Logic Production Systems (LPS)
Welcome to "LPS Corner", a public repository for open source development of LPS engines, examples and related tools. In additional to several experimental engines at Imperial College, there are currently two maintained implementations:
- lps.swi, based on SWI-Prolog+SWISH, running on Linux containers, maintained in this repository
The two engines execute slightly different LPS language versions; thanks to an ongoing academic project, over Spring 2019 lps.swi is expected to "cross-compile" LPS for lps.js.
Logical Contracts Server, maintained elsewhere, is a proprietary extension to lps.swi.
The simplest way to try LPS is to use a public engine instance rightaway:
Use the last items of its Examples menu, e.g. First Steps with LPS.
The second easiest way is to use Docker::
docker run -p 3051:3050 \ -v /MyData:/data \ -v /MyLogs:/home/lps/engine/logs \ logicalcontracts/lps.swi
The above starts up a SWISH server with LPS at port 3051 on your system; to use it just take your browser to
http://localhost:3051. Make sure you have Docker installed in your system. MyData and MyLogs should be two empty writeable directories.
For more details please see INSTALL_server.
Previous version: D-Day, 2018
- Bug fixes
- Legal contract examples
Previous versions: March 2, 2018
This is the first update since Summer 2017, and incorporates most improvements developed by LogicalContracts:
- LPS background servers
- LPS program hibernation / restore
- External (Prolog) fluents and events
- Simplified "fluent editing" actions
- Simulated real (calendar) time
- Bug fixes
.. and more, please see preliminary documentation at the LogicalContracts site.
What is LPS?
“Logic-based Production System" is a new computer language that combines the characteristics of an imperative programming language with those of a declarative database and knowledge representation language. It is the result of over a decade of research led by Bob Kowalski and Fariba Sadri at Imperial College London.
The kernel of LPS consists of a database, together with reactive rules of the form if antecedent then consequent. The database changes destructively in response to actions and external events, according to a domain-specific causal theory. Computation consists in making the reactive rules true, by performing actions to make the consequent true whenever the antecedent becomes true. In addition, LPS includes Prolog-like logic programs both to recognise when antecedents become true and to generate plans of actions that make consequents true.
Additional material about the kernel of LPS is available on the RuleML wiki. See also the slide presentation at 1st CLOUT workshop 2017. Other technical papers and bibliographic references can be found on Bob Kowalski's home page at https://www.doc.ic.ac.uk/~rak/, which also includes a copy of his 2011 book, which presents some of the philosophy underlying LPS.
Trying LPS online
See and try the language at http://demo.logicalcontracts.com:
Start with menu Examples / First Steps with LPS and links therein, as well as "LPS Examples". Refer to the LPS syntax.
For regular or intensive usage you can install LPS on your machine, using a number of alternatives:
Installing your own
Besides trying it rightway on your browser at http://demo.logicalcontracts.com, you can install LPS:
- as your own local instance of the lpsdemo web application, on Windows, Mac, Linux
- as a server in your own private or university cloud
- as a SWI Prolog program
Please refer to the INSTALL instructions.
XSB Prolog support is deprecated, but still available in the Aug 15, 2017 version
License and Copyright
All files in this repository are copyright Imperial College London and open source licensed with 3-clause BSD, except for files with other specific copyright and licensing notices, all being some sort of open source.
Main authors include Bob Kowalski, Fariba Sadri, Miguel Calejo and also contributions by Jacinto Dávila. The engine implementation in this repository is an evolved version of the core interpreter developed during David Wei's Master Thesis project at Imperial College, London, supervised by Fariba Sadri and Krysia Broda.
Core Engine + Prolog Studio (deprecated)
InterProlog Consulting's Prolog Studio adds somes facilities for LPS running over XSB Prolog: a semantic highlighting editor and a timeline visualizer. To install:
- Pull or download this whole repository into a new folder "LPS"
- Follow instructions at top of InterProlog Studio, including a recent XSB Prolog install
Other projects using LPS
See the robot game at https://bobthesimplebot.github.io
How to contribute
- See lps_corner/RoadMap.md
- Before you commit any changes to the LPS engine or tools, make sure you execute all tests with 'interpreter:test_examples'.
- Please comment your code, and add appropriate copyright/license header
- To edit this and other future wiki documents: learn Markdown
If you add a file to the examples directory and execute it with the make_test option, e.g.
go(MyFile,[make_test]), a test results file MyFile.lpst will be generated. If the program behaves correctly with the current version of LPS, commit this file too.
When you later execute go(MyFile,[run_test]), the program behavior will be compared with the previous test results; ditto when you run all tests, with 'interpreter:test_examples'.
Is there support? Who do I talk to?
The following contacts imply no obligation nor guarantee:
- Bug reporting and occasional support:
- To find out more about LPS/KELPS in general: