and storing the decision tree
information received from the user, wherein the decision tree is a
series of nodes, each node being one of a test node, a result node,
or an error node.
18. The method of claim 3 further comprising expanding and
collapsing the nodes for display.
19. The method of claim 3 further comprising adding a reference to
a data field in another system and a list of values.
Description
TECHNICAL FIELD
[0001] This application relates to determining whether events, such
as taxable events arising from commercial transactions, have
occurred using decision trees.
BACKGROUND
[0002] Computer systems are often used to determine the tax
consequences of a commercial transaction. Specifically, they are
often used to make a tax determination, which involves a
determination as to whether the commercial transaction resulted in
a taxable event, and if so, what type of taxable event has resulted
and to what taxing authority is a tax owed. In addition, computer
systems are often used to perform the calculation of the amount of
taxes that are owed.
[0003] There are many different types of transaction taxes and many
different taxing authorities. It is desirable in some cases to have
tax determination application software capable of handling tax
determinations for a variety of different transaction taxes and a
variety of different taxing authorities. To do this, a tax
determination application software module may be designed so that
it can be called from different applications that handle different
types of commercial transactions.
[0004] The desirability of having tax determination application
software being capable of handling many different transaction taxes
is important from the perspective of an end user who conducts a
wide variety of different commercial transactions that give rise to
the different taxes. This capability is also important from the
perspective of a software vendor who wants to make available for
customers a high degree of pre-configured tax determination
capability, even though any one customer may only use the software
to handle a limited number transaction taxes.
[0005] One challenge in designing tax determination application
software having broad functionality in terms of the number of
different transaction taxes it can handle lies in developing (and
perhaps even more importantly, maintaining) the software code that
embodies the tax rules for the many different tax determinations.
This is because tax rules are complicated and often change. A tax
expert may be needed to design, develop and maintain the software
that implements the tax rules, but in many cases tax experts do not
have much experience in coding software.
SUMMARY
[0006] Generally, the invention, in one aspect, provides the
capability to perform determinations as to the occurrence of a wide
variety of different events using executable decision trees that
are easily developed and maintained. One area where the invention
may find specific applicability is in determining the tax
consequences of commercial transactions. The invention may also
find applicability in any kind of decision which needs a high
flexibility of configuration and which typically has a high number
of variables on which the determination depends.
[0007] In one aspect, a computer program product, tangibly embodied
in an information carrier, determines whether an event has occurred
using a decision tree. The computer program product is operable to
cause data processing apparatus to proceed to and execute a first
of at least one test node of the decision tree. Each of the at
least one test node includes a comparison field and at least one
allowed value. The execution of each of the at least one test node
includes accessing a data value associated with the comparison
field, and determining whether the accessed comparison field data
value is equal to the at least one allowed value. If the comparison
field data value equals any of the at least one allowed value,
execution proceeds to a child node of the test node. Otherwise,
execution proceeds to a sibling node of the test node. The child
and sibling nodes are either another test node, a result node, or
an error node. The computer program product is also operable to
cause data processing apparatus to execute any further test node to
which execution of the decision tree proceeds until execution
proceeds to a result node or an error node.
[0008] In different implementations, the events that correspond to
the plurality of decision trees may be taxable events arising from
a commercial transaction, and the test nodes may represent a
question that forms the basis for a determination of whether or not
the taxable event has occurred. Alternatively, the event may be a
fiscal code as must be printed on invoices in Brazil, an
identification of a delivering plant according to certain
optimization rules, an identification of bank and payment method
for settlement. The computer program product may also be further
operable to cause data processing apparatus to identify for
execution a set of a plurality of decision trees stored in a
decision tree repository, wherein each decision tree corresponds to
a different event, and then execute each of the identified set of
decision trees as described. Each of the plurality of decision
trees stored in the repository may embody a form comprising a
mark-up language, such as an extensible mark-up language, or XML.
Each of the at least one test node may represent a question that
forms a basis for a determination of whether or not the taxable
event has occurred.
[0009] In another aspect, a repository of a plurality of executable
decision trees is provided in a system that determines which of a
plurality of different taxable events have occurred for a
commercial transaction. Each tree determines whether a different
taxable event has occurred. Each tree includes at least one
executable test node. Each executable test node includes a question
to be answered in a determination of whether or not the taxable
event corresponding to the tree has occurred. Also, each executable
test node, when executed using information about a specified
commercial transaction, yields an answer in one of a first form or
a second form. An answer in the first form causes a child node to
be executed, and an answer in the second form causes a sibling node
to be executed. Each child and sibling node is either another test
node, a result node or an error node. Each executable test node
also includes at least two result nodes corresponding to a
determination of whether the taxable event corresponding to the
tree has occurred.
[0010] Each decision tree may correspond, for example, to a
transaction tax type. The transaction tax type for one of the at
least one decision trees may be a value added tax or a use tax, to
name just a few examples. Also, a decision tree corresponding to a
transaction tax of a specified type is able to determine whether or
not a taxable event of the specified type has occurred in a
plurality of taxing locations.
[0011] In yet another aspect, a computer-implemented method
generates a user interface for developing and maintaining a
decision tree. The method associates a decision tree with a series
of user interfaces to request decision tree information from a
user, displays each user interface in the series of user interfaces
associated with a particular decision tree, receives decision tree
information from a user using at least one of the user interfaces,
and stores the decision tree information received from the user.
The decision tree is a series of nodes, with each node being one of
a test node, a result node, or an error node.
[0012] The user interface may be expandable and collapsible to
display the nodes in the decision tree. The method may include the
ability to add a reference to a data field in another system and a
list of values.
[0013] The details of one or more embodiments of the invention are
set forth in the accompanying drawings and the description below.
Other features, objects, and advantages of the invention will be
apparent from the description and drawings, and from the
claims.
DESCRIPTION OF DRAWINGS
[0014] FIG. 1 is a block diagram of a system incorporating various
aspects of the invention.
[0015] FIG. 2 is a flowchart of a process for making tax
determinations and calculations, which may be performed by a system
such as the system shown in FIG. 1.
[0016] FIG. 3 is a figure illustrating a generic structure for an
executable decision tree, which may be used in a system such as the
system shown in FIG. 1.
[0017] FIG. 4 is a figure illustrating a data structure for header
information and a test node for an executable decision tree, which
may be used in a system such as the system shown in FIG. 1.
[0018] FIGS. 5-9 are figures showing the structure of XML documents
making up an embodiment of a decision tree, which may be used in a
system such as the system shown in FIG. 1.
[0019] FIG. 10 is a diagram illustrating the execution of an XML
decision tree document, which may be used in a system such as the
system shown in FIG. 1.
[0020] FIG. 11 is a block diagram of a system for creating a
revising XML decision tree documents, the figure also illustrating
data flows that occur during such processes.
[0021] FIG. 12 is a flow chart of a method for creating decision
trees, which may be used in a system such as the system shown in
FIG. 11.
[0022] FIGS. 13-22 are screen snapshots of a user interface
displayed when decision trees are created or revised, which may be
displayed on a screen of a system such as the system shown in FIG.
11.
[0023] Like reference symbols in the various drawings indicate like
elements.
DETAILED DESCRIPTION
[0024] A transaction tax engine (TTE) 10, shown in FIG. 1, is a
computing system that makes a tax determination for a commercial
transaction. The tax determination includes determining whether a
taxable event has occurred, what type of taxable event has
occurred, and to what taxing authority a tax is owed. Some examples
of taxable events for a transaction include a value added tax
(VAT), a sales tax, and a use tax. In some cases, a transaction may
result in more than one taxable event. After making the tax
determination, the TTE 10 in the FIG. 1 example also calculates an
amount owed in taxes for each taxable event.
[0025] In the FIG. 1 implementation, the TTE 10 is a stand-alone
system that may be called from various calling systems 20. By way
of example, the TTE 10 may be called from a sales transaction
system or from a billing system, to name a few. As such, the TTE 10
may serve as a single system to make tax determinations and
calculations for a wide variety of different transaction taxes and
for a number of different taxing authorities (countries, states,
provinces, etc.). In other implementations, the TTE 10 may be a
part of a single system that both executes a transaction and
performs the tax determination and calculation.
[0026] A tax determination module 30 in the TTE 10 performs the
function of determining which taxable events have occurred. To do
this, the tax determination module 30 uses executable tax
determination rules stored in a repository 40. The rules stored in
the repository 40 are in the form of executable decision trees. In
one implementation, there is a different decision tree for each
type of transaction tax and for each taxing authority. For example,
there may be one decision tree for a VAT for Germany, one decision
tree for a sales tax in the U.S. state of California, one decision
tree for a use tax in the U.S. state of Washington, etc. The
repository 40 in this implementation is located within the TTE 10.
In other implementations, the repository 40 may be located
elsewhere in a networked environment, but accessible by the TTE 10.
For example, the repository 40 may be in a separate computer system
that is maintained by a vendor who develops decision trees for
various different transaction taxes, and then allows them to be
accessed. Alternatively, decision trees that are designed,
developed and maintained by vendors may be uploaded to the TTE 10
and stored in repository 40, as will be described later.
[0027] A tax calculation module 50 in the TTE 10 performs the
calculation of taxes. To do this, the tax calculation module 50
uses different calculation procedures for each of the different
taxes that may need to be calculated. Also, the module 50 accesses
information stored in a tax rate repository 60 and in a tax
exemptions repository 70. In the FIG. 1 example, the tax
calculation module 50 is a part of the TTE 10. In other
implementations, the module 50 may be located elsewhere, and called
by the TTE 10. An administration module 80 is also a part of the
TTE 10 in the FIG. 1 implementation. The administration module 80
allows maintenance to be performed on the TTE 10, such as the
creation and revision of the tax decision trees that are stored in
the rules repository 40. This is described in more detail later.
Module 80 also controls the performance of tax simulations.
[0028] Referring to FIG. 2, a method 200 that the TTE 10 in FIG. 1
may use to determine and calculate transaction taxes starts with a
remote function call from the calling system 20. When this occurs,
the TTE 10, at step 210, compiles an input document. The input
document contains all of the information needed to make the tax
determinations and to make any tax calculations that may need to be
made. The input document may be compiled, for example, by
extracting data from the calling system 20 or from any other system
that contains the needed data.
[0029] The next step 220 is to conduct the tax determination. To do
this, the tax determination module 30 (FIG. 1) first uses the input
document to determine which of the tax decision trees should be
executed. For example, if the input document indicates that the
transaction occurred within the jurisdiction of a single taxing
authority (for example, entirely within Germany), then the tax
determination module 30 will identify all of the tax decision trees
for that taxing authority for execution. If, however, the
transaction occurs within several jurisdictions (for example,
because the parties to the transaction are from different
jurisdictions), then the tax determination module 30 will identify
all applicable decision trees. For example, if there is a
transaction involving both Germany and France, all applicable
German and French decision trees will be taken into account,
although of course not all of the decision trees need lead to a
taxable event. For example, in the example transaction involving
Germany and France, the result may be that there is a taxable event
in Germany, but not in France. These results are determined by the
content of the decision tree, and are taken into consideration by
the designer of the tree. After identifying the decision trees to
be executed, the tax determination module 30 causes the execution
of each of the identified tax decision trees (in this example, each
of the tax decision trees for Germany transaction taxes) to
determine if any taxable events have occurred.
[0030] Next, at step 230, the TTE 10 determines which tax
calculation procedure is to be executed to make the tax
calculation. After this has been done, the tax calculation module
50 is called, and more specifically the determined tax calculation
procedure is called, and then the called procedure makes the tax
calculation (at step 240). At step 250, an output document is
compiled, which may contain, for example, information regarding the
type of tax that is owed, the taxing authority to which the tax is
owed, and the amount in taxes due. For some transactions, more than
one type of tax may arise, and thus the output document may provide
a summary of information for all taxes due. The output document may
be sent back to the calling system 20 or to another system designed
to receive the information.
[0031] FIG. 3 is a flowchart for a generic version of an executable
tax decision tree 300. FIG. 3 illustrates a design methodology for
the decision tree, which may be applied to all decision trees
developed for a particular system. For example, the tax
determination rules repository 40 (FIG. 1) may include many
different tax decision trees, all designed according to the
methodology illustrated in FIG. 3. As discussed previously, there
may be, for example, one tax decision tree for a VAT for Germany,
one decision tree for a sales tax in the U.S. state of California,
one decision tree for a use tax in the U.S. state of Washington,
etc.
[0032] The decision tree 300 has header information 302, which
uniquely identifies the decision tree so that it may be called for
execution from among all decision trees in the repository 40. In
the FIG. 3 example, the header information 302 also points to a
test node 304, which serves as a starting point in a tax
determination using the tree. Test node 304, similar to other test
nodes in the decision tree 300, includes a question that needs to
be determined as part of the tax determination. For example, in the
case of a tax determination for Germany VAT, one question may be
whether the transaction is a service transaction.
[0033] The form of the answer to the question for a test node is
either "yes" or "no." If the answer to the test node is "yes,"
execution of the decision tree proceeds to a child node. If the
answer is "no," execution proceeds to a sibling node. Therefore, if
the answer to test node 304 is "yes," then the execution of the
decision tree proceeds to another test node 306, which is a child
node to test node 304. On the other hand, if the answer to test
node 304 is "no," then execution proceeds to test node 308, which
is a child node to test node 304. Test nodes 306 and 308 are
similar to test node 304 in that they include a tax related
question in a "yes" or "no" format, and execution similarly
proceeds from each of these nodes to a sibling node if the answer
to the question is "no," and to a child node if the answer is
"yes."
[0034] In the case of test node 308, a "no" answer causes the
execution of a sibling node, which in this example is an error node
310. An error node indicates that the answer can lead to no valid
result. If an error node is reached during execution of a decision
tree, an error indication will result. In the case of test node
306, a "yes" answer leads to the execution of test node 312, and a
"yes" answer to that node leads to the execution of a result node
314. A result node indicates for example, that there has been a
taxable event, or there has not. Execution of the nodes of the
decision tree 300 proceeds through test nodes until reaching either
a result node or an error node.
[0035] FIG. 4 shows an example data structure 400 for a decision
tree's header information 410 and a test node 420. In this example,
the header information data structure 410 includes a decision tree
identifier, for example, a key to the decision tree, which uniquely
identifies the tree from other decision trees. The header data
structure 410 in this example also includes an initial node
identifier, which points to the first test node that is executed
(for example, as is shown in FIG. 3). In some implementations, an
initial node identifier is not needed, for example, if the decision
trees are implemented in XML.
[0036] The header information data structure 410 also includes a
textual description of the decision tree, for example, "Tax
Determination for Germany/VAT." Also, the header information data
structure 410 includes a country field (or the taxing authority), a
tax type field (for example, VAT, sales, use, etc.), and a result
field which indicates the possible results that may occur upon
execution of the decision tree (for example, whether the tax is
due, not due, suspended, deferred, etc.). In addition, the header
data structure 410 may include context information which will be
described later, status information such as whether the decision
tree is active or in development, and version information, which
may indicate, for example, the version number of the tree and
perhaps the date the version was created.
[0037] The test node data structure 420 includes a node identifier,
so that other nodes may point to the node for execution, which
again, may not be needed depending on the implementation. Next, a
display text field contains a textual description of the question
for the node. This description may be displayed, for example, to a
designer of the decision tree, as will be shown later. A comparison
field specifies a field in the previously discussed input document
(or alternatively a field in a calling system) where information
needed to answer the question of the node is located so that it may
be retrieved. A value list field identifies the values that the
comparison field must equal for a positive result. In other words,
a value in the comparison field is compared to one or more values
in the value list field. A positive node identifier and a negative
node identifier each identify the node to be executed if the answer
to the node is, respectively, positive ("yes") or negative ("no").
Positive and negative node identifiers are not needed in all
implementations; for example, they are not needed in an XML
implementation.
[0038] A decision tree interpreter (included in the tax
determination module 30 shown in FIG. 1, for example) evaluates the
input document in the test nodes. The basis for comparison is a
field catalog (which is described in detail in connection with FIG.
11) which contains a list of allowed fields to be used in the
decision tree. When designing the tree, any of the fields from the
field catalog may be used to define tests (questions). The input
document contains all of the fields included in the field catalog.
It is the duty of the calling application 20 (FIG. 1) to provide
all of these fields defined in the field catalog, and map them
accordingly in the case where the data model of the calling
application 20 is different.
[0039] The decision tree interpreter, in one implementation, has
two parts: a generic part for the processing of the nodes (for
example, if a test node returns "true," go to the sibling node,
else to the child node, etc.), and an implementation-specific part
which provides a link between the input document and the decision
tree interpreter. For example, if one test node contains the
question "Is the product a service?", then technically a parameter
describing the product property of being a service is compared to
the reference parameter in the test node. The
implementation-specific part has a method for each parameter of the
field catalog, and returns parameters to the decision tree
interpreter for comparison. However, there are some parameters
which are not uniquely defined (that is, the information exists as
one single parameter in the input document), but rather may be a
list of parameters. For example, the input document contains a list
of all "business partners" involved in the transaction (that is,
all of the parties to the transaction), and typically there will be
two such business partners involved in the transaction.
[0040] If a test node needs to evaluate a business partner property
(for example, the nationality of the business partner), then a
"Context" parameter may be needed to define which of the business
partners for which the parameter is required (for example, that of
the supplier or that of the recipient). In this example, the
context parameter for the test parameter "Nationality" may be, for
example, the partner role "Recipient." It is possible to have
multiple context dependencies as well. In many cases, parameters
depend on the country and the tax type, for example, the taxability
of the product. For example, one product may be classified as
"Fully Liable to US Sales Tax," "Not Liable to US Service Tax,"
"Fully Liable to Mexican VAT," and "Partly Liable to Mexican Luxury
Tax." The input document would contain a list of all product
taxability information with country and tax type as key fields. By
specifying the context parameters country and tax type, the
decision tree interpreter knows which parameter of the list in the
input document to use for the comparison.
[0041] In one implementation, the tax determination decision trees
are implemented as XML documents, and may be processed by an XML
processor. An example of a tax determination XML document is shown
in FIGS. 5-9. The XML document depicted in FIGS. 5-9 may be
implemented by a single XML document that comprises the decision,
or may alternatively may be implemented by several XML documents
that collectively comprise the decision tree.
[0042] Referring first to FIG. 5, the XML tax determination
document contains <Targets> tags and <GlobalContexts>
tags. The first element 504 after the root element 502 defines the
"Targets" of the tax determination, that is, the field names of the
desired results of the determination. In this example, the fields
"TaxIncidence," "TaxEventType" and "Tax Event" are the "Targets"
that should result from the tax determination. It is also possible
to define additional fields, for example, a fiscal code or some
textual information. The desired fields are described
in<Target> child-elements having the attribute "Field," or if
a result may yield more than one possible value, a child-element
having the attribute "Table." In this example, there may be several
"Tax Events" to which a single "Tax Incidence" belongs, and hence
the "Tax Events" result is expected in table form.
[0043] The "Tax Incidence" result is a description of the tax
situation pertaining to one specific tax type. For example, the
result may be that a certain tax is due, or not due. The "Tax
Event" result is an ensemble of "Tax Incidences" for all possible
tax types. The "Tax Event Type" result is a legally defined
distinct type of tax levied under a certain "Tax Event" (for
example, VAT, sales tax, mineral oil tax, withholding tax, etc.).
The "Tax Event" result characterizes a business transaction from a
taxation point of view (for example, domestic sales tax, zero-rated
export, etc.), and defines which "Tax Event Types" are due and how
they are calculated (which determines the tax procedure, as
referred to in step 230 of FIG. 2). The "Tax Event" result may also
be used for tax reporting, and is a summary of its underlying "Tax
Incidences."
[0044] The next element 506 of the XML document defines "Global
Contexts" that are constant for the whole tree. In this example,
the fields "Country" and "TaxType" are the "GlobalContexts" that
are used in the tree. In this example, the value for the "Country"
field is "DE," or Germany, and the value of the "TaxType" field is
"VAT." These global contexts are used in the execution of the
decision tree where applicable, such as a key field for accessing
dependent parameters. A "Context" is needed, as discussed
previously, for extracting the proper information from the input
document for comparison with the test parameters in the decision
tree. A "Global Context" is a kind of default value relevant for
the whole decision tree. In addition, each test node may require
the definition of "local" context variables, as will be described
later.
[0045] Referring now to FIG. 6, the counterpart to the
<Targets> tag shown in FIG. 5, which describes the desired
results, is the <Results> tag 602 shown in FIG. 6, which
contains the actual found values resulting from the tax
determination. The <Results> tag 602 contains one
<Result> for each <Target> defined in the
<Targets> tag 504 (FIG. 5). Each <Result> has either a
"Field" attribute and an associated "Value" attribute, or a "Table"
attribute and associated "Value" tags. For example, where the
<Target> allowed for a table of results (as is the case with
the "TaxEvent" target from FIG. 5), these results are stored
in<Value> tags, rather than being stored as a single "Value"
attribute. The <Results> tag 602 may only have <Result>
tags as child elements, and no sibling elements, since the
<Result> tags represent an endpoint of a decision tree.
[0046] As shown in FIG. 6, the value for the field "TaxIncidence"
is "due," and the value for the "TaxEventType" field is 1, which
may correspond, for example, to a VAT. The three values for the
table "TaxEvent" are "Domestic VAT," "Domestic VAT+OIL," and
"Domestic VAT+ALC." In this example, "OIL" may be a German mineral
oil tax, and "ALC" may be a German alcohol tax (both taxes are
excise duties).
[0047] If the tax determination leads to a non-valid result,
an<Error> tag 702, shown in FIG. 7, may be used instead of a
<Results> tag as an endpoint of the decision tree, as shown
previously in FIG. 3. Referring again to FIG. 7, the <Error>
tag 702 has the attributes "ErrClass," "ErrNo," and "ErrText,"
which provide, respectively, the class of the error, the number for
the error, and a textual description of the error.
[0048] The actual tax determination is done by sequential decisions
that are represented by <Test> tags, examples of which are
shown in FIGS. 8 and 9. Each <Test> tag allows for the
comparison of a specified parameter (identified by the "Field"
attribute) against a list of allowed values (identified by the
"Value" attributes). These allowed values may be defined in a fixed
way (for example, when checking against customized values), or the
allowed values might be determined dynamically (that is, comparing
with another parameter).
[0049] The outcome of the comparison determines the next node that
is processed. In the case of a positive result ("true," or in other
words, the specified parameter equals one of the listed allowed
values), the next child element of the <Test> element is
processed. For example, for <Test> element 802 shown in FIG.
8, if a positive result ("true") is obtained, then the child
element of <Test> element 802 is processed, which is
<Test> element 804. In the case of a negative result
("false"), the next sibling element of the <Test> element 802
is processed, which is <Test> element 806.
[0050] Each <Test> element contains the attribute
"Description," which may serve as documentation for the person
developing the decision tree, or may be used for tracing. In the
FIG. 8 example, the "Description" attribute for <Test>
element 802 is "Check, if Service or Delivery," which is a textual
description of the test that element 802 conducts. The "Field"
attribute for <Test> element 802 is "ProductTaxabilityType."
It is the value for this attribute for a transaction that is tested
against the allowed values specified in the "Value" tags for
<Test> element 802. The allowed values specified by the
"Value" tags for <Test> element 802 are "Service" and
"Catalog." As such, if, during execution of the decision tree, the
"ProductTaxabilityType" for the transaction equals either "Service"
or "CatalogService," then a positive result would be made for the
<Test> element 802 and execution would proceed to
<Test> element 804.
[0051] If a parameter being tested is not a unique among all
attributes for a given transaction, but may occur multiple times,
the usage of <Contexts> tags allows for the differentiation
of which parameter is to be tested. This is illustrated in FIG. 9,
which shows that for a <Test> element 902 has a "Field"
attribute 904 of "Country." In this example, the attribute
"Country" is not unique, and so the attribute is further defined by
"Contexts" tag 906, which defines that that the specific "Country"
attribute to be tested is the "Country" attribute with a context of
"Role=SF." In other words, the country that is tested is the
country of the entity, or "Business Partner," that is serving in
the role of "ship from." In addition, allowed values may also be
dynamically determined. Instead of using a fixed <Value> tag,
a <ValueOf> tag may be used to derive acomparison value.
Again, a <Contexts> tag can be used to further specify the
value. As illustrated in FIG. 9 where the <Test> element 902
has a <ValueOf> tag 908, the allowed value is the "Country"
of the Business Partner having a "ship to" role.
[0052] FIG. 10 is a diagram showing the processing of a decision
tree XML document during run-time. Referring also to FIG. 1, the
tax determination module 30 may include XML processing capability
to process XML documents as shown in FIG. 10. As discussed
previously, the tax determination module 30 identifies a decision
tree within the repository 40 for processing, and also may initiate
the XML processing capability. Referring now to FIG. 10, an XML
processor 1000 processes a decision tree XML document 1010. In so
doing, the XML processor 1000 will use information that was
compiled into an input document 1020, as discussed previously. The
XML processor 1000 generates an output document 1030 with results
of tax determination made using the decision tree document 1010 and
the information from the input document 1020. The input document
1020 and the output document 1030 may also be in an XML format, or
they may not be.
[0053] Turning now to the creation and revision of decision trees,
FIG. 11 shows an example system and process flows for performing
this function. In the example shown in FIG. 11, the decision trees
are implemented in XML. A maintenance processor 1100 controls the
process of creating and revising decision trees, and may reside,
for example, in a workstation that a developer uses. A video
display screen 1110 connected to the maintenance processor 1110
provides the developer with a graphical view of a decision tree
being created or revised.
[0054] In the FIG. 11 implementation, decision trees 1122 and
associated texts 1124 are stored in memory 1120. Also, the decision
trees 1122, in this example, are XML documents, as described
previously. As such, a global XML schema 1130 is available for
basic XML document validation. Each separate XML document (three of
which are shown in FIG. 11) may be a different decision tree;
alternatively, more than one XML document may make up a single
decision tree. The memory 1120 shown in FIG. 11 may be, for
example, the repository 40 shown in FIG. 1. Alternatively, the
memory 1120 may be memory residing in a design computing
environment, in which case the decision trees 1122 and associated
text 1124 may be uploaded to a transaction tax engine (TTE) 1140,
as shown by the arrows in FIG. 11. This upload may be made, for
example, to the TTE 10 shown in FIG. 1, and specifically to the
repository 40 of that TTE 10.
[0055] The XML documents 1122 are read and parsed by an XML parser
(not shown in FIG. 11). This provides a parsed document object
model (DOM) representation 1150 of the XML decision tree documents
1122. Data to be shown on the maintenance processor screen 1110 is
read from the DOM representation 1150, and conversely, the DOM
representation 1150 of decision trees is updated based on a
designer's input being made to the maintenance processor 1100.
[0056] A field catalog 1160, described previously, is provided that
defines which fields are available for use in a decision tree (when
defining node details). The fields that are available in the field
catalog 1160 may depend on a "decision tree type" indicator. The
"decision tree type" indicator specifies the use of the decision
tree. In the examples thus far described, the use of the decision
tree is to determine whether taxable events have occurred. Other
types of decision trees may be, for example, any kind of
determination which needs a high flexibility of configuration and
which typically has a high number of variables on which the
determination depends may be suitable. For example, the decision
trees may be used for determining a "Fiscal Code" for Brazilian
transactions. The Fiscal Code is an official code to classify
business transactions, and the proper Fiscal Code is required to be
printed on invoices. The determination of the Fiscal Code depends
on many parameters, including for example, the type of merchandise,
type of transaction, type of business partners, origin and usage of
the materials, etc.). Other examples of how the decision trees
could be used are for determining a delivering plant according to
certain optimization rules (geographical, availability, customer
preferences, etc.), and for determining a bank and payment method
for settlement according to certain criteria (such as vendor
location, contracts, cash flow optimization, etc.)
[0057] As such, the maintenance processor 1100 may support more
than one type of decision tree. Where this is the case, the
"decision tree type" indicator may be used to determine a
validation schema for the XML representation of a designed decision
tree, because the validation schema may be different depending on
the decision tree type. The "decision tree type" indicator may also
determine the contents of the field catalog 1160 that are available
for use in a decision tree being designed or revised.
[0058] FIG. 12 shows an example process 1200 that may be executed
by a maintenance processor (such as the processor 1100 shown in
FIG. 11) to develop a decision tree. The process 1200 begins, at
step 1205, where the processor allows a user to enter decision tree
information. This may include, for example, header information such
as that shown in FIG. 4. When the decision tree information is
entered, a starting node is generated and displayed on the
developer's screen (step 1210). This node will typically be a test
node for the decision tree, and specifically will be the first test
node of the decision tree that is executed at run time. For that
node, the process 1200 proceeds to a test node configuration
process 1215, where the processor receives user input of display
text for the node (step 1220), such as text describing the question
that the node is answering. As part of the node configuration step
1215, the processor also receives an identification of the node's
comparison field (step 1225) and the node's value list (step
1230).
[0059] After the starting test node has been configured, additional
test nodes may be selected and similarly configured by test node
configuration process 1215. Step 1235 determines if there is
another test node in the decision tree, and if so, the process 1215
may be performed for another node as indicated by the flow chart.
As discussed previously, the structure of test nodes is such that
each test node has an associated child node and sibling node.
Therefore, the user indication that there is another test node may
be made by a user's selection that a sibling or child node is going
to serve as a test node.
[0060] At step 1240, the result nodes of the decision tree may be
selected and configured. Result nodes will be either a child node
or a sibling node of a test node, or both. As such, receiving a
user's indication that a particular child or sibling node is to
serve as a result node may be the user action that selects a
results node. At step 1240, the user is also prompted to receive
information regarding the result (for example, tax is due, not due,
suspended, deferred, etc.)
[0061] At step 1245, the error nodes of the decision tree may be
selected and configured. As with the result nodes, the error nodes
will be either a child node or a sibling node of a test node, or
both. As such, receiving a user's indication that a particular
child or sibling node is to serve as an error node may be the user
action that selects the error node. At step 1240, the user is also
prompted to receive information regarding the error that is
indicated if the error node is reached during execution of the
tree. After all of the test, result and error nodes have been
configured, the decision tree is validated (step 1250) and stored
in memory (step 1255).
[0062] Persons knowledgeable in the art will understand that the
design process need not follow the steps shown in FIG. 12 in the
order indicated. For example, some result nodes may be configured
before all the test nodes are configured. In addition, decision
trees stored in memory may be recalled from memory and revised if
necessary, using a process similar to that shown in FIG. 12.
[0063] An example of a user interface that is displayed to a user
who is developing or revising a decision tree will now be discussed
in detail. FIG. 13 shows a screen snapshot 1300 of such a user
interface, which presents a logical outline of the decision tree to
the user. The user interface is capable of displaying the different
decision tree nodes that represent a test, a result or an error
condition, and displaying the interrelationships between the
various nodes of the tree. In FIG. 13, a left-side portion 1302 of
the user interface displays the decision tree outline using a
hierarchy control. A right-side portion 1304 of the user interface
includes sub-screens that display data associated with the nodes,
and allow the user to enter and revise the data associated with the
nodes.
[0064] When a designer uses the user interface to create a new
decision tree, a header node 1306 and a beginning test node 1308
are created and displayed. The header node 1306 is used to maintain
general information about the decision tree. By selecting
(clicking) the header node 1306, which highlights the display of
the header node 1306, input areas on the right-side portion 1304
are made available for entering or revising header information. For
example, a "Tree Description" field 1310 has an associated input
area for a textual description of the decision (for example, a "Tax
determination for Germany/VAT" in this example. A "Tree ID" field
1312 has an associated input area for a number ("100") that
uniquely identifies the decision tree from other decision trees. A
"Country" field 1314 identifies the country, or taxing authority,
for the decision tree, and a "Tax Type" field 1316 identifies the
tax type for the decision tree (for example, VAT, sales tax, etc.).
In this example, fields 1314 and 1316 are not yet filled in by a
user, but would be filled in with "DE" for the "Country" field, and
"VAT" for the "Tax Type."
[0065] The user may also input additional information about the
header node using data entry areas associated with a "Target" tab
1318. This tab 1318 in the FIG. 13 example is located below the
header field area described above, and is used to input information
about the "Target" fields, which are the fields to be determined by
the decision tree. A "Field Name" field is used to declare the
result fields that are to be determined by the decision tree. The
user can choose whether a result field will consist of a single
value or a list of values by selecting either the "Single Value"
button or the "Value List" button, respectively. The user may use
the "Description" field to input a description of a corresponding
"Field Name" field. For example, by entering different field names,
the decision tree may be configured to determine whether a tax
type, which is a legally defined distinct type of tax that is
levied (e.g., value-added tax, sales tax, mineral oil tax, etc.) is
implicated, whether a tax incidence, which describes the tax
situation pertaining to a particular tax type, has occurred or
whether a tax event, which characterizes a business transaction
from a taxation point of view, has occurred.
[0066] Referring to FIG. 14 and another screen snapshot 1400 of the
user interface, a tab 1420 labeled "Global Context" has associated
data entry areas for information about context fields that are
valid for the entire decision tree. A "Field" field is used to
indicate the name of the field that is valid for the entire
decision tree, a "Field Value" field is used to input the value of
the field that is valid for the entire decision tree, and a
"Description" field is used to input information describing the
"GlobalContext" field.
[0067] Referring to FIG. 15 and another screen snapshot 1500 shown
there, a "Status" tab 1522 is used to view status information about
the decision tree. The "Status" tab 1522 displays information such
as the decision tree's creation date and time, an identification of
the person who created the decision tree, a date that the decision
tree was last changed, a time when the decision tree was last
changed, and an identification of the user who last changed the
decision tree. This information provides a history of the decision
tree.
[0068] Once the configuration of the header node of the decision
tree, by entering information into the "Target" and "Global
Context" tabs is completed, a user may now configure the decision
tree to make a tax determination by adding and modifying nodes. The
tax expert may create a new node by right-clicking on a "yes" or
"no" answer node of a logical test node. A context menu 1524 will
appear and provide the user the option to insert a test node,
insert a result node or insert an error node. By selecting one of
these options, the user establishes the type of node that the node
will be.
[0069] Referring to another screen snapshot 1600 shown in FIG. 16,
when a user chooses to insert a test node, a new group of three
nodes will be created. A logical test node is created and displayed
to the right of the selected answer node. In the example shown in
FIG. 16, a logical test node 1626 is created next to a yes node
1628. A "yes" node 1630 and a "no" node 1632 are created below the
newly created logical test node 1626. When the new test node is
created, the sub-screen on the right side of the user interface
also changes to allow the tax expert to define the attributes of
the new test node 1626. On the attribute sub-screen the user may
input a description of the test in the "Description" field and the
name of the field to be tested in the "Field Name." Below these
fields are three sub-screen tabs. The user may use these tabs to
define the test node. By selecting the "Context" tab 1638, a list
of context options may be input by the user for the "Field Name"
field 1636. The user may enter the name of the context option
field, field values, and a description of the context option field.
In the example, the field named "Country" is tested for the context
option "Role" field whose value is "Ship From."
[0070] Referring to yet another screen snapshot 1700 shown in FIG.
17, it is shown that by selecting a "Value" tab 1740, a user may
define one or more acceptable values for the question field using
the "Field Value" and "Description" fields. Therefore, the decision
tree will determine whether the value of the field to be tested
equals the value of one of the allowed values among the list of
acceptable values provided by the tax expert. Referring to FIG. 18
and screen snapshot 1800, it is shown that by selecting a "Value
Of" tab 1842, a user may also configure the decision tree to
compare two dynamic field values. In the example, the field named
"Country" when the field "Role" has a value "Ship To" is compared
to the field named "Country" when the field "Role" has a value
"Ship From."
[0071] Referring to FIG. 19 and screen snapshot 1900 shown there, a
user may insert a result node next to a yes or no answer node by
right-clicking on the yes or no answer node and selecting Insert
Result from the context menu that appears. The result nodes of a
decision tree indicate that for the sequence of test questions and
answers that led to the result node, no more tests need to be
performed and the tax determination has found a valid result. In
the example, a result node 1944 has been created next to the yes
answer node 1946 of the test node "Is it a service transaction?"
1948. When the new result node 1944 is created, the sub-screen on
the right side of the user interface also changes to allow the tax
expert to define the attributes of the new result node. By
selecting the "Result Value" tab 1950 in the attribute sub-screen,
the tax expert can input a Result Values "Field" 1952, a value for
the field 1954 and a description 1956 of the field. For example,
the tax expert may configure this result node to indicate that the
sequence of test questions and answer nodes that led to this result
node resulted in the finding of a value "100" for the field
"TXEVENT," which indicates that the business transaction for which
data was provided to the decision tree results in a domestic output
tax being due. Alternatively, referring to FIG. 20 and screen
snapshot 2000, the tax expert may select the "Result Value List"
tab 2060 in the attribute sub-screen, to input a Result Value field
2062, list of values for that field 2062, and a description 2064 of
that field for those target fields (i.e., fields that must be
determined by the decision tree) that were configured as a value
list in the header node.
[0072] Referring to FIG. 21 and screen snapshot 2100 shown there, a
tax expert may insert an error node next to a yes or no answer node
by right-clicking on the yes or no answer node and selecting Insert
Error from the context menu that appears. The error nodes of a
decision tree indicate that for the sequence of test questions and
answers that led to the error node, the decision tree cannot find a
valid result during its tax determination and therefore returns an
error message. In the example, a new error node 2166 has been
created next to the no answer node 2168 of the text node "Is it a
service transaction?" 1948. When the new error node 2166 is
created, the sub-screen on the right side of the user interface
also changes to allow the tax expert to define the attributes of
the new error node. By selecting the "Error" tab 2170 in the
attribute sub-screen, the tax expert can input information about
the application area of the error message, the message
identification number, and message text to describe the error for
future use. Alternatively, the tax expert may use the Insert Test,
Insert Result and Insert Error buttons located in a button-type
menu 2172 of the user interface to create new nodes. Once a node of
a decision tree has been created, a user can modify the information
found in that node at any time by double-clicking the node to
access the fields in the attribute sub-screens. Once information
describing a decision tree (as mentioned above) is input using the
user interface, the transaction tax configuration program module
determines whether any information was received in the fields of
the attribute sub-screens. If so, the transaction tax configuration
program module updates the XML document, text table, and
maintenance table with this new information.
[0073] The user interface may provide a tax expert with additional
functions to modify a decision tree. The user interface may include
buttons that allow a tax expert to delete an existing branch of the
decision tree, cut or copy an existing branch of a decision tree
and paste that branch in another area of the decision tree. An
additional modification or maintenance function of the user
interface includes a check command. The check command may be used
to determine whether there are any open branches (e.g., branches
that are not terminated with either a result node or an error
node), whether each node is defined completely, and whether the XML
document is valid. The user interface also may include a function
that allows the status of every node to be displayed. The user
interface may also provide automatic versioning of decision trees.
When a decision tree is created it is automatically assigned
version "1". When the decision tree is being modified it is given a
maintenance version number. When the decision tree is ready for
productive use, it is released. The released decision tree is also
given a version number. When the released decision tree is
modified, a new version is created automatically by copying the
current version and incrementing the version counter by one. This
modified decision tree must be released again for productive
use.
[0074] As shown in screen snapshot 2200 of FIG. 22, a graphical
presentation of a sequence of logically related questions 2274 and
the answers 2276 to these questions. A decision tree may improve a
user's understanding of a complicated situation and help the user
make more informed decisions. A tax expert may use the user
interface to completely represent a complicated tax situation.
Because the yes and no nodes always represent the yes and no
answers, respectively, to the question located beside the test node
directly above them, it is easy to determine the answers associated
with each question by looking at the user interface. The tax expert
can collapse the nodes he in not interested in and obtain a clear
picture of the portion of the decision tree he is interested in by
using the tree control 2278 associated with a node to expand or
collapse that node.
[0075] The user interface also may allow a user to copy a previous
version of a decision tree into a new or current maintenance
version of a decision tree, copy an existing tree, save a current
tree under a new name, display the current released version and
maintenance version of the decision tree, navigate to a particular
node of a decision tree using a search capability (e.g., search for
a result node with a particular result value), display the complete
decision tree in a printable format, upload an XML document from a
local file into a new decision tree, view the XML source code of an
XML document and delete a decision tree. When a decision tree is
deleted the data is not removed but its status is changed to
deleted.
[0076] A number of embodiments of the invention have been
described. Nevertheless, it will be understood that various
modifications may be made without departing from the spirit and
scope of the invention. Accordingly, other embodiments are within
the scope of the following claims.
* * * * *