1.1. Field of the Invention
The present invention relates generally to collaborative document systems like message control systems, and more particularly to electronic mail control systems.
1.2. Description and Disadvantages of Prior Art
Collaborative document systems represent a well established technology allowing to share electronic documents within a community of a large number of distributed users using their interconnected workstations for access to these collaborative document systems. A common feature of this technology is that the electronic documents to a certain extent obey a predefined structure of interrelated data fields. Typically the particular data fields making up an electronic document can be freely defined; examples for instance are: project name, subject, author, list of team members representing the primary editors, additional editors and the like. Type and number of these data fields typically depend on the particular purpose the electronic documents are used for.
A particular example of such collaborative document systems are the digital communication systems, called messaging systems. One specialized variant of these systems are of the “electronic message” or “electronic mail” type, which often are simply called email systems. In such systems, several or even a very large number of work stations are interconnected by a system which allows users at the work stations to send messages to each other. Such messages are the electronic equivalent of letters and memoranda.
In a typical messaging system there will normally be a standard format defined for such messages (or, often, a number of slightly different formats for different purposes). The format will normally have several fields, including a sender field, a date field, an addressee field (the TO: field), and a message body field; in addition, there will often be further fields, including in particular a “copy-to” field (the CC: field) and a “subject” field. The addressee and copy-to fields will often each be capable of holding more than one identifier, so that a message can be sent to several addressees and copied to several more. The “addressee” and “copy-to” fields are normally functionally equivalent as far as the system is concerned, with the distinction between the two being of significance only to the users.
In a typical prior art system, when a user generates a message, the system will automatically transmit it to the addressees (including the copy-to's). At each of the stations to which it is sent, it will be stored in a folder (which may well be termed the “in-tray” folder). When the users at that station choose to manually inspect the contents of the in-tray, they will extract the various messages from it one after another, inspect each, and take whatever action they think appropriate, such as deleting the message, printing it out, or storing it in some other folder.
In practice, it is found that such a procedure can become burdensome, because the number of messages which accumulates in the in-tray can become so large that a considerable amount of time and effort is required for the user to work through the accumulation of messages in the in-tray. Because the usage of such systems nowadays is so widespread and their handling has become quite simple the number of messages the typical user is receiving every day may be quite large; one may even expect that this number will increase further.
To cope with these increasing numbers of messages various technologies have been suggested within prior art.
Certain collaborative document systems, like for instance Lotus Notes, allow a sender to attach some type of sender specific priority to a certain message. The “Urgent” flag within Lotus Notes allows a sender to mark messages with the following levels of urgency: “Urgent”, “Normal”, “Just for information”. A serious disadvantage of such approaches is that this priority expresses only the author's (e.g. sender's) perception of urgency rather than the reader's individual perception of urgency, not to talk about a sender who willfully “oversells” the importance of his message. In particular if the document is targeted for several users, this mechanism cannot work.
Other prior art approaches consist in establishing filters which automatically filter out documents based on attributes occurring within a message (like for instance author, subject, size). This is an approach typically taken to filter out spam messages, i.e. messages advertising for some product sent to a mailing list or newsgroup.
Other teachings, like for instance U.S. Pat. No. 5,377,354, provide technology for prioritizing a plurality of incoming electronic mail messages for a user by applying user created rules, which are stored in a rules-store, to said user's messages. After execution of these rules priority numbers are assigned to the messages which permit to sort said user's list of messages in accordance to this computed priority number.
To improve the handling of these rules the teaching of W. W. Cohen in “W. W. Cohen, W. W. 1996a. Learning rules that classify e-mail. In Papers from the AAAI Spring Symposium on Machine Learning in Information Access, 18-25” discloses to dynamically create such rules based on a proposed learning process.
Despite of all of these advances further progress is still required as users of such collaborative systems, especially messaging systems, are not satisfied with the current state of the art.
1.2 Objective of the Invention
The invention is based on the objective to improve the handling of messages directed to an user of an electronic messaging system.
The objectives of the invention are solved by the independent claims. Further advantageous arrangements and embodiments of the invention are set forth in the respective subclaims.
The present invention proposes a computerized method and means for providing an improved handling of messages directed to an user of an electronic messaging system.
The suggested method comprises a first step of determining user characteristics capable of identifying the user as an addressee of a message.
The method further comprises a second step of analyzing a certain message and determining its message characteristics exploiting the information contained in one or a multitude of message fields.
The method finally comprises a third step of applying at least one deduction rule of a potential set of deduction rules to said certain message. The deduction rule assigns based on the user characteristics and the message characteristics a role of said user to said certain message. The role is capable of revealing said user's relationship to said certain message.
According to a fundamental observation of the current invention the recipient's role with respect to a certain message identifying the user's relationship to the particular message is a most important element to decide how to handle that message. Knowing the user's role with respect to a certain message the user can decide, in most cases without knowing the details of the message contents (i.e. without having read and having understood the message), how and when to deal with that particular message.
Being able to derive a recipient's role with respect to a certain message automatically saves a lot of working time for the recipient. Users of messaging systems report that they have to spend a lot of their time only for the purpose to understand how to deal with a certain message and how to handle it. Providing the recipient's role according to the current invention makes it in many cases unnecessary to spend time consuming inspection steps for deciding how to handle that message; i.e. as a result a recipient has to read a particular message only once. Thus the proposed technology supports a user to much more selectively work with a multitude of messages.
It is a further fundamental observation of the current invention that the recipient's role with respect to a certain message most clearly expresses the sender's expectation how that recipient has to handle a message (from the sender's perspective). It is the recipient's role which easily allows to decide for example whether a recipient is required to take notice of that message only (representing a “for your information only” message), whether a recipient is expected to open a message and read it carefully, whether a recipient is expected to perform some actions described within the message or whether a recipient is expected to reply with a further message back to the sender.
The teaching of the current invention is orthogonal to other prior art approaches for the same problem, i.e. it enhances the current technology in a new direction such that it may be combined with other known technologies, like for instance prioritization approaches, filtering approaches and the like. This provides the important advantage of synergistic improvements.
The suggested approach has the further advantage that no message documents are filtered away, thus no data has been lost. A user is never losing control and the user can identify at a glance based on the role assigned to each message those messages where an action is required and thus someone else may depend on. This can be done without doing manual scanning of each document; instead, the individual role is computed automatically.
In the drawings and specification there has been set forth a preferred embodiment of the invention and, although specific terms are used, the description thus given uses terminology in a generic and descriptive sense only and not for purposes of limitation. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the appended claims.
The present invention can be realized in hardware, software, or a combination of hardware and software. Any kind of computer system—or other apparatus adapted for carrying out the methods described herein—is suited. A typical combination of hardware and software could be a general-purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein. The present invention can also be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which—when being loaded in a computer system—is able to carry out these methods.
Computer program means or computer program in the present context mean any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following a) conversion to another language, code or notation; b) reproduction in a different material form.
The current invention is illustrated to some extent based on the commercially available collaborative document system “Lotus Notes”. Of course the current teaching applies also to any other type of system, which offers comparable collaborative document functionalities in general or messaging capabilities in specific. Even if the described embodiment of the current invention focuses on messaging systems the current teaching may be applied to all collaborative document systems as well.
4.1 Introduction
A message to be sent from a sender to one or more recipients comprises several parts also called message fields. Preferably and specifically, the following fields, which can be exploited by the current invention, can be identified:
A Sender Field
The sender field contains the identity of the sender, i.e. a string representing the (worldwide unique) userid of the author. This identity is a unique identifier used by the messaging system. Depending on the specific design of the system, the identifier may be chosen in a form which also has a mnemonic significance, or the system may include conversion tables which convert the identifier to a human-recognizable name.
A Date Field
An Addressee Field—the TO: Field
This contains the identities of the addressees to whom the message is being sent, in the form of identifiers like the sender identifier. There can be more than one addressee. This field even may contain a symbolic distribution list. The field represents the primary addressee(s) of the mail. In messaging systems this field even may be empty in case the fields CC: or BCC: described below are filled in.
A Copy-to Field—the CC: Field
This field contains one or more strings, each one representing either an addressee or a symbolic distribution list (comprising a multitude of addressees). This field represents the secondary addressee(s) of the message to whom a complimentary copy of that message will be sent.
A Further Copy-to Field—the BCC: Field
This field contains one or more strings, each one representing either an addressee or a symbolic distribution list (comprising a multitude of addressees). This field represents the secret secondary addressee(s) of the message to whom a complimentary copy of that message will be sent. No other addressees are aware of secret secondary addressees, not the addressees on the TO: and CC: fields and also no other secret secondary addressees—giving this field its name: “Blind Carbon Copy”.
A Subject Field
This will contain a short heading identifying, to the addressees, the subject matter of the message.
An Urgency Indicator
This field carries optional information that the sender expresses especially high (or sometimes especially low) “urgency”. It defines the priority of the message according to the view of the sender.
A Body Field
This will contain the main text of the message, and will be of highly varied length; it represents the actual payload, i.e. the text transported from the sender to the addressees.
A Greeting Field, also Called Salutation Field
This field is optional and contains the greeting, e.g. for instance the first name(s) or nickname(s) on a separate line. In many implementations (e.g. Lotus Notes) this field does not exist as a separate message field, but any greetings are stored in the body field (see above) instead.
A Closing Field, also Called a Footer
This field may contain a closing phrase (e.g. for instance “Best regards”) followed by a short footer or stamp containing things like a phone number or address or other information for describing the sender like business entity the sender is working for.
A typical scenario a user of messaging systems is confronted with might be the following. The messaging system proposes to the user a list of messages (e.g. the set of incoming messages) addressed to him/her. Usually an overview panel lists a one-line description for each message; this line might comprise the sender of the message, the sending date, the size of the message and may be also a subject line, defining the subject as viewed by the sender. Especially if the number of these messages is quite large a user's problem is to handle the multitude of messages in a most time-efficient way.
For some of the messages a swift response by the recipient is required, for other messages only a careful read is necessary while other messages could remain unopened allowing the user to come back to it if required at a future point in time. Based on the information made available by the current state of the art messaging systems the user has almost no other chance for deciding how to handle a particular message than to read into each of the messages to understand the particular situation and then to decide how to handle it. This type of scanning activity needs some time, which may add up to significant amounts of time per user and year.
Current state of the art solution approaches allowing a user to deal with this problem are the following:
As known solutions approaches might improve the problem situation but nevertheless remain unsatisfactory the current invention is targeting at an orthogonal direction compared to prior art approaches based on the following observations:
Based on these observations the following overall methodology is suggested for improving the handling of messages directed to an user of an electronic messaging system:
The particular user is that user who logged into the messaging system with a unique (personal) identity key (e.g. userid and password). From this identity key the user specific attributes, called the user characteristics, may be determined. These user characteristics may comprise one or several of the following elements:
These user characteristics turn out to be quite successful, of course other user characteristics may be used instead.
According to the observations mentioned above for each of these messages the expectations to the user (the recipient) may differ with the role he/she has in this message mail (e.g. user should respond, or user should just take notice of the content without response). Thus the role is reflecting the expectations of a sender with respect to the addressee; automatically determining the addressee's role is most helpful to understand these expectations without having to inspect the particular message. Some of the expectations make the processing of a message/document time critical, since they imply that others wait (e.g for a response), while other expectations are not time critical, since no other individuals wait for an immediate reaction. Each message document may be made available to several users, where different users may have different roles for the same document.
The following automatically determined roles turned out to be very helpful:
When discussing the deduction rules in more detail further below the current specification will return to the roles once more. Of course above mentioned roles represent examples only, depending on the specific application area other roles can be used instead.
The computer system renders the message to the user, according to the user's role wrt. to this message. Preferably, the role for each message is already shown in the overview panel that displays a set of messages at a glance (e.g. one-line text or one icon per document). The role can be encoded as icon, or expressed by any other attribute (e.g. using font, size, color, bold, italics, underline attributes, blinking).
Alternately, the role may only rendered implicitely by sorting/arranging the messages in the overview page in an appropriate order (e.g. the messages where the reader has to act/respond displayed first). Optionally, the rendering can be done in the already opened document as well; even though it would be best to save the user opening of the document at all for finding out the role, even a small improvement to prior art would be to assist the user in determining his/her role faster. This can be done e.g. by highlighting any occurrence of the user's name and/or userid when displaying the content of the document; for example in an message document, it will speed up the “manual” scanning when the user's name is highlighted inside a list of other user's names. In addition, the computer system can do a simplification of the found strings (e.g. in a messaging system, it may strip redundant trailers like “/Germany/IBM@IBMDE” from all userids that have the same trailer as the current user—like you can strip country code and area code in phone numbers that have the same country and area code).
4.3 Deduction Rules
The above mentioned rules typically are made up of a condition part and an action part. The condition part may be expressed as an Boolean predicate combining some of the user characteristics and/or some of the message characteristics. If the condition part evaluates to true, the action part will be fired. The action part of a deduction rule finally assigns a particular role to a message.
In the most simple case the proposed method suggests to search for occurrences of the user's identity key, user-specific attribute strings available in the system (like mail-userid U, full name N and nickname NN) and restricts itself to match only strings in certain fields with strict semantics (e.g. in a mail system, it may look only for occurrences of U, only in the fields “To:” and “cc:” and “bcc:”).
Preferably however, the computer system uses all information in the message/document, conventions, heuristics, and user preferences for distinguishing documents with immediate actions required for the individual user from documents without immediate actions required.
The following deduction rules have been proven to be quite successful:
The advantage of above mentioned rules is that the condition part is having a low complexity only. Of course more complex condition parts are possible. The example deduction rules given below actually represent heuristics for discovering wrong use of conventions with respect to messages. Since many users receive more and more electronic mail documents than they can afford to read carefully, more and more mails are discarded without being opened. Thus, in anticipation of such an attitude some senders “oversell” their mail documents (i.e. they try to make the mails seem more attractive than they are) in order to improve the odds that the addressee(s) read their mail document which would be discarded unread without overselling.
Some successful deduction rule examples of a more complex structure are the following:
This deduction rule is targeted at situations of overselling messages by generating mass mail (spam) with personalized address field (To field). The selectivity of this rule can further be enhanced if a search for the user's name and/or nickname in the message body is performed; if no such name can be found within the body of the message a role of DISTRIBUTION-LIST is assigned.
Also this deduction rule is targeted at situations of overselling messages by generating mass mail similar to the preceding rule. To determine an existing business relationship the electronic address book of that user could be analyzed and established business relationship could be made part of that user's characteristics.
The features given below represent optional enhancements which can be combined with the technology outlined so far for achieving further advantages.
These points in time for execution of the methodology are ordered in sequence of decreasing resources consumptions as well as decreasing responsiveness.
Step 100 indicates the beginning of the proposed methodology. Step 101 is dedicated to the determination of the user characteristics of the particular user; in one embodiment of the current invention this could mean to determine the user's userid his/her (full or partial) user name or even the nickname. Within the next step 102 a particular message is analyzed to determine that particular message's characteristics. For the embodiment of the current invention within an e-mail oriented messaging system this could mean to determine the contents of the TO:, CC: or BCC: message fields, the body of the message, the salutation and/or the footer within the message or similar types of data. Step 103 comprises the application of one or a multitude of deduction rules, wherein based on the user characteristics (determined within step 101) and based on the message characteristics (determined within step 102) said user's role is determined and assigned to that message. This role is capable of revealing said user's relationship to the message. Finally within step 104 the determined role is rendered and thus visualized to the user. Repeating the proposed methodology along path 105 would mean to execute the rendering activities in a single step 104 for all assigned roles; while alternatively it is possible to perform the rendering for each individual assigned role and start repetition of the methodology once this has been completed along path 106.
More details on the rendering process is given by
A second type of visualization is given in column 206. According to this second approach for each of the assigned roles a different icon is visualized.
As already pointed out above the most important advantage of the current invention by being able to derive a recipient's role with respect to a certain message consists in saving a lot of working time for the recipient. The recipient's role with respect to a certain message made available to the user automatically eliminates the necessity of the recipient to manually inspect the message for deciding what to do with that message. A further advantage is the recipient's role with respect to a certain message most clearly expresses the sender's expectation how that recipient has to handle a message (from the sender's perspective).
When a sender sends an electronic message document to one or more addressees, there are certain expectations what the addressees should do with the mail. In most cases the addressees are expected to perform one or more of the following activities:
Even assuming for a moment that all addressees would have enough time so each addressee could perform each of these activities for each mail document, this usually is not even desired by the sender. In other words, different addressees of one mail document will usually only perform a subset of these activities. This will depend on the concrete topic of the mail document as well as on the job description of the involved users. In other words, it depends on the recipient's role (i.e. the recipient's relationship) to said mail document to understand what the sender's expectations are. If the recipient's role can be a derived automatically then also the sender's expectations on the recipient have been revealed.
The following table gives a rule of thumb how the expectations of the sender can be derived from the attributes present in a mail document which led to a certain role assignment, based on the conventions implied by a typical electronic mail system. The leftmost column indicates the role, mainly defined by the fact in which field of the mail document the userid of the current user can be found, and whether it is the only content of the field or just one of two or more userids.
The different roles imply quite different expectations; thus, it seems plausible that automatically deriving and displaying the role as important feature of the mail promises benefits.
In above table, there are several table elements that have the value “maybe”; this means to compute these values automatically from mail attributes more enhanced technologies are required. These expectations do depend on sender's and addressee's context, job description etc. For their automatic derivation more enhanced technologies like those of text mining and/or data mining have to be exploited: using these technologies enhanced user characteristics and/or enhanced message characteristics can be determined which then will allow the set of deduction rules to assign roles of the recipient to the individual messages which even closer match the sender's expectations.
When sorting mail documents by role, it is evident that the sort order will slightly depend on the weight one puts on the different expectations. If possible, the enduser should be able to take influence, for example to specify how documents received in the bcc role should rank in comparison to other roles.
While the current invention to a certain extent has been described based on an e-mail system, this has to be understood as an example only. Actually the proposed teaching may be by applied to all systems which exchange messages, or in other words to messaging systems in general. Other variants of messaging systems (also comprising the concept of addressees) to which the current teaching beneficially may be applied to are:
Number | Date | Country | Kind |
---|---|---|---|
02102724 | Dec 2002 | DE | national |
Number | Name | Date | Kind |
---|---|---|---|
5794039 | Guck | Aug 1998 | A |
5999932 | Paul | Dec 1999 | A |
6052709 | Paul | Apr 2000 | A |
6779021 | Bates et al. | Aug 2004 | B1 |
7136920 | Castell et al. | Nov 2006 | B2 |
20020120689 | Kang et al. | Aug 2002 | A1 |
20020138581 | MacIntosh et al. | Sep 2002 | A1 |
20020188689 | Michael | Dec 2002 | A1 |
20030069900 | Hind et al. | Apr 2003 | A1 |
20030177190 | Moody et al. | Sep 2003 | A1 |
20040054733 | Weeks | Mar 2004 | A1 |
Number | Date | Country |
---|---|---|
505859 | Oct 2002 | TW |
509852 | Nov 2002 | TW |
511014 | Nov 2002 | TW |
Number | Date | Country | |
---|---|---|---|
20040114735 A1 | Jun 2004 | US |