WO2004088544A1 - Query processor for classifiable items - Google Patents

Abstract

Computer-searchable catalogues hold information as predetermined hierarchical classified descriptors for inherently classifiabe items. Descriptors are selected from an hierarchical search board menu in order to build a search strategy. Currently in-use terms and their interrelationships are always visible on the search board, to minimise difficulty in searching. The searcher may easily modify and repeat a search. An unclassified keyword search is optionally available. Current searches may be stored in cookies and/or within the server. Applications include Web e-commerce and providing information about items like lists of species of animals or plants from within a solitary computer.

Description

TITLE QUERY PROCESSOR FOR CLASSIFIABLE ITEMS.

FIELD

This invention relates to software for information retrieval such as by means of access using the Internet or an intranet; to browsing optimisation, to a form of "temporary bookmark management", enhancement of navigation and to web site content organisation and management, thereby facilitating access by a user to managed data held in an organised database.

BACKGROUND

This specification addresses several types of "enhancement of navigation" by a person within a computer database, largely in order to facilitate the end purpose of the search be it e- commerce, information retrieval, access to a file, or similar tasks. The act of looking for an item in a database, or a particular computer file within a directory, is a frequently carried out task for a computer user, yet the quality of the human interface, providing assistance to the user, is usually poor. The inventors are involved with an e-commerce venture. Development of the present invention was primarily stimulated by the existence of frustrating and hard-to- use interfaces within presently existing search engines, web sites, programmes, and the like.

E-commerce catalogues are not as easy to use as their corresponding printed paper catalogues, though each has specific advantages. Often a printed catalogue (typically 1500 pages in size in our business) must be available to supplement an electronic form. Most catalogues grow by using more hyperlink search levels, meaning that the searcher will encounter greater difficulties with the search. Many searchers revert to a printed catalogue (if still available) as a store of information.

Collections of data which are hard to search are clearly sub-optimal for meeting their intended purpose. If a web site was constructed to offer items for sale as an e-commerce venture, one consequence of reduced searching efficiency is a lowered rate of sales. The inventors have noted that web sites purporting to replicate a printed catalogue of items for sale almost invariably exhibit this problem. Any user would want to quickly and easily either (a) locate and assimilate the information of interest, or (b) be satisfied, without undue effort, that it is absent. Several solutions to the "enhancement of navigation" requirement may be combined in order to raise the frequency of satisfactory search procedures, such as:

1. Helping the user; enhancement of navigation by holding the latest tested position. a) Nearly all list presenting routines lack means to hold the "last entry viewed" in a kind of invisible/virtual bookmark, so if an initial selection is wrong the user will usually have to start over again from the beginning. The majority of existing web sites that provide for user searching offer search engines in which a fresh, empty entry screen appears every time the user steps back from a bit, and all the entered data (which may include complex, non-intuitive strings entered from a keyboard) is lost. Widely used programmes such as Adobe's "Photoshop" (R) version 7 do not retain any kind of automatic bookmark to the last-opened image file under the File-Open command (although the "recently opened" list is an approximation). This type of problem becomes more acute if the file names are not descriptive - they may be digital camera sequential names only, held in time order. b) In the event of not hitting the desired block of information by means of a search term, if the search strategy or search terms have disappeared, the user has no indication of what was wrong with the search. To continue, the user must re-enter some or all the search parameters before executing another search. In some cases the search term is a relatively complex string; hard to enter without operator error, and as a result the level of difficulty in using the database rises. c) Some sites, such as the USPTO or the ESPACE patents search engines, do retain the parameters in a user-modifiable fonn after a search, but that search is not hierarchical, and is unstructured because the user can enter any string. (In this case that may well be the most appropriate design for a search engine though many of this type maintain immense indexes including everything but "stop" words too insignificant to be descriptive). Helping the user; leading the user to a "permissible scope or range" during a search. a) The software of many searchable sites, including e-commerce sites offers no guidance to the naive user as to what can be asked for in a free-form search window. Well- developed librarian-type skills may be required to search databases of that type with confidence, although customers engaged in on-line purchasing on a variety of sites are unlikely to be experts and it is a weakness of the site to expect such skills. The user has no indication of what was wrong with his enquiry because the last search strategy has disappeared. It may be a matter of using American or English spelling, it may be a matter of including a space in a parts code. Perhaps the thing is not available at all. b) A free-form search window has the problem that the user may not know what terms are available for matching. If "BC109" returns a some "no hits" answer, how does the user know ,to try "BC109AX" instead? Some search engines display the number of hits involved in a Boolean search, helping the user to see which terms are too restrictive or too general. c) Many computer programs and many sites on the World Wide Web offer a form of table-based search engine with hierarchical search screens. Hierarchical hyperlinks are commonly employed to guide the user through a number of separate displayable screens in order to complete a search input. Alternatively,

3. A search that uses a previously constructed classification which forces the user to stay within a limited range of available parameters tends to lead the searcher towards a result as he or she selects from the available list, moving down a hierarchical stracture towards a more closely described set of search descriptors and a more limited number of genuine choices. A search that is not based on prior classification of this type and which lacks the pre-ordained search terms leaves the searcher unaware of the actual range of available items until the search is run (at which time (in general) a computer programme will attempt to match the parameters possessed by each existing item with the set of parameters specified by the user and report those items that matched). In many cases a user unfamiliar with a specific e-commerce web site cannot discern what is on offer, or why he or she failed to find any item of a desired type.

PRIOR ART No relevant patent publications are known within the prior art. One currently (April 2, 2004) active web site which shows, in part only, similar functions is that of the City Lit in London <www.citylit.ac.uk> which has a subject then day selection code for students to use. Currently active patent search web sites such as <http://ep.espacement.com> do handle at least partially classified items in the database (using IPC and/or ECLA classes) but there is no apparent linkage between search terms while they are being entered into the boxes and no requirement to traverse levels of an hierarchical indexing or classification system in order to create a search.

DEFINITIONS

Hierarchical Classified Data Set CDS: the type mainly considered during the development of this invention, is one in which a hierarchy of super-groups, groups, sub-groups etc each hold Descriptors that reasonably describe a single item or a group of items. The Descriptors have been defined and created by a cataloguer charged with organising the item data. Preferred Descriptors are industry standard words that would impart meaning to a user versed in the industry being catalogued. In Fig 4, the CDS comprises the variable contents of boxes 405 - 100 Ordered Classified Data, the type mainly considered during the development of this invention, presents a classified group to the user in a predetermined structure.

A Descriptor is a generic name for a field in a data table for the time being displayed in the search board. In Fig 4, boxes 405 to 408 plus box 409 comprise the Descriptors.

List driven means the searcher is bound to a restricted data set of valid items.

105 The terms item and information are used to describe an entity that (a) has been considered worth placing into a catalogue or (b) is a discrete thing, product or service subject to inspection and acquisition by a user.

A data table is a means for holding related information.

A classified catalogue holds ordered information pertaining to sets of items such as stock in a 110 shop or warehouse, books or music, stock catalogues, or museum collections or species of animals or plants in the wild, or the like.

In the context of this invention the following data tables will be discussed:

1. The classification data table: a) For a commercial example, an equivalent to a printed catalogue of goods, organised by 115 catalogue name, catalogue index, product type and product families would typify the classification system. b) For a scientific example, the affinity of organisms to be subject to a systematic taxonomy (class, order, family, species, and sub-species) provides a valid classification basis.

120 2. The vendor data table:

The vendor data table would hold vendor name and other details

3. The item data table:

The item data table would include item details and (internal) links to the classification data set and may usefully include fields providing additional data filtering options.

125 For a commercial example the data field 'vendor' provides one filtering option.

For a scientific example the data field 'country' provides one filtering option. 4. The user data table:

The user data would hold the login information, password, search data and other details.

5. The Selection Data Set (SDS) data table: 130 This table would hold historical SDS data.

A Field is one part of a record in data table.

A Record is a complete field data set held in the data table

A Data Base is a collection of data tables.

Hyperlink: A recognised method of jumping to new information on the same web page or 135 another web page. Widely used on and outside the World Wide Web and is implemented in for example HTML (Hyper Text Markup Language).

Search Board: A screen menu panel displaying the Selection Data Set. (See 402 - fig 4).

Selection Data Set: The user choice of Descriptors for the time being displayed on pull-down menus and freeform entry boxes and tick boxes within the Search Board

140 In the non-limiting example used to illustrate the principles of the invention, the Selection Data Set has the following elements: a) Classification Data Set: In this example, a hierarchical, ordered list of classified items that appear in pull-down menu boxes labelled Catalogue, Index, Product Type and Product Family. (This is otherwise known as a set of tables, to be used in a top-

145 down sequence). b) Vendor: A list driven group of entities offering item(s) for trade accessed by a pull down menu box. (Fig 4 - 409). c) Keyword is a free form Descriptor, which the searcher reasonably expects to find used within the item data set. (Fig 4 - 410)

150 d) Two Tick boxes alias "radio buttons" are used to limit the scope of the Keyword search. (Fig 4- 416 indicates one tick box).

Hence, in this example, five Descriptors form the Selection Data Set.

(This example offers a keyword search that is restricted to a single word. Overcoming this limitation is a standard programming exercise.)

155 A search engine is a device for locating items within a database table, starting with the selection data set entered by the user.

A SDS cookie is a variable data string containing the user SDS for the time being. A browser may be set to accept (and record), or to reject a cookie.

An interface for a search engine is a software bundle capable of displaying choices to a user by 160 means of a display device, accepting input from the user, handling the input according to rules, and passing the input to the search engine itself, optionally after some modification.

A terminal is a computer hardware device typically including display means and user control inputs such as a keyboard and/or a pointing device and controlling software; in most cases comprising a common digital computer running browser software.

165 A network: Includes local area networks or "intranets" and wide-area networks. The Worldwide Web or Internet is named herein as an existing and preferred wide area network; however equivalents of the Internet developed in the future as replacements or as alternatives should be regarded as equally preferred options.

OBJECT

170 It is an object of this invention to provide an improved searching/finding interface for catalogued items and/or information held in electronic form, or at least to provide the public with a useful choice.

STATEMENT OF INVENTION

In a first broad aspect, the invention provides software for use within a first digital computer 175 for allowing a person to query or search for a desired selection of items from within a catalogued collection of items represented as entries within at least one database, wherein the software includes

(a) means to create a Classification Data Set or CDS (as herein defined) comprising a hierarchically arranged group including at least one member, each member holding one or more

180 predetermined search terms,

(b) entry of hierarchically arranged data pertaining to particular items; the data being mainly taken from the CDS,

(c) means for presenting the CDS to the person so that a search term may be selected by the person from within each member of the group in relation to any previously selected terms,

185 and

(d) means to show and retain a display of the cunent hierarchically arranged set of selected search terms (comprising part of a "selected data set" or SDS as herein defined) so that they remain available for use at least during the person's search, and

(d) means for displaying at least one resulting set of zero or more hits resulting from use of 190 the SDS together with the displayed SDS, all so that a search in the catalogued collection may be carried out more effectively and more conveniently.

Preferably the invention also provides means for minimising the number of different pages to be displayed by avoidance of a construction using a plurality of links (hyperlinks) between different pages and instead retaining a consistent display comprising a search board including 195 the cmrent SDS beside the or each display of hits.

Preferably the person is also provided with an option to invoke one or more non-predetermined search terms outside the Descriptors by the inclusion of means that provide for a search based on unique strings such as keywords.

Preferably the invention also provides means to retain the current SDS during an interval after 200 an earlier search and before a later search, so that the SDS remains visible, and remains functional, so that the SDS from the earlier search may be used during the later search.

Preferably the invention also provides means to retain the SDS during an interval of indefinite length between an earlier search session and a later search session, so that the SDS remains visible and remains functional, so that the SDS from the earlier search session may be used 205 during the later search session.

A preferred means for retaining the SDS comprises a storable and recallable entity holding data and an example entity is a SDS cookie (as herein defined).

Preferably the catalogued collection of items represented as entries within at least one database are held within a second digital computer or host/server used to store and dispense the infor- 210 mation to at least one first digital computer, each used as an interrogating terminal, and the software used in the host includes means for co-operating with the software (such as an existing browser) used in the terminal.

A preferred example of co-operation is that the software used in the host is capable, on demand, of communicating executable software suitable for use within the terminal to provide 215 means for allowing a person to query or search within a catalogued collection of items represented as entries within at least one database, in order to retrieve a selection of items from the group.

Optionally the software for use within the terminal and the software for use within the host are provided as a combination package capable of ranning within a single computer. 220 Optionally the search terms are stored remotely from the terminal and may be recalled for use - at a later date.

Preferably a plurality of terminals are connectable by means of a data network to the host thereby providing access to the group of items for a plurality of persons

Preferably (in the case of person who has become logged in to the software) the SDS cookie is 225 held within the host; otherwise the SDS cookie is held within the terminal.

Preferably the display of found items includes further descriptive information.

Preferably the further descriptive information may be accessed by means of at least one icon alongside each found item, selection of which causes a corresponding function to be enabled.

Preferred corresponding functions for the case of a warehouse catalogue or other e-commerce 230 application include the functions of: (a) providing a purchase price, (b) seeking a quotation, or (c) providing further information, perhaps in the form of a separate but linked file.

In a second broad aspect, the invention provides software for use within at least one digital computer; the software providing means for allowing a catalogued collection of items represented as entries within at least one database to be prepared for later retrieval of 235 information by a person, wherein the software includes means for:- a) maintaining at least one set of controlled Descriptors (the Classification Data Set as herein defined) belonging to a hierarchical level for the catalogue; the at least one set being created by classifying a plurality of items intended for inclusion in the catalogue and adding zero or more Descriptors to each hierarchical level of the controlled set of Descriptors for

240 each item; and b) maintaining the catalogue in the form of a plurality of records holding at least one Descriptor taken from each controlled set of Descriptors, at each hierarchical level, for each of the plurality of items for inclusion in the catalogue, so that a hierarchically arranged group, the group having more than one member, of predetermined, available search terms

245 is generated and may used in searching.

In a third broad aspect the invention provides a first method for classifying or cataloguing at least one item in order to render the item findable by a searcher from within a catalogue-like database holding a plurality of catalogued items; the method including: a) creation of a classification system having more than one hierarchical level for describing the 250 items to be included in the catalogue-like database wherein each level is provided with an associated list of Descriptors reflecting the options available at that level, b) (for each item to be included in the catalogue-like database) addition of a Descriptor, if not already present, to the list of Descriptors within each hierarchical level, and c) (for each item to be included in the catalogue-like database), making an internal link 255 between the item and a corresponding Descriptor at each level, so that subsequent selection of a partial or complete hierarchical set of Descriptors will retrieve at least one corresponding item.

In a fourth broad aspect, the invention provides a further method for retrieving information concerning the said catalogued item from within the database by a searcher operating a 260 terminal, wherein the further method includes the steps of: a) placing a selected search term in zero or more of a hierarchical range of selection boxes displayed at the terminal, and b) optionally entering a search modifier selected from a list of permitted modifiers or entering a free-form modifier, and

265 c) causing execution of a search within the database employing the entered search data set, and d) inspecting a display of a set of item-related information for each found item, if any, for which a coincidence between the selection data set and the item Descriptor occurred, and e) the invention further providing retention and display of all search terms used so that

270 f) one or more of the selection data set can be edited for re-use by the searcher so that the search can be repeated.

Preferably this invention provides a search board interface for a catalogue held as an electronic data base wherein a user may use a search board to search within the catalogue for one or more items each represented by a record and having more than one field for each record, 275 wherein a) more than one selection box each conesponding to one of a hierarchy of Descriptors is shown on the display device, and b) the user selection of a Descriptor for loading into a hierarchical selection box is limited to valid Descriptors in that field, and

280 c) each Descriptor is presented again within the corresponding selection box after the search is run, and d) each Descriptor is available for modification in hierarchical order prior to ranning another search, and e) all user searches, regardless of the number of levels of hierarchy, take place within a single 285 frame set, so that the interface allows the user's selection of any one or more Descriptors to be inspected and optionally edited for re-use after each run of a search procedure, and so that the interface is more easily and more accurately selected for repeated use.

Preferably the invention provides a procedure for classifying or cataloguing and then finding at least one item, the data being stored within a computing means as a database concerning a 290 plurality of items; the procedure including a) a classification method for items and information for use during the creation of the data base, the method including at least one level of hierarchy and the creation of one or more Descriptors that indicate a hierarchical structure; b) the input of one or more items including their classification from the range of permitted 295 Descriptors within that hierarchical classification data set in order to specify at least one or more attributes possessed by an item to be catalogued, and c) a finding method for retrieving information concerning the said catalogued item from within the database by a searcher operating a terminal, wherein the finding method includes the steps of

300 i) entering none or the highest one or as many of an hierarchical range of selection boxes displayed at the terminal, and ii) optionally entering a search modifier selected from a list of permitted modifiers or entering a freeform modifier, and iii) causing execution of a search within the database employing the entered search data set, 305 and iv) inspecting a display of a set of item-related information for each found item, if any, for which a coincidence between the selection data set and the item Descriptor occurred, and v) the invention further providing retention and display of all search Descriptors used so that 310 vi) one or more of the Selection Data Set can be edited for re-use by the searcher so that the search can be repeated.

In an alternative aspect this invention provides a search board interface for a catalogue held as an electronic data base wherein a user may use a search board to search within the catalogue in order to locate one or more items each represented by a record and having more than one 315 field for each record, the interface providing that: a) more than one selection box each corresponding to one of a hierarchy of Descriptors is shown on the display device, b) the user's selection of a Descriptor for loading into a selection box is limited to valid Descriptors in that field,

320 c) each Descriptor is presented again within the corresponding selection box after the search is run, d) each Descriptor is available for modification in hierarchical order prior to ranning another search, and e) all user searches, regardless of the number of levels of hierarchy, take place within a single 325 frame set, so that the interface allows the user's selection of any one or more Descriptors to be inspected and optionally edited for re-use after each ran of a search procedure, and so that the interface is more easily and more accurately selected for repeated use.

In a fifth broad aspect, the invention provides a digital computer wherein the digital computer is ranning software in a stand-alone form, as previously described in this section.

330 In a related aspect the invention provides a host computer connected through a network to at least one terminal, wherein the host computer and the terminal are ranning software in a form capable of communicating with each other across a network, as previously described in this section.

PREFERRED EMBODIMENT

335 The description of the invention to be provided herein is given purely by way of example and is not to be taken in any way as limiting the scope or extent of the invention.

DRAWINGS

Fig 1: is a flow chart showing a procedure for generating and mamtaining an hierarchical array of lists of standard terms for items included in a catalogue.

340 Fig 2: is a flow chart showing a procedure for generating and mamtaining a database of items, holding the standard terms selected for each item, and any extra information required by the catalogue and underlying purpose.

Fig 3: is a flow chart showing a sequence of interactions between a host (server holding a catalogued database) and a searcher at a requesting computer when searching a 345 database of items according to the invention, and taking useful steps at the completion of the search. Fig 4: is a typical screen layout as seen by a searcher at a requesting computer during a search.

Fig 5: reproduces a captured typical screen layout as seen by a searcher at a requesting 350 computer during a search. (Colour would be used in practical screen layouts).

Fig 6: reproduces a typical screen layout as seen by a classifier when loading information relation to searchable items.

Fig 7: reproduces a typical product-description screen layout as seen by a classifier when loading specific information relating to a searchable item.

355 In brief, this invention first creates classified catalogues consisting of any information pertaining to things that are inherently classifiable, such as stock in a shop or warehouse, books or music, museum collections, lists of animals or plants (especially where the hierarchy of species names as developed by Linnaeus is used) or the like. The classified catalogue is then made available for searching by means of Classified Data Sets within the structure of a

360 search board menu, all of which is displayed by an appropriate programme either ranning within a single computer or partly ranning within a user computer and partly within a server computer (which usually transmits the user software to the user computer (terminal) on each occasion).

The user runs, modifies, and re-runs a multi-level search while retaining a full view of the 365 logical structure of all search elements.

According to the invention, the set of attributes that relate to a particular item to be included within the electronic catalogue or database are (a) described in terms of a hierarchy ranging from broad to narrow Descriptors, and (b) are selected from a deliberately limited pre-deter- rnined range. This CDS may be supplemented by for example key-word searches (Fig 4: 410) 370 or search on provenance, manufacturer/vendor 409, or the like.

EXAMPLE 1

This Example relates to use of the invention within an e-commerce web site, in which an electronic equipment vendor offers items from a catalogue that exists in electronic form. At this time, most browsers can support routines written in Java, HTML, XML, and the like. 375 Although the cunently preferred language for use in a requesting computer 302 (fig 3) when browsing is HTML (hypertext markup language) the invention may be written using any applicable software in one or more languages as are commonly used in such situations. The visual appearances of the examples are given by way of example and are also non-limiting. CLASSIFIED CATALOGUE CREATION

380 In Fig 1, 100 is the starting point of a process for creating and mamtaining lists of standard terms; the Classification Data Set. This process can be regarded as a first thread to be run from time to time, even while item classification (Fig 2) is continuing. At the start 100, an item to be included in a catalogue is taken 102, and for each category (such as hierarchical level or vendor) a classifier reviews the cunent list of standard terms and ensures that one of

385 the terms provided in that list can reasonably be applied to the article. If not, a new standard term can be created (104) although it is generally preferable to keep the number of standard terms in any particular list as low as possible. All the terms used in a list will be offered to a searcher at a later date for use in a search strategy. At the end of a check, a decision 105 as to whether to re-start with another item or finish (106) is made. Fig 6 shows a computer screen

390 in use during item classification.

(As a development, a count can be kept of the number of times that any particular standard term has been used within a catalogue as created in Fig 2). A low or zero count may indicate that the term is inappropriate while a high count may suggest that the term is too broad and cannot distinguish items.

395 In Fig 2, cataloguing of items is described. Assuming for the purpose of illustration that a conventional database having one record including one or more fields to represent each item is used, the process starts (200) by creating a new record for a new item (201). Starting with the broadest category a data entry person selects a most appropriate standard term from a list displayed to the person, and loads that term (202) into the "broadest category" field. That

400 loaded term will lead to a more specific list of terms, each of which depends on that broader term, for the next lower level. The process continues through further categories 203 - or levels in the hierarchy until the narrowest category is done. At that time, the record in the database will have a single term (usually) for each level although for the sake of more perfect recall, more than one search term option may be stored. At box 205, extra information, such as

405 links to text files, data sheets, publications, and the like can be included. The item may have a price to be included at this point. The provenance or the manufacturer/vendor can be included. An illustration can be included. Other fields may also be included but not made visible to the public, as is common in the art. The flow chart of Fig 2 also terminates with a decision (206) as to whether to re-start or finish (207). It will be appreciated that although

410 prior-art catalogues of this type have been exploited in commerce, it is the searching environment (see below) that distinguishes this invention. Fig 7 shows an in-house classifier's screen as used when data relating to products are being added or edited. Here, 701 indicates an area set aside for text Descriptor entry, 702 indicates an area for illustration loading, and 703 indicates an area for entering pricing details. Along the top, 704 indicates an open one of 415 several drop-down menus for use in picking one Descriptor from the choices offered at each level thereby classifying the item. Because the same choices are offered to searchers at a later date, there should be no uncertainty in relation to finding what is present.

SETTING UP A SEARCH IN CLASSIFIED CATALOGUES

Fig 3 shows in a top-down format a typical dialogue 300 as may occur from start (303) to 420 finish (320) when a person makes use of the searchable database. For e-commerce applications there is usually a remote server, hosting a database 301 within a server computer forming part of an Internet service provider or other host. There is a requesting terminal or computer 302, operated by a user who wishes to make use of the catalogue or catalogues on offer. Use of the World Wide Web is assumed, although any of the communications media such as 425 intranets known to those skilled in computer communications could be used. 301 and 302 may on occasion comprise logical, not physical divisions of the same computer but for the example we shall assume a separate server.

The first step is to ascertain the URL (uniform resource locator) of the server 301 in order to establish a connection 305 to the web site of the information provider. As is well-known in 430 the art, browser-type software will accept graphics and page layout code 306, including search board layout from the server 301. Fig 4 shows such a page layout. Although the entire searching process could be done with text alone, people have become highly accustomed to a graphical user interface involving frames, search boxes, display areas, and the like with user control through a pointing device such as a computer mouse, supported by a keyboard.

435 SDS MANAGEMENT

The software places particular importance on holding the user SDS for display and re-use over a period of time. The authors consider that managing the SDS contributes to the "look and feel" of the software. Users fall into one of four usage modes. More serious users tend to often return to the e-commerce site for further quotations and/or purchases of a similar type

440 of product, so automatic recovery of the SDS is a time and effort saver. The SDS preserved from the most recent search enables the user to carry on from his or her last visit. A casual user would be less likely to go through the logging in procedure, which is at present treated as optional. The four different usage modes relate to past and current SDS presentations, based on whether the browser is set to record ("enable") or reject cookies, and whether the user

445 takes the active step of logging in to the web site as presented by the server, or not. Whether a cookie 307 is present or not depends on whether the computer has ever been searched with previously in this catalogue, and on whether the browser has been set up to record cookies from web sites.

1. If cookies are not accepted, and the user does not log in, then each time the user clicks the 450 Submit button (411 - Fig 4) and sends the SDS (comprising the search parameters) from the user's terminal to the server, the software will return the SDS from the server to the terminal together with the search results, thereby refreshing the search board. While the user remains within the catalogue search, the search board always shows the most current SDS. Upon leaving the catalogue search, such as when the user visits a web page outside 455 the Catalogue Search, the SDS is lost owing to the absence of a recording medium. Lacking recall of the SDS from an old session, the first use of the search board shows an empty SDS.

Each user who enters Catalogue Search and makes a selection on the Search Board and clicks the Submit button causes the SDS to be transmitted to the server. The server 460 activates a search engine to select the data requested within the SDS and returns the SDS and data requested for display to the user's computer together with a SDS cookie. The user's browser determines from preference settings or the like whether it will record or reject the cookie.

If the user has completed user login and is logged in, the SDS cookie is also recorded in 465 the user data area within the server. From this sequence, various opportunities arise.

2. If cookies are accepted, and the user does not log in, the SDS will be recorded during a search within a SDS cookie held on the user's computer. The SDS cookie is read and updated for each new catalogue search. The user may leave Catalogue Search, visit other web pages and on return will find the SDS appears, to automatically re-establish the last

470 search from the SDS cookie. If the SDS cookie is deleted before the next search, (as may be caused by a time-dependent flush of the cache in the user's computer) the SDS information is lost and the first use of the search board is on an empty SDS.

3. If cookies are recorded, and the user logs in as a known user, the user name and details will exist within the database in the server. Upon submitting a catalogue search, the SDS

475 is added to those user details, over-writing any previous SDS stored there.

4. The user may log in at any computer that accepts cookies and the server, upon accepting the login attempt, will use the SDS data recorded in the database to retrieve information and to update the SDS cookie. Hence, on visiting the catalogue the last user search is automatically re-established. Also, the user has access to commercial features of the site

480 including a message area. If the "keep me logged in" option was ticked, information in the cookie is used to automatically log the user in again. This restores the SDS for the last search but has the possible disadvantage that any other person using that computer is assumed to be the recognised user.

Use of a "keep me logged in" option is likely if the same person consistently uses the same 485 computer but is less advisable if other people use the same computer and the same web site because they would be logged in as the original person. This may lead to a security breach.

If a user evolves a successful SDS he or she may choose to store it under a user-selected identifier for later use, because otherwise the invention will allow only the most recent SDS available to be re-used.

490 SEARCH PROCESS

A Classification Data Set (CDS) may be presented using a selected one of multiple techniques. The selected technique may depend on the size and complexity of the CDS. One technique involves sending all elements of the CDS within a single (HTML) file which loads all selection options in a single upload. A second technique involves uploading the highest

495 one hierarchy of the CDS, and then uploading the next hierarchy subset as determined by the selection within the senior hierarchical set. The second technique has the advantage of being capable of handling a very large CDS, and the disadvantage of there being a momentary pause while the next subset of the CDS is uploaded before a subsequent selection can be completed. Both techniques have been implemented, with the preferred example (Example 1) using the

500 first alternative.

Assuming a fresh catalogue search is begun, the first step is for the requesting computer to nominate a specific database if more than one database is available from that server. Once that is done, the Classification Data Set including all the search terms (309) for all categories within that catalogue is passed from the server to the requesting computer in one file. It is

505 practicable and workable even if several working catalogues are in use together; each of which currently requires that up to about 30K of HTML (ASCII) characters are sent. That does not take long. The available search terms for each hierarchy are displayed on the search board portion of the page layout (Fig 4 at 405) for selection by clicking with a mouse (or any similar type of selection device). Software in the requesting computer selects the appropriate search

510 terms from the downloaded search terms, which are tagged with Descriptors, according to what has already been selected at the next higher level (if any). The hierarchy may comprise two to ten or even more levels though typically three to five hierarchical levels is sufficient for any kind of catalogue in order to refine a search to the relevant items only, so that a reasonably "pure" set of hits is returned. More than five or six may test the cunent speeds of

515 dial-up modems operating at say 56 Kbits/s to the extent that searching becomes impractical, by raising the numbers of search terms involved, although data compression will reduce the scale of the problem. For the specific application of Example 1, we believe that four hierarchical layers are almost always sufficient.

Preferably the system allows any category to be made maximally broad by selecting a "select 520 all terms" response, which will also disable any existing selection of particular terms within lower hierarchies. In Fig 4, the "+ Select All +" option, which allows all the terms currently available in the subset to be included in the SDS is being shown in boxes 408 and 409. Use of hierarchies helps impose order on the items that the searcher is preparing to search through in order to find a relevant item. In practice, a searcher who is new to the site is quickly and 525 easily guided to appropriate items in existence by means of the hierarchical classification.

After the search terms are established, the "Submit" button is activated. The current SDS (314) is returned to the server (301). A database search is made up from the SDS and is executed within the server. The current set of search terms (the SDS) is refreshed from the version held in the server to the searching computer so that the SDS is maintained on the

530 search board 402, remaining on display both during and after search execution and ready for optional alteration by the user. The inventors consider display of the SDS to be a significant advantage of their invention. Many prior-art search engines clear (or fail to reload) the entire set of terms when one term is modified or when a search is executed, and the user must re- enter them all. Under current browser and web-page related software constraints, the SDS on

535 display has to be deliberately refreshed from the server automatically and transparently by software after each pressing of the "Submit" button (411) or after each receipt of material to be displayed. In addition the SDS cookie is passed from the server to the searching computer.

A list of hits (if any) 315 is received from the server (301) and displayed at 413 for inspection. If the searcher notes an interesting hit 414, it may be explored and further details requested

540 (318) as are appropriate for the kind of item and the type of service offered. Further details may include an image, a data sheet, an analysis, a review, a price, details of provenance, and the like. A purchase may be made (319). If the list lacks interesting hits, one or more of the search tenns in the SDS that are still on show (405-410), may be modified as previously described and the search re-executed. The SDS is conditionally saved for re-use at the end of

545 a search as described in Table 1 (see above).

Additional search terms can be used. The user may search by vendor (manufacturer) 409 or an analogous type of Descriptor in other catalogues, such as provenance, age, common name, tribe, or materials for museum specimens. Further, the user may use a key-word search 410 as for many prior-art search engines, either alone or implicitly in a Boolean "AND" combi- 550 nation with the SDS. In the components e-commerce example described here, the user is offered two options;"key-words through all fields" (416) or "key-words just within order codes".

Fig 4 shows a preferred screen layout as generated by the preferred software upon a display device 400 during a search. The GUI is conveniently written in a language that allows frames. Within Fig 4, the bar 401 is used to select "Catalogue search" from several pages available on

555 the web site. The Search Board at left (402) is a frame containing several labelled entry boxes arranged in a top-down hierarchy. The first box 405 is entitled "Catalogue" by analogy with the idea of physically pulling one or more printed catalogues down from a bookshelf in order to search through them. Box 406 is called "Index" by analogy with the act of opening the index pages in the selected printed catalogue. Box 407 is called "Product Type" by analogy

560 with the idea of selecting a topic from the catalogue index. "Product Family" (box 208) is used to narrow the choice from within the index. A searcher may have no idea of the existing Product Family search terms until such time as she opens the box (405) by clicking on the triangle marker (405A) at the right, to expose the list. In this case it will be noted that the string currently inside the box, "-(-Select A11+" is a default or "maximising" entry that has the

565 effect of instracting the search engine to "ignore entries in this field" and report all the items present in the next higher search box. Otherwise there is an implied Boolean "AND" in operation between selection boxes holding data. The same default string "+Select A11+" is here shown inside the "Vendor" box 409 in order to permit display of all available vendors. One advantage of this invention over a set of paper catalogues is the capability for fusing

570 many paper catalogues (each representing a separate vendor's range) by displaying all similar products together, while at any time the searcher can extract the contents of a single catalogue by making a choice under "Vendor" . The hierarchical Descriptors (ternis) comprising the SDS shown here have been selected from a range limited during catalogue creation to just those terms for which a findable item exists, as previously described with reference to Figs 1 and 2.

575 In order to speed access using short cuts, the alternate Keyword Search box 410 allows entry of key word or product code strings for direct access to expected items although there need not be a mechanism here to impose a predetermined limit on the choice. "Search through" radio buttons (416) direct the keyword search through either all fields/tables or through order codes. Finally the "Submit" button 411 is activated in order to execute the search and at this

580 point the constructed strategy is sent to the host computer, typically as a SDS of selected terms. Although the searcher may press the "Submit" button 411 at any time before completing all possible fields, the result list from too broad a search may be unmanageably large. Note that, as shown here, the search terms in use remain visible to the user during construction of a search, while he/she is inspecting the resulting hits, and the search terms

585 may be edited before doing another search (see below). An entire database relating to all items is likely to be too difficult to search without an imposed organisation, if a searcher knows little of the usual organisation of the items covered by the database.

In this Example the user has searched for monochrome scanners within an existing catalogue called CAD (computer-aided design) Peripherals. At the right side the result list 412 displays

590 three hits at 414 (where the words "summary of specifications" is shown in the diagram, the real screen shows details). Items may be illustrated at this point. A price and a shopping basket icon are included for each, and a summary shopping-basket block is shown at lower left of Fig 4. The shopping basket frame 415 is available for monitoring selections made from the results list 413 in the conventional way. Further details including downloadable data sheets

595 may be acquired by the user.

Figs 5 and 6 illustrate evolution of a search for tilt and inclination sensors. In Fig 5, the user has selected only the broad field of "Sensors" from the available list as a catalogue title and 67 hits (representing all sensors included in the database at the time) are to be shown to the user in 4 frames. This would be tedious to read through. In Fig 6, the Index has been defined as 600 "Tilt & inclination" by selecting that term from the Index box. Further, the product type has been specified here as "Fluid filled inclinometers" and the right-hand frame is currently showing 4 complete frames out of 8.

Because this Example is used in a business that often provides quotes for bulk orders, the shopping basket procedure conveniently has the options of making an immediate payment or 605 requesting a quote for an order in bulk. The "user area" button 403 opens a user's personal page within this web site where the most recent past searches are displayed and any quotes requested are, once prepared by an executive, made available to the user for acceptance or refusal. Preferably each set of user area contents is held under password protection in a computer at the source of the web page.

610 Fig 5 is a screen copy of a similar search and this example is for rotary switches of a certain manufacturer's family. Note that the software can show images within the list of hits. Fig 6 shows a copy of a screen used during preparation of a catalogue including means to accept image files (cf Fig 1). Note that the terms comprising the Catalogue Data Set extend across the page and have the same headings as the Selection Data set. Descriptions, illustrations,

615 and prices are also entered here.

EXAMPLE 2

This is a Searchable Library of birds, for use in finding a species of bird within a collection of known birds. This library (database plus software according to the invention) may be usefully held entirely within a laptop computer carried into the field. In 'stand-alone" cases such as

620 this, there is no requirement for communication over the World Wide Web or other network. Some differences from example 1 include that more than one catalogue may be generated during initial classification of the same group of individual species. This provides alternative views of the same group; useful to different people or at different times. The Linnaean zoological classification system from phylum down to sub-species is of course hierarchical, but

625 might comprise the item to be searched for rather than the search terms to be entered. Habitat, behaviour, and zoological classification are examples of three catalogues, one of which may be selected as the first step in a search. The specific names and number of some other hierarchical selection boxes to be shown on the screen would follow from the catalogue. For example, under "habitat" there may be type (mountain, plains, ocean), location (country,

630 terrain or latitude), then bird-specific Descriptors like size, colour, call, or type of flight. Advantages of this method for management of the raw data include that a user is guided towards entering a valid search Descriptor and is more likely to quickly find an entry, especially given that if the first pass is nearly right it is easy to modify the search and repeat it. Again, prior classification by an expert will assist the novice user who may not usually select

635 the preferred words. A similar, multi-catalogue approach may be used for plants or trees or library books (for example).

VARIATIONS

Other purposes to which the principles of the invention may be applied include: any programme in which the user has to find the desired file to open, but the file names are not 640 sufficiently informative, any other requirement to search through a group of items, or a search engine for use on the World Wide Web. The invention could be used with (for example) any other warehouse holding a range of item for sale, documenting for retrieval a collection in a museum, or a book shop.

Many variations may be made to the displayed screen. For example the interface may show a 645 user how many items remain in the selected group during the process of making a more and more specific selection. Speech may be used instead of a pointing device. When a file is downloaded from the server, the system may construct a directory structure like that of the host system within the requesting computer, for the sake of better organisation.

COMMERCIAL BENEFITS and ADVANTAGES

650 The inventors consider display of the SDS to be a significant advantage of their invention. In the present invention, each of the search terms remains visible within its same selection box (405-410) after performing a search and while the searcher is inspecting the items found (413), so that the searcher can see how he or she has arrived at a certain set of retrieved items 315. A first pass at searching may be only approximately correct, and the searcher will want to 655 repeat a similar search. Given that some fields may involve keyboard entry of long non- intuitive strings (such as parts codes) that have an inherently high error rate, it could be quite frustrating to force a user to re-enter data many times over during a search procedure, although that is the norm. Some searchers may have disabilities in relation to pressing keys on keyboards, problems in bringing relevant data (possibly in different volumes of handbooks)

660 together on the desk for re-entry, or may be coping with difficult communications links. Many prior-art search engines clear the entire set of terms when one term is modified or when a search is executed, and the user must re-enter them all. Under current browser and web- page related software constraints, the SDS terms on display has to be deliberately refreshed from the server automatically and transparently by software after each pressing of the

665 "Submit" button (411) or after each receipt of material to be displayed.

Specific advantages include:

1. Anyone unfamiliar with a relevant classification system for a group of items can benefit from the planning and thought involved in classification of the items as provided in classified catalogues, (see Figs 1 and 2).

670 2. In order to facilitate its use by likely customers inexperienced in the art of on-line searching, a preferred "user assistance" feature of the invention is that the user is only able to apply a search term if an entry for that Descriptor exists, implying that an item has been classified and is available for finding. As a result the user cannot search for things that are not on offer. Inspection of the list immediately shows what is held in the database.

675 3. Differences in naming items by people from different language, cultural, or educational backgrounds are minimised. Mis-spelling or faulty entry of non-existent code numbers is not permitted (except within the keyword search field).

4. Correct names (such as scientific names for common plants) may be encouraged.

5. The invention more closely mimics the action of opening one or more printed catalogues 680 in order to select an item, yet it retains the known advantages of pubUshing on the Web, such as no actual printing and distribution costs, better recording of user activity/interest, and a trivial updating procedure. More than one printed catalogue can be merged into one electronic catalogue. The closer resemblance to a paper catalogue search provided in software according to the invention means that a user finds the search engine easier to use, 685 especially when browsing. That greater ease of use means that a vendor (or web-site owner) finds that more buyers complete the process of searching and will place orders.

6. Internet traffic is kept relatively low during a search because the search terms are trans- mitted only once to the user's terminal. Further, most searches will result in a speedier, successful conclusion. The programming approach used here (not to use hyperlinks) 690 means that only one search-related screen (Fig 4 / Fig 5) is displayed (though with necessary variations of content). Had hyperlinks been used, many screens would be shown during a search along with necessarily greater amounts of Internet traffic, and delays).

Finally, it will be understood that the scope of this invention as described and/or illustrated herein is not limited to the specified embodiments. Those of skill will appreciate that various 695 modifications, additions, known equivalents, and substitutions are possible without departing from the scope and spirit of the invention as set forth in the following claims.

Claims

We claim:

1. Software having the technical effect of facilitating the discovery, by a person using a terminal, of at least one desired item from within a catalogued collection of items 700 represented as entries within a database held within a digital computer, characterised in that the software:

a) provides a database structure wherein each database entry includes a corresponding hierarchical organisation of controlled classification terms with more than one level of hierarchy; the levels ranging from broad to narrow;

705 b) provides a classification step comprising means for the creation and maintenance of a first set of controlled classification terms (a Classification Data Set or CDS as herein defined); one finite range of terms being created for each level of the hierarchy; one or more of the terms in each finite range being capable of distmguishing items within the catalogued collection of items from each other, and terms in a first level of the hierarchy

710 being linked to corresponding terms in an adjacent level of the hierarchy;

c) provides a subsequent loading step comprising means for classifying an item to be loaded into the database, wherein for each level of the hierarchy the item is matched with one or more appropriate terms for that level, thereby creating a database holding, at each level of the hierarchy, a subset of the finite range of classification terms created for

715 that level comprising the classification terms that have been matched at least once to the items loaded within the database, thereby creating an item data table;

d) provides a subsequent searching step including means for display of a search board (as herein defined) at the terminal and means for placing a set of search parameters into the search board, thereby creating a Structured Data Descriptor (SDS), wherein the person

720 at the terminal preparing to search within the database is taken in sequence through the hierarchy and at each level those classification terms from within the subset remaining available for selection, in relation to selections made by the person at any broader level are displayed as selectable search parameters (including a parameter having an effect of including all the terms remaining available at that level of the hierarchy), so that the

725 software includes means to indicate at the time of preparation whether any items as specified are present within the database and so that the parameters can be changed prior to execution of the search according to the SDS; e) provides means for display to the person of details of any items discovered within the database as a result of execution of the search according to the SDS, and

730 f) provides means to retain the current SDS after an earlier search so that the SDS remains visible and capable of amendment and re-use, so that the retained SDS may be re-used in order to explore a range of items held in the database, so that the person can more quickly discover a most suitable item.

2. Software as claimed in claim 1, characterised in that the software is adapted in relation to 735 loading step (c) in order to accept further classification terms into the CDS during performance of that step should an appropriate classification term be absent from the finite range of controlled classification terms.

3. Software as claimed in claim 1, characterised in that the software is adapted in relation to the first set of controlled classification terms in order to accept further alternative sets of

740 controlled classification terms terms at corresponding hierarchical levels, while still referring to the same catalogued collection of items.

4. Software as claimed in claim 1, chamcterised in thai the software is adapted in order to provide means for retaining the SDS in a storage medium under control of the software in a form comprising a storable and recallable entity holding data, so that a person can recall a

745 search strategy at a later time.

5. Software as claimed in claim 1, chamcterised in that the software is adapted in relation to step (d) in order to offer a facility for accepting at least one key word from the person; the at least one key word not forming part of a set of classified terms, the at least one key word being selected by the person as a search parameter, so that queries using uncontrolled

750 key words are not excluded.

6. Software as claimed in claim 2 or in claim 3 or in claim 4, characterised in that the software is adapted for operation within a single digital computer that holds the database and also serves as a terminal for facilitating the discovery of items within the catalogued collection of items held within the database.

755 7. Software as claimed in claim 2 or in claim 3 or in claim 4, characterised in that the software is adapted for operation across a network (as herein defined), the software including (a) software within the first computer that includes host database provider means and to hold the catalogued collection of items, and (b) derivative software generated within the first computer by the software in accordance with the person's instructions and 760 communicated to the remote terminal, thereby providing functional compatibility between the first computer and the remote terminal.

8. Software as claimed in claim 4, characterised in that the software retains the SDS in a storage medium within the host database provider; the SDS being linked to an identifier describing the person who created the SDS, so that a person can recall a SDS from any

765 remote terminal.

9. Software as claimed in claim 7, characterised in that the software is particularly adapted for discovery of one or more items from within the database and acquisition of the selected items by the person; the adaptation including means for:

a) inclusion of facilities for the selection by the user, after execution of the hierarchical 770 search strategy at least once, of one or items that have been placed in the database,

b) generation of a list of the one or more selected items, and

c) information leading to a process for acquisition of the one or more selected items. Preferred corresponding functions for the case of a warehouse catalogue include the functions of: (a) purchase, (b) quotation, or (c) provision of further information.

775 10. Software as claimed in claim 6, characterised in that the software is particularly adapted for discovery by the person of information concerning items previously entered into the database.

11. A digital computer characterised in that the digital computer is ranning software as claimed in claim 6.

780 12. A first computer and a terminal, characterised in that the first computer and the terminal are ranning software as claimed in claim 7.

13. A method for classifying or cataloguing at least one item in order to render the item findable by a searcher from within a catalogue-like database holding a plurality of catalogued items; characterised in that the method includes:

785 a) creation of a classification system having more than one hierarchical level for describing the items to be included in the catalogue-like database wherein each level is provided with an associated list of Descriptors reflecting the options available at that level, b) a process for classifying each item and means for addition of a Descriptor, if not already present, to the list of Descriptors within each hierarchical level, and

790 c) making an internal link between the item and a corresponding Descriptor at each level, so that subsequent selection of a partial or complete hierarchical set of Descriptors will retrieve at least one corresponding item.

14. A method for retrieving information specifying an item catalogued according to claim 13 from within the database by a person operating a terminal, characterised in that the further

795 method includes the steps of: a) placing a selected search term in zero or more of a hierarchical range of selection boxes displayed at the terminal, and b) optionally entering a search modifier selected from a list of permitted modifiers or entering a free-form modifier, and

800 c) causing execution of a search within the database employing the entered search data set, and d) inspecting a display of a set of item-related information for each found item, if any, for which a coincidence between the selection data set and the item Descriptor occurred, and

805 e) the invention further providing retention and display of all search terms used so that one or more of the selection data set can be edited for re-use by the searcher so that the search can be repeated.

15. A search board interface for a catalogue held as an electronic data base as claimed in claim 13 characterised in that the person may use a search board to search within the catalogue

810 for one or more items each represented by a record and having more than one field for each record; the interface providing that: a) more than one selection box, each corresponding to one of a hierarchy of Descriptors, is shown on the display device; b) the user selection of a Descriptor for loading into a hierarchical selection box is limited 815 to valid Descriptors in that field; c) each Descriptor is presented again within the corresponding selection box after the search is ran; d) each Descriptor is available for modification in hierarchical order prior to ranning another search; and

820 e) all user searches, regardless of the number of levels of hierarchy, take place within a single frame set, so that the interface allows the user's selection of any one or more Descriptors to be inspected and optionally edited for re-use after each run of a search procedure, and so that the interface is more easily and more accurately selected for repeated use.

825 16. A procedure for classifying or cataloguing and then finding at least one item as claimed in claim 13, the data being stored within a computing means as a database holding a plurality of items, characterised in that the procedure includes a) a classification method for items and information for use during the creation of the data base, the method including at least one level of hierarchy and the creation of one or

830 more Descriptors that indicate a hierarchical stracture; b) input of one or more items including their classification from the range of permitted Descriptors within that hierarchical classification data set in order to specify at least one or more attributes possessed by an item to be catalogued, and c) a finding method for retrieving information concerning the said catalogued item from 835 within the database by a searcher operating a terminal, wherein the finding method includes the steps of i) entering none or the highest one or as many of an hierarchical range of selection boxes displayed at the terminal, and ii) optionally entering a search modifier selected from a list of permitted modifiers or 840 entering a freeform modifier, and iii) causing execution of a search within the database employing the entered search data set, and iv) inspecting a display of a set of item-related information for each found item, if any, for which a coincidence between the selection data set and the item Descriptor 845 occurred, and v) the invention further providing retention and display of all search Descriptors used so that vi) one or more of the Selection Data Set can be edited for re-use by the searcher so that the search can be repeated.

850 17. A search board interface for a catalogue held as an electronic data base as claimed in claim 13 wherein a user may use a search board to search within the catalogue in order to locate one or more items each represented by a record and having more than one field for each record, characterised in that the interface provides that: a) more than one selection box each corresponding to one of a hierarchy of Descriptors is 855 shown on the display device, b) the user's selection of a Descriptor for loading into a selection box is limited to valid Descriptors in that field, c) each Descriptor is presented again within the corresponding selection box after the search is ran,

860 d) each Descriptor is available for modification in hierarchical order prior to running another search, and e) all user searches, regardless of the number of levels of hierarchy, take place within a single frame set, so that the interface allows the user's selection of any one or more

Descriptors to be inspected and optionally edited for re-use after each run of a search

865 procedure, and so that the interface is more easily and more accurately selected for repeated use.