Browse Prior Art Database

Method to detect user's preferred computational operation and adjust scheduling accordingly

IP.com Disclosure Number: IPCOM000239342D
Publication Date: 2014-Oct-31
Document File: 2 page(s) / 33K

Publishing Venue

The IP.com Prior Art Database

Abstract

Determination of a user's preferred computational operation by analysis of user interface interactions and subsequently automatically adjust the operation's priority on access to resources such as CPU time or disk reads and writes.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 50% of the total text.

Page 01 of 2

Method to detect user's preferred computational operation and adjust scheduling accordingly

Disclosed is a method for automatically prioritising computational activity, such as a process or thread, by analysis of user interactions. Scheduling of computational operations, such as a thread or process, does not take into account the user's current preferred operation. While a user can assign a priority to a process within an operating system, or a thread within a program that allows for this, it does not take the user's current behaviour in to account. For example, the user tells the system to prioritise operation A, but then while working on operation B they wait for this to complete. Operation A is still running but the user has forgotten they prioritised this, so operation B is taking longer even though the user would prefer this operation to get more processing time.

    A real-world example would be, a user of ApplicationXYZ has started an import of some documents which can take a long time. While the import is ongoing, the user realises they need to export a dataset to give to a colleague or they will block their work. The export is also going to take a long time. It is not clear to the user which of these operations has priority, but clearly the user would prefer the export to take priority. If the export takes too long, the user may be tempted to cancel the import so the export can use all of ApplcationXYZ's processing resource. This means work relying on the import will be delayed, but also cancelling an import could lead to a broken dataset that the user would have to spend time cleaning up.

    Existing non-standard prioritisation techniques include:
 Operating systems let you adjust process priority. This does not persist across sessions by default, but there are tools that can remember the priority you assign, e.g. http://www.ghacks.net/2008/03/01/automatically-adjust-the-process-priority/
 There are tools available to automatically "re-nice" a process (therefore lower it's priority) after a certain condition is met, e.g. ticks on the CPU
( http://and.sourceforge.net/ )
or application type ( http://thermal.cnde.iastate.edu/ŝdh4/verynice/ )
 Analysis of thread interdependencies so priority can be given to threads a stalled thread depends on to continue. http://www.stanford.edu/~mebauer/pdfs/paper.pdf

    While the known solutions provide good scheduling, they do not take in to account the user's preference. The proposed solution is to monitor the system user's actions and detect patterns in the behaviour that could indicate they have a preferred operation that they would like to complete first. While not suitable for server environments, if implemented in a desktop environment it could make a user's experience much better.

    The proposed solution detects a user's preferred operation by analysis the user's UI interaction. Simply put, if the user repeats an action that has no side-effect or no perceived value then the system can as...