The present invention relates to visualization and/or management of the organisation of data within data processing systems.
Users of data processing systems receive or access information on various subjects, from a number of sources and in various formats. Users also generate or modify documents, images, and other types of information which are saved in local storage associated with the user's computer. This enables the user to view and modify the data without needing to be connected to the network. The different types of data are stored in various places on a user's computer system and are viewed and managed by a variety of application programs—each of which is best suited to one particular type of data. Users also work with data and applications which are distributed over two or more computer systems. The data or the applications, or both, may reside on a computer separate from the user's computer system.
Each of the data types is normally organised into its own hierarchical information space. The user creates a taxonomy of folders in which to store copies of e-mails. The user also generates a similar (but seldom identical) taxonomy for organising bookmarks in their chosen browser. Furthermore, the user divides their local drives into hierarchical directory/folder structures. All these structures were influenced by the user's mental model of how things are related, and it is frequently the case that each structure has some features in common with the other data structures that the user has created for other types of data. Nevertheless, the hierarchical data organisation is seldom (if ever) identical across multiple applications.
Maintaining these different types of data and organising the data that is received from the different applications is an onerous task, unless the different data can be organised in a consistent manner so that navigation and management is easier for the user. One approach to try and achieve consistency and hence reduce future workload is to attempt to unify all received data into one format and then access it using one particular application program. This approach of transforming data to a single format (which could be labelled as “coercion”) would require parsing and transformation of most data feeds including data conversion. The coercion approach would present a significant processing overhead and would probably run into conversion problems.
It is also known to “federate” various data sources, so that they appear to be one (logical) data source. Federation is a technique which is used with multiple data sources which differ slightly in terms of detailed representation but not in terms of gross characteristics. Heterogeneous database management systems are a good example of such sources. Federation can be achieved by interposing an additional layer of abstraction between the physical data and one application program which is used to view and manage it. However, the addition of this extra level of abstraction adds undesirable complexity and processing to the user's computer system and could suffer from similar problems to those that afflict the coercion approach. Furthermore, automated federation typically combines independent data sources without taking account of similarities between each source's data organisation.
Both of the previous approaches suffer from the problem that new or modified data formats and advances in the application programs normally used to manipulate each type of data are unlikely to be readily available in either a coerced or a federated system. For these reasons, it remains preferable to use the “native” application program for each type of information. (Examples of ‘application programs’ in this context are Microsoft's Internet Explorer or Netscape's Navigator for Web browsing, Microsoft® Outlook for e-mail, or Lotus Notes® software for e-mail and data management. Lotus Notes is a trademark of International Business Machines Corporation, Microsoft is a trademark of Microsoft Corporation in the United States, other countries, or both).
There exists a need for a solution which simplifies the management of data by providing consistency of data organisation for different data types, without the disadvantages of known coerced or federated approaches.
According to a first aspect, the invention provides a method for managing distribution of subscription data within a network of publish/subscribe messaging systems, the method including the steps of; maintaining, in storage associated with a first messaging system, information representing a taxonomy of subscription topics of the first messaging system; forwarding information representing at least a part of the taxonomy of subscription topics to a connected publish/subscribe messaging system; comparing, using topic name matching, forwarded information with a taxonomy of subscription topics maintained for the connected publish/subscribe messaging system, to identify differences and similarities; and in response to identification of differences between the taxonomies, using the forwarded information to generate a modified taxonomy of subscription topics for the connected publish/subscribe messaging system which increases consistency between the taxonomies.
The method preferably includes maintaining, in storage associated with the first messaging system, a change history recording changes to the taxonomy of subscription topics of the first messaging system; forwarding at least a part of the change history to the connected publish/subscribe messaging system; and generating the modified taxonomy by applying, to the taxonomy of subscription topics of the connected publish/subscribe messaging system, a part of the change history corresponding to the identified differences.
According to a second aspect, the invention provides a method for managing data organisation for a computer system, the method including the steps of: generating a reference taxonomy representing the organisation of stored data for one or more computer programs of a computer system; accessing storage associated with a message broker to obtain a taxonomy of message topics used by the message broker to control routing of messages to subscriber computer systems; comparing the reference taxonomy with the taxonomy of message topics to identify matching and non-matching features within the reference taxonomy and the taxonomy of message topics; and in response to a selection of a preferred taxonomy by reference to the matching and non-matching features, storing the preferred taxonomy as a replacement of at least one of the reference taxonomy and the taxonomy of message topics.
The term ‘taxonomy’ is used in this description to refer to a data organisation, such as an hierarchical organisation of message topics, files, folders and directories, or an organisation of bookmarks. A generated reference taxonomy can be used to help a computer user to identify and understand the logical relationships between existing data structures and message topics for a set of managed programs including a publish/subscribe message broker. In preferred embodiments, the reference taxonomy can be updated in response to differences between the reference taxonomy and a broker's message topic taxonomy. The comparison results may also aid users in updating their subscriptions to a publish/subscribe messaging service.
In one embodiment, the generation of the reference taxonomy preferably involves accessing storage associated with a computer program to obtain an application taxonomy for the program, the application taxonomy comprising information defining the organisation of stored data items of the program. A user may then select the application taxonomy as a reference taxonomy—possibly selecting from a number of different taxonomies of different computer programs a taxonomy that most closely resembles the user's mental model of how subject areas are related.
In an alternative embodiment, the reference taxonomy is a user-defined data organisation created using a graphical user interface (GUI) of a data organisation manager program. The data organisation may combine user selected features (e.g. nodes and/or links) of existing taxonomies and user created features. The user-created reference taxonomy is then compared with existing application taxonomies to identify similarities and differences.
The method according to a preferred embodiment includes the step of generating a modified reference taxonomy which aggregates features of the initial reference taxonomy and the compared taxonomy of message topics, wherein one or more new nodes representing data structures are created within the reference taxonomy which correspond to particular topics of interest. The modified reference taxonomy is preferably generated as a candidate reference taxonomy which is presented to a user for the user to reject, accept or partially accept. An accepted candidate reference taxonomy, or the accepted parts, are then stored as an updated reference taxonomy.
The invention can be implemented to help a user to increase consistency between their reference model of how data is organised and the taxonomy of message topics used by a message broker. This can be implemented as an automated or semi-automated synchronization of taxonomies. An example is where a folder is automatically created in a Windows file system in response to identification of a message topic of interest that is identified by comparison between the taxonomies. Different levels of automation can be provided for, and controlled via configuration settings.
The step of comparing taxonomies preferably uses string matching techniques to identify matching node names and topic names, and preferably compares qualified names (representative of node names and relationships between nodes within a taxonomy) to identify nodes and topics having both matching names and relationships within their respective taxonomies. For example, a reference taxonomy and a topic taxonomy may both contain a node whose path is “/computing/scientific/molecular modelling”.
Each application taxonomy is preferably obtained by means of an adapter for the respective computer program. Use of a set of adapters for accessing data storage for a message broker and for application programs having different data structures and data formats, together with a generic computer program which interfaces to the adapters, provides an extensible solution since new adapters can be added to support additional programs as required. The organisational data obtained for a computer program is preferably converted to a generic format when required (such as XML), before being compared with the reference taxonomy. This enables taxonomies for applications having different data types and formats to be compared. Each taxonomy is preferably displayed to a user via a GUI. The GUI preferably enables user selection of a preferred taxonomy and/or preferred features (nodes and links). The GUI preferably also enables the user to create new nodes and links within a reference taxonomy.
One embodiment of the invention provides a method for both visualizing and modifying the organisation of hierarchically organised sets of data structures. The method enables a user of a data processing apparatus to increase consistency between the organisation of data structures for various types of data, associated with different programs within the apparatus. The organisation of a set of information spaces that correspond to different data types is managed to provide increased consistency between the different organisational structures, simplifying or automating subsequent management of data storage and navigation (i.e. locating data via paths through the organisational structures).
The invention is beneficial where the data items associated with each of a set of different programs all conform to some extent to a common hierarchical organisation. Similarities and differences between the hierarchical organisations are identified and the user is prompted to modify the hierarchies if required. In preferred embodiments, certain differences can be disregarded when comparing two hierarchies (e.g. by specifying exceptions). An example where this is likely to be useful is where some elements of a particular data taxonomy are only relevant to a subset of the data types or a single application program.
Methods and systems according to preferred embodiments of the invention are able to handle data in its native formats using the native application program of choice for each data type.
Preferred solutions enable definitions of data structures to be automatically or semi-automatically gleaned from existing data storage structures and then compared, by processes running in a computer. The data structures can then be modified in response to user interactions with a graphical user interface. Further automation of the step of modifying a data organisation is also possible.
Methods as described above may be implemented in computer program code, designed to control the performance of operations in a data processing apparatus on which the program code executes, and the program code may be made available for electronic transfer or recorded on a recording medium.
Further aspects of the present invention provide a data processing apparatus including: a data processor; data storage; one or more computer programs for controlling the operation of the data processor to perform operations on data items stored in the data storage; and a taxonomy manager for managing organisation of stored data in association with the one or more computer programs, in accordance with one or more of the methods described above.
Embodiments of the invention are described in detail below, by way of example, with reference to the accompanying drawings in which:
System Overview
The data processing apparatus typically runs several different computer programs, such as an operating system 20, an e-mail program 30, a Web Browser 40, a word processor, etc. The data processing apparatus has a number of separate data storage structures associated with the different programs, such as files, folders and database tables. The corresponding different types of data are saved into the appropriate data storage structure by a user. Copies of e-mails can be stored into an hierarchical set of folders within an e-mail program, and bookmarks containing URLs for accessing Web pages can be organised into hierarchical folders within a Web Browser program. Similarly, the computer's local drives can be divided into directories and folders containing document and image files, managed by the operating system's file system 50.
Each desktop program has its own hierarchy of folders or directories in which the various files and bookmarks are stored. Each of these hierarchies is known as an application taxonomy. There are often some similarities between the different application taxonomies, since the user-created hierarchies each reflect the user's personal view of how stored data items are related. However, there are typically also many differences between the various hierarchical data structures, and these differences add complexity to tasks such as navigation and management of stored data.
An embodiment of the invention is described below as a set of computer program components which cooperate to control the performance of operations of a data processing apparatus when loaded and executed on the apparatus. It will be apparent to a person skilled in the art that the individual steps of the method described below can be implemented in computer program code and that a variety of programming languages and coding implementations may be used to implement the methods described herein. Moreover, the computer program is not intended to be limited to the specific control flow described below and one or more of the steps of the computer program may be performed in parallel rather than sequentially as described. One or more of the operations described in the context of a computer-program-controlled implementation could alternatively be implemented as a hardware electronics component.
The computer program components may be stored on any computer readable medium, including storage devices such as magnetic or optical disks, memory chips, or other storage devices suitable for interfacing with a general purpose computer. The computer readable medium may also include a hard-wired communication medium such as exemplified in the Internet system, or wireless medium such as exemplified in the GSM mobile telephone system.
Where steps or features in any of the accompanying drawings are referenced by the same reference numerals, those steps and/or features have the same or similar functions or operations in the context of the present description (unless the contrary intention appears).
Taxonomy Manager
A taxonomy manager computer program 60 is installed on the data processing apparatus. The word ‘taxonomy’ as used in the present specification refers to an organisation of data items. The taxonomy manager 60 helps the user to understand his or her own data organisation, by means of a graphical user interface (GUI), and allows management of data storage structures to increase the consistency between the separate application taxonomies associated with a number of the programs running on the apparatus.
Referring to
Adapters
The set of adapters 120 allow the taxonomy manager 60 to read and initiate creation of data structures—such as folders in a Windows operating system's file system or Unix® directories. Other adapters enable reading and creation of data structures used within e-mail and database programs (for example, Lotus Notes information management and communication program from Lotus Development Corporation, or Outlook Express communication software from Microsoft Corporation) or within Web Browser programs (such as Microsoft's Internet Explorer or Netscape's Navigator). The adapters 120 are plug-ins to the taxonomy engine 110 and can be used to traverse directories or other data structures to extract (210 in
In general, a different adapter is used for each different operating system or application program, since different file formats and application programming interfaces (APIs) are used by the various programs. Example data structures include the Internet Explorer “Favourites” folder and Unix operating system directories, and an example data format is Lotus Notes database format.
Relatively simple adapters are used to read file and folder names and organisational information for a file-system-based structure (such as Windows file system folders and directories). Since the hierarchical data structure is necessarily available to applications interacting with the operating system, known interfaces can be used or known parsing techniques can be applied to access the organisation information. For example, the WIN32 API can be used to browse the directory structure within a Windows operating system. Since bookmark folders used by a Web Browser are simply XML files within a directory, the adapter used for a Web Browser incorporates an XML parser which can identify mark-up tags indicating certain types of data structure.
For certain other applications, such as a database or e-mail program, the adapter needs to access folder structures within the application. Adapters for such applications are written with an implementation of product-specific APIs or an ability to interpret product-specific data formats, to access information regarding the internal data structures of each application. For example, a database client program can be used to access database tables within a relational database, and a Lotus Notes client program can be used to access database folders used by the Lotus Notes program.
The storing of data structures, and management of stored data structures, by relevant applications and operating system file systems is well understood by persons skilled in the art and so will not be described in detail herein.
Referring to
An adapter 120 does not retrieve a full copy of an application's data and pass this to the taxonomy engine. Instead, only the information relating to the organisation of data within the data storage used by the application (e.g. names of files, folders and directories and their relationships within an hierarchical file system) has been retrieved 210 and passed 220 to the taxonomy engine. This avoids the duplication of large amounts of data and the significant data processing that is required by many federated systems which rely on transformation of an application's data items to a non-native format.
Taxonomy Engine
The taxonomy engine 110 receives the various organisational data from the adapters 120. Within the organisational data, each node has a name string which is unique within the set of children for a single parent node —i.e. a parent cannot have two children with the same name. Each node therefore has a unique path within the taxonomy, where the path is the fully-qualified name expressed relative to the root of the tree.
As is well known, many programs make use of an hierarchical organisation of data. Although data may be organised in many different ways, including linked tables within a relational database or a flat (one level) array of independent files, the following description relates to the specific example of a taxonomy manager for managing the organisation of data within an hierarchical taxonomy. In the example of hierarchically organised data, the graphical representation is typically a two-dimensional hierarchical ‘tree’ representation (such as shown in
In the description that follows the data organisation taxonomy specific to each separate application is termed the ‘application taxonomy’. The ‘reference taxonomy’ is the master data organisation taxonomy with which the separate application taxonomies are compared. The reference taxonomy represents the user's preference for organising his/her data on the machine.
A candidate for the reference taxonomy may be generated in one of three ways (as described below). Regardless of which of these methods is used, the user is then free to modify the candidate reference taxonomy manually to obtain the final reference taxonomy he or she desires. The user may also use some or all of the three methods. For example, the user could select one of the application taxonomies as a good starting point (method 2 below), manually modify it (method 1 below) and then compare it with some or all of the application taxonomies (method 3 below) to complete the final reference taxonomy.
The three methods of generating a candidate reference taxonomy are:
The Taxonomy Engine can compare any application taxonomy against the reference taxonomy. This will show up similarities and differences and these relationships are remembered by the Taxonomy Engine:
Based on the comparison results, the taxonomy engine then generates 300 a new candidate reference taxonomy—in this case a taxonomy representing an aggregation of data structures and interrelationships within the two sets of organisation information. The candidate taxonomy takes account of similarities between the two sets of organisational information, firstly by representing a pair of matching nodes (one from each of the taxonomies being compared) as a single node in the reference taxonomy. The candidate reference taxonomy takes account of differences between the two sets by aggregating the hierarchical structures of the two sets so that the reference taxonomy retains a representation of those parts of each taxonomy which do not match with the other taxonomy. The aggregation is performed subject to a confirmation of acceptance 310 by the user, and is limited by any exceptions specified by the user in advance (such as parts of a pre-existing taxonomy which are to be disregarded when generating the reference taxonomy). If accepted, the aggregated taxonomy is saved 320. If not accepted, the user may perform modifications 330 via the GUI or simply reject the aggregated taxonomy.
The taxonomy engine can iterate over each of the other application taxonomies in turn, performing simple string comparisons of the node names. This may involve accessing stored application taxonomies or generating them dynamically via the adapters. Where a match is found, this suggests an equivalence relationship—and the user is prompted to confirm or reject the suggested equivalence relationship. Where multiple nodes match by node name (alone), their paths are compared 520 by stepping back through the taxonomy information from the ‘matching’ node to the parent, to achieve further context. These fuller paths are compared (by string comparison) and a number of suggested equivalence relationships are presented 530 to the user. These suggested alternatives are sorted into an order according to the closeness of match between the strings (i.e. using the full context information). Once again, the user can accept or reject the suggested matches at this stage. Where a suggested equivalence is rejected by a user, a confirmation of the rejection is stored and applied to future comparisons to avoid presenting the user with a suggestion that has already been rejected.
In addition to the accepted relationships, the system enables users to specify other relationships (such as equivalence, child, parent) that the user wishes to adopt despite the absence of a match resulting from a string comparison of node names or fuller path names. These relationships can be defined by interactions with the GUI—drawing links and moving nodes—and are stored persistently. In addition to storing the reference taxonomy, together with saved accepted and rejected relationships, a history of changes to the reference taxonomy is also stored persistently.
For each node of a first taxonomy identified as matching a node of the second taxonomy, the qualified-name string matching determines whether the matching node has a matching child node and/or a matching parent node in the data repositories of both of the managed programs. If a matching node has a different set of child nodes in one program from the set identified within the other program (for example, if there are no child nodes in one program), the taxonomy engine generates an aggregated set of child nodes and links which is an aggregation of the two different sets. This step of aggregating sets of nodes and branches can be performed for all matching nodes within the two data taxonomies being compared, and typically results in an expanded tree. An example of this combining of nodes and branches is shown by example in
If a set of nodes at different levels of the taxonomy are identified as matching, the aggregate taxonomy which is generated by the merge operation includes all of the sub-trees of the two separate hierarchies overlayed on each other (unless the user specifies exceptions). Referring to
Note that the proposed reference taxonomy is only a candidate at the time of its creation—a representation of a possible organisation of data which has not yet been applied to the data of either of the compared programs. Secondly, no attempt is being made at this stage to identify matching data content stored in association with the two data storage hierarchies—the only comparison is between the named hierarchical data structures (files, folders and directories) and their relationships (as represented by their qualified-names which specify unique paths within the taxonomy) within each of the two hierarchies corresponding to the data organisation of the two managed programs. The node labels D1 and E1 in
The reference taxonomy is displayed by the GUI as a candidate reference taxonomy. The user can then decide 310 whether or not the candidate reference taxonomy of
In response to an accept or save instruction, the taxonomy engine 320 generates 300 a new XML file for storage of the reference taxonomy, and saves the reference taxonomy persistently to disk. As a separate operation from saving the reference taxonomy (although not necessarily separate in alternative embodiments), the taxonomy engine deploys the reference taxonomy to whichever one or more of the managed programs the user specifies. The deployment flow is shown schematically in
When applied to the data organisations of one or more applications, the taxonomy engine 110 initiates creation, renaming and moving of folders, bookmarks and directories for each of the managed programs for which the reference taxonomy was approved by the user. The taxonomy engine instructs 400 the relevant adapter 120 for an application 30 (for example, an e-mail program, operating system file system, or Web Browser), and the adapter 120 controls 410 the application 30 to create, rename or move data structures 420 to conform to the new taxonomy 170.
In particular, the taxonomy engine invokes a method in the adapter's interface and passes to the method a specification of the operation to perform together with the target node in the application taxonomy on which the operation is to be performed and additional information required for the operation. For example, a move operation would specify to the adapter the current and required paths of the node to be moved. The adapter would use these paths to, firstly, locate the corresponding data element in the application's data storage hierarchy and, secondly, to initiate a move of that data element to a new location in the hierarchy that corresponds to the required path. The actual implementation of the adapter would depend on the application type. An example would be the deletion of an element from one location in an XML bookmarks file and the insertion of an equivalent element into an alternate location in the XML file.
The result of this step is that the data organisation of one or more of the managed programs will conform to the newly generated reference taxonomy. Nevertheless, exceptions can be catered for such that the reference taxonomy aggregates selected parts of the taxonomy of multiple applications, but other selected parts are maintained application-specific and are not reflected in the taxonomies of other applications. These exceptions can be handled by requiring user-confirmation of all updates during the deployment step, but can equally be handled by an initial step in which the user selects (for example, by interaction with the graphical representation of an application taxonomy) any taxonomy features that the user wishes to exclude from the operation of the taxonomy manager. Exclusions of parts of a taxonomy can speed up operation of the taxonomy manager, and is likely to be desirable if the user only wishes to modify a small sub-tree of two application taxonomies and to retain differences for other parts of the taxonomies.
The addition of new data structures and moving of data structures can be advantageously applied to two separate applications to achieve consistency-between them (i.e. the data organisation of both applications is updated with reference to the other). However, subject to the user's requirements, it may be useful to apply the reference taxonomy to only a selected one or more of the applications that were compared (i.e. features of the organisation of data within a first application are applied to a second user-selected application, without changing the data organisation of the first application). The current embodiment can achieve this effect, because the user has an option to reject the addition of any node for any application taxonomy, but in other embodiments an option for applying a reference taxonomy to only a subset of the compared taxonomies may be selected at the beginning of deployment.
An example of the result of deploying the reference taxonomy is where a reference taxonomy contains a data structure for “\personnel\compensation”, which resulted from the existence of an e-mail folder of that name within the data storage structure used by an e-mail application. When the reference taxonomy is deployed to update the local file system, the absence of a file system folder named “\personnel\compensation” triggers an event to create such a folder (subject to confirmation by the user). If the user confirms the event, the adapter for the file system data storage structure is invoked to create the appropriate folder.
In a second example, a bookmark folder “news/weather” may cause a node “/weather” to be added to the reference taxonomy. Other selected applications will then have a node “/weather” added to their taxonomies (subject to confirmation by the user).
When an execution of the taxonomy manager identifies new data structures in one of the sets of organisational information being compared, a new candidate reference taxonomy is generated and the user is prompted to confirm whether the new candidate should be accepted and applied to the managed programs. If accepted, the previous saved reference taxonomy will be updated and corresponding data structures will be introduced into the application taxonomy of each managed program for which the new taxonomy was accepted.
The above description refers to periodic or user-prompted execution of the taxonomy manager. For example, synchronisation of taxonomies could be scheduled to take place once per day or once per week. In an alternative embodiment, described later, the system is configured such that the taxonomies of all of the applications that use the reference taxonomy are updated automatically in response to addition of data structures to one of the taxonomies, without reliance on a timer expiry or user-request to trigger the update of the set of applications. As described above, automated updating is implemented with the user-configurable option that updates must be confirmed by the user since many users will require this safeguard.
The user may decide that some new data structures should be kept application-private and not applied to other managed programs. Any data structure designated as application-private by the user will remain as such in future comparisons of application organisational structure unless the user explicitly requests that the status of the data structure is changed for re-inclusion.
The taxonomy engine also enables modification of the data organisation information in response to user interactions with the GUI, as described in detail below.
Graphical User Interface
The GUI interprets data feeds from the taxonomy engine and generates a graphical representation of the current taxonomy, as described above. In addition, the GUI implements drag-and-drop capabilities and other known graphical user interaction mechanisms (such as a right mouse button click or keyboard shortcut for delete)—enabling the user to add, prune, rename or move sub-trees within the taxonomy.
The user can select a node of the candidate reference taxonomy and drag it to a new location—either overlaying an existing node (see below), or at a new position in the GUI corresponding to an ancestor, sibling or descendant relationship to other nodes, or to create an independent data structure. When a parent node is moved, all descendant nodes are moved with the parent unless the user specifies an alternative action.
When the user drags a node over another they are prompted for whether this operation is a replace or a merge operation. If it is a merge, the node is given a new name with defaults being either of the existing node names. The child nodes are aggregated and the data is aggregated by the adapters if the data stored by the separate applications are stored in similar formats. Where data may be over-written the user may be prompted with dialogues for confirmation; e.g., “File conflict, which do you want me to keep?” Or, “Shall I version them?”.
If it is a replace operation then the node is by default given the name of the replacing node and the content of the replaced directory (Windows Explorer for example) can optionally be deleted or retained, but if retained then the directory is no longer represented in reference taxonomy.
The user can then save the reference taxonomy that results from a set of user-controlled changes to an initial candidate reference taxonomy generated by the taxonomy manager. When saving such a resulting taxonomy, the taxonomy manager also saves a history of all changes. This provides an opportunity to replay the changes against the individual types of data that pertain to the various application programs that the user wishes to manage.
The program code implementation of GUI operations such as described above are well understood by persons skilled in the art, and so are not described herein in detail.
The features of the GUI and taxonomy engine of the taxonomy manager 60 which respond to a user's graphical interactions make it possible to use the taxonomy manager to generate a new taxonomy without necessarily starting from a comparison of two existing taxonomies. This new taxonomy could then be imposed on chosen applications, including the possibility of exporting the taxonomy to a new machine and then deploying the taxonomy there to update chosen applications on the new machine. The options for selective deployment of a reference taxonomy—including exporting to different systems—provides the possibility of a desired taxonomy being made available to all selected applications on a new system at an early stage.
Usage Scenario
As an example of a usage scenario, let us consider an experienced computer user who is developing a software product. The developer uses his Web Browser to locate information on competitive products, new technologies, analyst reviews of the marketplace, and so on. The developer marks useful pages by saving bookmarks/favourites within the Browser. The developer also sends and receives e-mails seeking and supplying information on a number of areas related to the new product, and they store the e-mails in folders within an e-mail client program. The developer also generates product specification or design documents or source code which are typically stored on the C: drive of the developer's computer, organised in directories or folders. Typically, the developer also receives documents or presentations as attachments within e-mails and detaches the attachments to save on the developer's hard disk.
The example user now has three different but loosely related information hierarchies into which different types of data have been saved. The user can work with the inconsistent set of information spaces, but the user's productivity is adversely affected by the inconsistencies—navigation is slowed down and tasks such as making a backup of all material on a particular subject or relating to a particular project are complicated.
However, reorganizing the set of different data structures to increase the consistency between them would, in the absence of any assistance, be an onerous task.
The taxonomy manager program 60 is used to examine the user's e-mail folders within a Lotus Notes database, Browser bookmarks and directory structures. The taxonomy manager program compares names of data structures within the directory structures to identify similar names. Sufficiently close similarity implies that a positive match has been found, and then the data structures are used to generate a ‘reference taxonomy’, as described above. The GUI presents the user with a graphical representation of the reference taxonomy, that the user might find intuitive and be able to remember.
As described above, the reference taxonomy is not necessarily an aggregate of all features of the compared taxonomies, since each addition is subject to confirmation by the user—and in some embodiments exceptions can be specified in advance of generating a candidate reference taxonomy. In the present embodiment, the ‘reference taxonomy’ is a taxonomy which is generated to represent an overall data organisation and the application taxonomies are each mapped onto the reference taxonomy.
Since the reference taxonomy is typically derived from hierarchies generated by the user, it is likely that the reference taxonomy will provide a reasonable representation of the user's view of data relationships. Furthermore, the GUI provides an easily viewable graphical form of the reference taxonomy which is very similar to well known application-specific folder structures. The user can then use operations provided by the GUI which allow re-organisation of the reference taxonomy to suit their preferences, or the user can indicate acceptance of the candidate reference taxonomy.
In response to the acceptance, the taxonomy manager creates, renames and moves folders, bookmarks and directories within one or both of the compared taxonomies to fit the user's accepted reference taxonomy. Where folders, bookmarks or directories are no longer needed as a result of merging or modifying the taxonomies, the taxonomy manager prompts the user for a decision regarding whether to remove unnecessary structures. No data items stored within the data structures have been deleted—instead data has been reorganised into a set of independent but consistent data structures.
The result of the above-described method is that the user is provided with a harmonised organisation of their information. When an e-mail is subsequently received with a large attachment on “The Performance Characteristics of Product1” the user decides to save the e-mail in a ‘Products/Product1/Performance’ mail folder. The user then decides to detach the document, and the user can do so directly into the ‘C:\Products\Product1\Performance’ folder on their local drive. The user should benefit from the harmonisation when subsequently needing to remember where the document was saved, because the path to the document will be similar to the folder in which the original e-mail was saved.
Distributed Management of Information Storage
In addition to the local use of a taxonomy manager program as described above, a taxonomy generated as an XML file by one taxonomy manager program can be communicated to other taxonomy manager programs running on different computers within a network and then applied to programs running on the different systems. This is shown schematically in
The desire for consistency of information structures is equally valid when information is distributed across a number of different computers. For example, a set of bookmark folders on one machine can be synchronized with a similar set on a different machine when a bookmark folder is created or moved on one machine. By enabling other designated machines to be automatically updated, a user's laptop computer can be harmonized with their desktop system. A project manager can design an appropriate data organisation for use by all members of the team, and this in turn may help to associate different levels of security or different Access Control Lists with different data structures such as data directories of folders.
Messaging
The taxonomy manager 60 can be implemented as a client of a messaging system 140, and used to communicate with the other computers 10′ running taxonomy manager programs 60′ to provide distributed operation. To achieve this, the above-described taxonomy manager architecture is extended in two ways. Firstly, a messaging system adapter 120 is provided as a messaging client application which couples the taxonomy engine 110 to the messaging system 140. Secondly, a listener 145 is provided which receives messages from other taxonomy managers 60 across the network and inputs the messages into the receiving-end taxonomy manager 60′. The receiving taxonomy manager then effects changes in information structures on the receiving computer 10′.
The individual taxonomy managers can use point-to-point messaging or, for example at a departmental level within an organisation, can use the publish/subscribe methodology to send and receive updates to a commonly agreed taxonomy. In a publish/subscribe solution, the listener component 145 can be implemented within the adapter 120′.
For example, consider the case of a departmentally agreed hierarchical taxonomy of projects and components of those projects. Computers (users) register an interest in the synchronisation of additions or changes to any part of that component taxonomy by subscribing to the relevant topics. Any machine which makes a change to the taxonomy publishes that change on a matching topic, allowing all interested (subscribed) taxonomy manager instances to pick up the change. Access Control Lists (ACLs) of the publish/subscribe broker are used to create an authorisation environment and prevent unauthorised changes from being published.
Described above is an example use of publish/subscribe communications for propagation of taxonomy information. A further problem can be solved using the above-described architecture of a taxonomy engine, GUI and adapters. This is the problem of distributing message routing information within a network of messaging systems that use the publish/subscribe communication methodology. In this application of a taxonomy manager, the taxonomy which is to be visualized and possibly modified is the taxonomy of topic names within a publish/subscribe message broker.
Publish/subscribe messaging systems allow producers of messages to specify a “topic” to which a message relates and consumers of messages to specify sets of topics in which they have an interest. A message broker stores a set of subscriptions in the form of topic names identifying topics to which users wish to subscribe. The broker then matches interested subscribers to published messages, by matching the string of topic names of an incoming published message with the topic names in the stored subscription lists, and delivers the published messages to the relevant subscribers. The producers and consumers are typically decoupled—such that they do not know of each others' existence. Topics can be organised into tree-based hierarchies, which may allow whole sub-trees to be selected by the use of wildcards. For example, a topic taxonomy might contain a root node of “/news”, which in turn contains sub-topics of “sport” & “weather”. Within each of the sub-topics there could be other sub-topics, such as “hockey”, “football”, etc. A subscriber can express an interest in a specific topic such as “/news/sport/hockey”, or in a set of topics (an area) such as “/news/sport/*”, in which case they will receive all messages on sub-topics of sport.
Known examples of publish/subscribe messaging systems suitable for use with a taxonomy engine, GUI and messaging adapters as described above include IBM® Corporation's WebSphere® MQ Integrator and MQSeries® Integrator family of software products. (IBM, WebSphere and MQSeries are registered trademarks of International Business Machines Corporation.) Such publish/subscribe messaging software products implement message brokering capabilities that provide intelligent message routing and integration of applications—typically routing messages according to business rules and knowledge of different application programs' information requirements, using message topic information (as described above) which is typically contained in message headers. Although not essential to the present invention, such publish/subscribe messaging systems also typically provide the ability to transform message formats with reference to the format requirements of target applications or systems. It is also known to distribute the processing load of message brokering functions across multiple processes, machines and geographical locations using a multi-broker topology. In addition to performance improvements via load sharing, such a distributed approach can also reduce the impact of a particular server failure. Distributing publish/subscribe message broker capabilities across a network requires the communication between the brokers of updates to message routing information that result from changes of subscriptions or, in some cases, changes of publish/subscribe authorisations. A taxonomy manager can complement the existing publish/subscribe message routing capabilities of a message broker to provide an efficient mechanism for the distribution of such information between publish/subscribe messaging systems.
The taxonomy manager 60 finds out what is being published, for example by subscribing to the root topic (expensive in terms of processing overhead, since all publications would then be sent to the taxonomy manager) or by the system administrator receiving a topic catalogue from which selections can be made. The taxonomy engine is configured to subscribe to a portion of the topic space, where it matches portions of the user's local taxonomy. This matching step uses the same matching techniques and rules as described above to relate application taxonomies, and is also subject to user confirmation. There is no guarantee of correlation between a user's taxonomy and the topic taxonomy used within a publish/subscribe system. As a result, the taxonomy manager may not be able to use simple name similarity to generate “matches”. In this case, the taxonomy manager either requests a topic catalogue or monitors topic strings of actual publications as mentioned above. The taxonomy manager could use this information to learn part of the structure of the actual topic taxonomy and use this taxonomy in the same way that it would use knowledge of a user's folders or bookmarks to suggest ways in which their local information might be structured.
The topic space corresponding to the stream of publications can thus be used as another source of data structures. The topic space can be compared and related to the user's overall taxonomy in the same way as described above for an e-mail application's folder structure. Therefore, the topic space can be used to influence the overall taxonomy that the user sees.
The user can define relationships between a topic sub-tree and a tree in the overall taxonomy which can then be used to drive the creation of a corresponding folder on their machine. This ability to define new relationships has the advantage of facilitating harmonisation of data organisation (enabling the user to increase the correlation between the topic taxonomy and the user's own folders etc) even in cases where there would otherwise be insufficient correlation for the taxonomy manager to identify any similarity between the global topic space and the local taxonomy. Once created, a topic-taxonomy relationship would stay in place despite changes within the sub-tree of topics below that topic in the taxonomy.
Optionally, the taxonomy manager can register subscriptions on the user's behalf due to the existence of a node in the overall taxonomy. If the user so chooses, they can subscribe to the publish/subscribe broker on a topic that is related to that specific node.
A further option is for the taxonomy program to pro-actively read the message destinations (e.g. message queues) relating to subscriptions, and to receive saved messages into the appropriate folders. In this way, the folder structure is automatically populated with data received from the network.
Extensions Including Data Backup Extension
A generated reference taxonomy can also be used together with the taxonomy manager's GUI to enable user-initiation of operations on data structures of the application taxonomies. Examples include the data backup and data retrieval extensions described above. A particular embodiment implements user-selectable commands within the taxonomy manager's GUI. The GUI commands initiate existing functions of the application program corresponding to an application taxonomy and provide parameters indicating which data structures of the application taxonomy are to be operated on.
If one or more applications referenced by the taxonomy manager have the capability to backup their data then it can be useful to enable the user to ask the taxonomy manager to backup selected nodes (for example, by putting a tick in a checkbox next to a node in the GUI representation of the reference taxonomy). Then all applications which are capable of backing up data could back up the data in their corresponding nodes.
In this way the user is able to work with the applications through the taxonomy manager itself rather then using it merely to organise data. Since a user-selected reference taxonomy represents the organisation of data that corresponds most closely to the user's view of how data is related, a GUI representation of the reference taxonomy can provide a highly intuitive interface for the user to initiate operations on the stored data and/or data structures of a set of application programs. The taxonomy manager can similarly be extended to manage other features or behaviours (such as data compression or encryption) that apply to one or more applications.
Data Retrieval Extension
An additional application of the taxonomy manager and merged taxonomy described above is as an index for accessing data which is distributed across the set of managed applications. A single node of the reference taxonomy created using the taxonomy manager as described above may correspond to a folder in each of the individual application-specific taxonomies, and this is advantageous since the consistency of data organisation helps the user to find the set of relevant folders within the set of applications. However, data retrieval efficiency can be improved still further. A function is added to the taxonomy manager via a plug-in component interfacing between a managed program and the taxonomy engine. This enables a user who is working with data items within one of the managed programs to specify a node of the reference taxonomy for which folder contents are required. The taxonomy manager receives the request via the plug-in component from the requestor application, and sends a corresponding request to each managed application requesting them to output the contents of their folders corresponding to the specified node. An example of this is where an e-mail within an e-mail folder has a corresponding text document which has been detached to a related folder within the computer's local file system or elsewhere in the network. The user is enabled to select the e-mail and to trigger a request for retrieval of the associated document from the related file system folder.
A simple implementation of the data retrieval extension can be used to show which applications have data related to a particular node or subtree in the reference taxonomy. This provides the option to perform a targeted cross-application search for particular information. The search is launched from the taxonomy manager and scans the folders and other data structures that are represented by nodes of the reference taxonomy subtree of interest.
Summary
Thus, taxonomy manager programs can be used for visualizing and/or managing data organisation, for re-ordering data structures and generating new data structures within an overall data management structure to increase conformity between different types of program or different instances of a program distributed across a network. Taxonomy managers can be used to receive new information relating to data organisation or subscriptions, and to publish such information. Having created new relationships between data structures or subscription topics, as described above, a taxonomy manager can be used to monitor activity on a user's computer. When an item is saved in a particular folder, the taxonomy manager can publish that item to a message broker. A GUI of a generated reference taxonomy can be used as an interface for initiating operations on application program data structures.
Alternatives
The above description of one or more embodiments includes a step of parsing an application's taxonomy and extracting the organisational information and converting the organisational information into a form that can be processed by the taxonomy engine. A number of data organisations are then compared to identify similarities and differences, and a modified taxonomy is generated as a separate file. In alternative embodiments, the extracted organisational information may be stored in a format such as XML or in a format other than XML. In another alternative embodiment, the reference taxonomy may comprise meta information about relationships between nodes in different taxonomies, the meta information being stored together with the application taxonomies.
In the above description, a brief mention was made of an alternative implementation in which the user has an option of configuring the system for automated updating of the taxonomies of all of the applications that use the reference taxonomy. This is implementable as an automated triggering of the taxonomy manager in response to a modification of any one of the taxonomies. Each of the managed applications includes a function for calling the taxonomy engine, via the corresponding adapter, in response to any update to their taxonomies. The call may be a simple request to initiate the ‘compare taxonomies’ step described earlier, or the call may include information describing the new changes to the calling application's taxonomy. For example, if a new bookmark folder is created in a Web Browser, a corresponding folder may be added automatically to an e-mail application's taxonomy and/or an operating system file system taxonomy. Where a user has accepted an apparent match between application folders and a corresponding node within a candidate reference taxonomy, the creation of a subfolder within the reference taxonomy to maintain consistency with a new application subfolder is likely to be useful to the user. The user can configure the system to control whether and when such an update is propagated to the taxonomies of the other applications. Thus a node in an application taxonomy has an attribute that dictates whether a change to it, by its own application, is automatically reflected in the reference taxonomy at the time of the next scheduled synchronisation operation. A node in the reference taxonomy has an attribute that specifies whether a change is automatically propagated to the applications. These attributes may be specific to add, delete or move operations.
Number | Date | Country | Kind |
---|---|---|---|
0315190.9 | Jun 2003 | GB | national |
Number | Name | Date | Kind |
---|---|---|---|
5794039 | Guck | Aug 1998 | A |
6442545 | Feldman et al. | Aug 2002 | B1 |
6625619 | McClendon et al. | Sep 2003 | B1 |
6711585 | Copperman et al. | Mar 2004 | B1 |
6980984 | Huffman et al. | Dec 2005 | B1 |
7047236 | Conroy et al. | May 2006 | B2 |
7177859 | Pather et al. | Feb 2007 | B2 |
7181468 | Spring et al. | Feb 2007 | B2 |
7200592 | Goodwin et al. | Apr 2007 | B2 |
20030050915 | Allemang et al. | Mar 2003 | A1 |
20040024739 | Copperman et al. | Feb 2004 | A1 |
20040181471 | Rogers | Sep 2004 | A1 |
20060224683 | Briscoe et al. | Oct 2006 | A1 |
Number | Date | Country | |
---|---|---|---|
20050010458 A1 | Jan 2005 | US |