You can subscribe to this list here.
2000 |
Jan
|
Feb
(2) |
Mar
(54) |
Apr
(44) |
May
(22) |
Jun
(24) |
Jul
(35) |
Aug
(34) |
Sep
(14) |
Oct
(11) |
Nov
(77) |
Dec
(21) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2001 |
Jan
(13) |
Feb
(27) |
Mar
(32) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2022 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
|
Sep
(1) |
Oct
(1) |
Nov
(1) |
Dec
(1) |
2023 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
(2) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
2024 |
Jan
|
Feb
(1) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Paulo M. <pm...@lo...> - 2024-02-01 00:11:58
|
Hi, Logtalk 3.74.0 is now available for downloading at: https://logtalk.org/ This release improves the linter warnings for unsound constructs in grammar rules and for if-then-else and soft-cut control constructs; improves the semantics and implementation of several multi-threading features; fixes compilation of "phrase/2-3" goals where the first argument is a "(::)/1-2" or "(^^)/2" control construct with a variable grammar rule body argument; fixes a regression where a compiler error when a source file term is not callable would be delegated to the backend; improves the Handbook documentation of several directives and built-in predicates; improves the Handbook documentation on dynamic predicates and grammar rules; adds a "listing" library (replacing the old undocumented one); adds a "recorded_database" library to help porting legacy Prolog code; fixes "json" library bugs; updates the "help" tool support to browse the Handbook and APIs documentation from the top-level; changes the default output directory for the "diagrams" tool to "./dot_dias"; updates the "tutor" tool for all recent linter updates; fixes a syntax error in the "logtalk_tester.ps1" script; improves the portability of the "fcube" port; updates the Advent of Code 2023 examples and several of the multi-threading examples; adds additional tests for the de facto standard "format/2-3" predicates and for the multi-threading built-in predicates; updates support for several text editors and syntax highlighters; updates the manual installation script to inform the user about the "INFOPATH" environment variable; fixes the Bash shell embedding scripts to actually use the default settings file when no "-s" option is passed; and includes portability updates for CxProlog, LVM, SICStus Prolog, SWI-Prolog, XSB, and YAP. Thanks to Yurii Rashkovskii for his bug reports and suggestions. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2023-12-12 18:00:24
|
Hi, Logtalk 3.73.0 is now available for downloading at: https://logtalk.org/ This release adds linter warnings for deprecated arithmetic predicates and functions; adds warnings for comparing numbers using unification; adds support for using backend-declared deprecated built-in predicates in linter warnings; improves checking the availability of predicates in "user" for "uses/2" and "use_module/2" directives; avoids false positive linter warnings about non-terminals called as predicates when the caller is a phrase-like predicate declared in the backend adapter file; improves compiler reporting of term-expansion errors; fixes unknown and undefined predicate call warnings when the calls occur in an included file to report the actual location instead of the main file; fixes printing of grammar rules linter warnings to respect the "grammar_rules" flag; adds adapter files support for deprecated built-in predicates and for declaring phrase-like predicates that call non-terminals; improves the Handbook grammar section now uses W3C-style EBNF syntax compatible with the [Railroad Diagram Generator](https://www.bottlecaps.de/rr/ui), also fixing typos and omissions; improves the Handbook section on parametric objects; improves the documentation of the "wrapper" tool; adds an experimental "mutations" library for generating random mutations of terms of selected types (intended for eventual fuzz testing support); adds a "tsv" library for reading and writing TSV files; adds new predicates and non-terminals to the "types", "grammars", and "random" libraries; improves the performance of the "term_io" library predicates; includes updates and fixes to the "lgtunit", "tutor", and "wrapper" tools; fixes the PowerShell documentation scripts to avoid an error when converting XML files; improves the "logtalk_tester" scripts detection and reporting of broken test sets due to backend bugs; adds new "haunted_wasteland", "scratchcards", and "trebuchet" examples (solving Advent of Code 2023 problems); adds new tests, updates, and fixes issues with the "poem", "profiling", "self_vs_this", "errors", "bench", and "benchmarks" examples; adds additional tests for the "phrase/2-3" built-in methods and fixes an issue with a "setof/3" built-in method test; improves the macOS installer; fixes the "logtalk_user_setup.ps1" PowerShell script to use a valid path for the backup directory; and includes portability updates for B-Prolog, CxProlog, ECLiPSe, LVM, and SWI-Prolog. Thanks to Domingo Alvarez Duarte and Yurii Rashkovskii for their contributions to this release. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2023-05-30 13:11:24
|
Hi, Logtalk 3.66.0 is now available for downloading at: https://logtalk.org/ This release adds new keys to the "logtalk_load_context/2" built-in predicate for use with the term-expansion mechanism; adds two new meta-messages to the message printing mechanism to support user-defined printing goals; adds new linter warnings for DCGs, lambda expressions, predicate directives and backends without a module system; improves reporting of warnings when compiling auxiliary predicates; improves the Handbook and the documentation of several libraries and tools; adds a new "ulid" library for generating Universally Unique Lexicographically Sortable Identifiers; provides improvements and fixes to several tools, libraries examples, and contributions; adds additional compliance tests for Prolog standard predicates; fixes a registry issue when running the Windows installer; and provides portability updates for LVM, Quintus Prolog, Scryer Prolog, and, Trealla Prolog. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Michael K. <ki...@cs...> - 2023-05-02 18:53:50
|
<html style="direction: ltr;"> <head> <meta http-equiv="content-type" content="text/html; charset=UTF-8"> <style id="bidiui-paragraph-margins" type="text/css">body p { margin-bottom: 0cm; margin-top: 0pt; } </style> </head> <body bidimailui-charset-is-forced="true" style="direction: ltr;"> <p>Coherent Knowledge Systems (<a class="moz-txt-link-freetext" href="http://coherentknowledge.com/">http://coherentknowledge.com/</a>) is happy to announce the first open-source release of its flagship ErgoAI (<a class="moz-txt-link-freetext" href="https://github.com/ErgoAI">https://github.com/ErgoAI</a>) software, an advanced tool for enterprise knowledge representation and reasoning (KRR). ErgoAI greatly extends the well-known open-source Flora-2 system (<a class="moz-txt-link-freetext" href="https://flora.sourceforge.net/">https://flora.sourceforge.net/</a>) and, with this release, Flora-2 as a distinct system is being retired. ErgoAI is a best-of-breed system for logical AI, that is based on fundamental research advances. It has been used to develop applications in a number of commercial and government organizations. It shines at decision analytics & automation tasks in financial, legal, defense, and other domains, such as compliance with policies, regulations, contracts, and guidance, where knowledge is complex, high accuracy is crucial, and good explainability is required. <br> <br> ErgoAI is highly composable with other AI and non-AI systems, and has connectors to major programming languages and data formats, including Python, knowledge graphs, and many others.<br> One can install ErgoAI most simply by using its automatic installers for the latest official stable release (<a class="moz-txt-link-freetext" href="https://github.com/ErgoAI/.github/releases">https://github.com/ErgoAI/.github/releases</a>). Or install the system and its source code by cloning the Github repositories and compiling the sources <a class="moz-txt-link-freetext" href="https://github.com/ErgoAI#installing-ergoai-from-the-sources">https://github.com/ErgoAI#installing-ergoai-from-the-sources</a>). ErgoAI comes with extensive documentation, tutorials, and examples. Please see <a class="moz-txt-link-freetext" href="https://github.com/ErgoAI">https://github.com/ErgoAI</a> for more details.<br> <br> Please join our discussion forum at <a class="moz-txt-link-freetext" href="https://groups.google.com/a/coherentknowledge.com/g/ergoai-xsb-forum">https://groups.google.com/a/coherentknowledge.com/g/ergoai-xsb-forum</a> and give ErgoAI a Github star (at <a class="moz-txt-link-freetext" href="https://github.com/ErgoAI/ErgoEngine">https://github.com/ErgoAI/ErgoEngine</a>).<br> <br> </p> </body> </html> |
From: Paulo M. <pm...@lo...> - 2023-04-27 15:24:17
|
Hi, Logtalk 3.65.0 is now available for downloading at: https://logtalk.org/ This release documents the no longer experimental "(@)/1" control construct; adds a new "since" key to the "info/2" directive; adds two new values for the number of proofs to the "mode/2" directive; improves the Handbook sections on debugging and documenting predicates; reduces the number of atoms created by some tools and libraries; fixes the "git" library when working with paths containing spaces; includes significant fixes and improvements to the "debugger" and "lgtunit" tools; improves tests reporting; includes fixes for the Java examples and testing automation on Windows; adds notes for most of the skipped tests; includes new and revised tests; and provides portability updates for LVM, Scryer Prolog, and, Trealla Prolog. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2022-12-20 16:41:13
|
Hi, Logtalk 3.61.0 is now available for downloading at: https://logtalk.org/ This release provides updated for several Prolog backend adapters and integration scripts for use in the Jupyter kernel for Logtalk (https://github.com/LogtalkDotOrg/logtalk-jupyter-kernel); adds a new "ids" library for generating identifiers (Base64 encoded) with parametrizable text representation and number of bytes of randomness; adds new "pairs" library object predicates; adds "packs" tool support for using "git archive" for installing packs, providing a better user experience when the packs registry server uses Single Sign-On (SSO) for authentication; and includes additional tests for library predicates and standard arithmetic functions. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2022-11-15 23:16:51
|
Hi, Logtalk 3.60.0 is now available for downloading at: https://logtalk.org/ This release adds a new linter check for the availability of predicates (and non-terminals) listed in `uses/2` and `use_module/2` directives; improves the readability of default flag values printed at startup; updates all backend adapter files to recognize `.pro` as a valid Prolog file name extension; includes Handbook, library, and tools documentation improvements; updates the manual installation instructions for `zsh` users; adds `json` library support for customizing the term representation of JSON objects and JSON pairs; adds new `reader::line_to_chars/2-3` library predicates; adds new and improved linter checks to the `lgtdoc` tool; adds a new `lgtunit` tool `file_path/2` predicate; includes fixes for the `logtalk_doclet` and `logtalk_tester` scripts; includes additional tests for several Prolog standard predicates; updated the `logtalk_user_setup` scripts to prevent misuse; and adds support for escape sequences inside double-quoted terms to most text editors and syntax highlighters. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2022-10-18 14:13:36
|
Hi, Logtalk 3.59.0 is now available for downloading at: https://logtalk.org/ This release adds a new linter warning for naked meta-variables; adds linter warnings to the "lgtdoc" tool; updates the Bash shell integration scripts to be aware of the default Homebrew installation directory on macOS; updates all PowerShell scripts to check for a compatible PowerShell version; improves several Handbook sections; improves the documentation of the "random" and "types" libraries; includes fixes for the "logtalk_tester.ps1" script; updates the "lgtunit" tool to print a warning on invalid test options; fixes "lgtdoc" issues; updates and adds new Prolog standards compliance tests; adds new library predicates; and includes portability updates for Ciao Prolog and SICStus Prolog. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2022-09-27 09:40:17
|
Hi, Logtalk 3.58.0 is now available for downloading at: https://logtalk.org/ This release adds a new linter warning for non-terminals called as predicates; adds a compilation warning when compiling a "logtalk_load_context/2" goal with an invalid key; fixes the compiler to not throw an instantiation error when the body of a clause is a variable; improves documentation on third-party dependencies to build documentation final formats and to use the developer tools; improves Handbook, libraries, and tools documentation; provides fixes and improvements for the "os", "types", and "arbitrary" libraries; includes a new "lgtunit" version with improved QuickCheck support; adds experimental predicates to the "help" tool to browse and search inline at the top-level the Texinfo versions of the Handbook and APIs documentation; fixes a missing message tokenization in the "packs" tool; adds new Net-Clause Language (NCL) examples; adds new compliance tests for standard Prolog predicates and operators; adds a port of FCube, a prover for intuitionistic propositional logic; adds Emacs code snippets, contributed by Paul Brown; provides an embedding script for Ciao Prolog and improves the documentation of the embedding scripts; provides a PowerShell script for building the Handbook; updates the release build script to include all final formats for the Handbook and APIs documentation in the generated archives and installers; and includes portability updates for LVM, Trealla Prolog, and YAP. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Paulo M. <pm...@lo...> - 2022-07-26 15:47:45
|
Hi, Logtalk 3.57.0 is now available for downloading at: https://logtalk.org/ This release improves the linter tool with new and refined warnings; improves the compilation of Prolog modules as objects; refines compiler error reporting; improves support for making predicates listed in "uses/2" and "use_module/2" directives available as scoped predicates for the entities containing the directives; adds Handbook sections on documenting predicate exceptions and on ports and contributions distributed with Logtalk; improves Handbook sections on Prolog integration and migration; adds links from library APIs documentation to the Handbook library descriptions; adds a "genint" library supporting application global and object local named generators for increasing positive integers; updsates the "gensym" library with support for both global and object local symbol generators; adds new predicatyes to the "types" and "statistics" libraries; includes fixes and improvements to the "csv" library; adds new parametric objects to the "hook_objects" library; fixes the "arbitrary" library implementation of shrinkers when passed terms of unexpected types; includes an improved "packs" tool with new predicates, refined semantics, and support for registries that require personal token authorizations for access; fixes "lgtdoc" generated Sphinx documentation cross-referencing between applications and Logtalk distribution libraries; improves "tutor" tool feedback on sone common linter reports; fixes "diagrams" tool cases where files listed as excluded would still be included in generated diagrams; adds new tests for Logtalk and Prolog features; and includes portability updates for CxProlog, LVM, SICStus Prolog, SWI-Prolog, and XSB. For details and a complete list of changes, please consult the release notes at: https://github.com/LogtalkDotOrg/logtalk3/blob/master/RELEASE_NOTES.md You can show your support for Logtalk continued development and success at GitHub by giving us a star and a symbolic sponsorship: https://github.com/LogtalkDotOrg/logtalk3 Happy logtalking! Paulo ----------------------------------------------------------------- Paulo Moura Logtalk developer |
From: Massimo Z. <mas...@ti...> - 2001-03-29 21:46:21
|
Hi, http://jude.sourceforge.net Up to date this isn't a working system :-( If you are interested read the detailed documentation and give me some feedback. In particular you should read this chapter: http://jude.sourceforge.net/jude/node5.html Frankly speaking the mechanism that add active-database features to XSB isn't full implemented and tested and I believe that it should be too slow for real-world systems. Sorry for my bad English. On Thursday 29 March 2001 22:31, you wrote: > I would like to find out if there are any projects that use XSB as part of > an Agent Infrastructure. Would you know of any, or be able to point me in > the right direction? > > Thanks, > Bill > > _______________________________________________ > Xsb-users mailing list > Xsb...@li... > http://lists.sourceforge.net/lists/listinfo/xsb-users Best regards Massimo. |
From: <wil...@lm...> - 2001-03-29 20:31:30
|
I would like to find out if there are any projects that use XSB as part of an Agent Infrastructure. Would you know of any, or be able to point me in the right direction? Thanks, Bill |
From: LECONTE Jean-m. <web...@la...> - 2001-03-28 17:08:51
|
But where is it in the XSB help ?? I was thinking constraints like in Gnu prolog, where you define the domaine of X (an example) =2E=2E=2E=2E etc Thanks Jean michel LECONTE ----- Message d'origine ----- De : "David S=2E Warren" <warren@cs=2Esunysb=2Eedu> =C0 : "LECONTE Jean-michel" <webmaster@lambda-systems=2Ecom> Cc : <xsb-users@lists=2Esourceforge=2Enet> Envoy=E9 : mardi 27 mars 2001 21:38 Objet : [Xsb-users] Constraints in XSB ? > LECONTE Jean-michel writes: > > Are constraints going to be implemented in XSB ?? > > Actually, XSB already supports real constraints (partially)=2E What did > you have in mind? > > Regards, > -David |
From: Randall R S. <rrs...@cr...> - 2001-03-27 22:22:44
|
Michael, If you don't know about it already, the bit masking and shifting that implements memory cell tags (in cell_xsb.h) is sensitive to processor type, OS type, and possibly even the OS port or version. I made this work (mostly by trial and error) for PowerPC Linux running on a Macintosh G3. I never got around to submitting the changes back to the XSB project (and haven't used that version of XSB for some time now (it's XSB 2.1 probably--but I'm sure those changes would still be valid). If you'd like, I can dig it up and send it to you. As I said, because MacOS X is a FreeBSD port (with a microkernel, no less) and LinuxPPC where I did the XSB port, uses a monolithic Linux kernel, there's a good chance that they used different address ranges and / or bit field assignments for the kernel, shared libraries, stack, initialized and unitialized data, shared memory segments, etc., so thus it would be fairly lucky if the LinuxPPC parameters I derived will work under MacOS X. RandyS At 20:53 3/27/2001, Michael Kifer wrote: >Mac OSX is too new, I suppose, so the configure script doesn't know about it. >We'll check if the latest version of configure knows something about OSX. >However, there might be other issues related to porting, and for this we >need help from the users who have access to this system. > >You can try to specify generic-redhat-linux and see what happens. >Most likely won't work, but ... > > > Just downloaded XSB 2.3 and I was wondering if it is possible to > > run this on Mac OSX. I ran ./configure from the build directory and the > > respnse is: > > checking host system type... configure: error: can not guess host type; > > you must specify one > > > > What type of UNIX host does Mac OSX use? It is very possible I'm just > > wasting my time, but it was worth a try. If XSB Prolog won't work on > > OSX's kernel, do you know if a port for XSB to OSX is planned for the > > future? I'm sure it wouldn't be that hard :) > > > > Thanks for your time, > > Ben Voigt > > > > _______________________________________________ > > Xsb-users mailing list > > Xsb...@li... > > http://lists.sourceforge.net/lists/listinfo/xsb-users > > > >_______________________________________________ >Xsb-users mailing list >Xsb...@li... >http://lists.sourceforge.net/lists/listinfo/xsb-users |
From: David S. W. <wa...@cs...> - 2001-03-27 21:39:48
|
LECONTE Jean-michel writes: > Are constraints going to be implemented in XSB ?? Actually, XSB already supports real constraints (partially). What did you have in mind? Regards, -David |
From: David S. W. <wa...@cs...> - 2001-03-27 21:36:14
|
Joe, Can you send your interface to this person (email above)? Jean-michel, you need to have the most recent release of XSB to run it, so please download and install it if you haven't already. Regards, David LECONTE Jean-michel writes: > can i have the good code to call XSB from a vb program cause i have Visual > Basic 6.0 and when i call the program , tht's crashing ! > > > Jean Michel LECONTE > > > _______________________________________________ > Xsb-users mailing list > Xsb...@li... > http://lists.sourceforge.net/lists/listinfo/xsb-users > |
From: LECONTE Jean-m. <web...@la...> - 2001-03-27 21:06:17
|
Are constraints going to be implemented in XSB ?? Jean michel LECONTE _ / \ / \ http://www.lambda-systems.com / \ / \_/ SYSTEMS |
From: LECONTE Jean-m. <web...@la...> - 2001-03-27 21:03:36
|
can i have the good code to call XSB from a vb program cause i have Visual Basic 6.0 and when i call the program , tht's crashing ! Jean Michel LECONTE |
From: <ki...@cs...> - 2001-03-27 20:59:48
|
Mac OSX is too new, I suppose, so the configure script doesn't know about it. We'll check if the latest version of configure knows something about OSX. However, there might be other issues related to porting, and for this we need help from the users who have access to this system. You can try to specify generic-redhat-linux and see what happens. Most likely won't work, but ... > Just downloaded XSB 2.3 and I was wondering if it is possible to > run this on Mac OSX. I ran ./configure from the build directory and the > respnse is: > checking host system type... configure: error: can not guess host type; > you must specify one > > What type of UNIX host does Mac OSX use? It is very possible I'm just > wasting my time, but it was worth a try. If XSB Prolog won't work on > OSX's kernel, do you know if a port for XSB to OSX is planned for the > future? I'm sure it wouldn't be that hard :) > > Thanks for your time, > Ben Voigt > > _______________________________________________ > Xsb-users mailing list > Xsb...@li... > http://lists.sourceforge.net/lists/listinfo/xsb-users |
From: Ben V. <br...@ho...> - 2001-03-27 20:53:20
|
Just downloaded XSB 2.3 and I was wondering if it is possible to run this on Mac OSX. I ran ./configure from the build directory and the respnse is: checking host system type... configure: error: can not guess host type; you must specify one What type of UNIX host does Mac OSX use? It is very possible I'm just wasting my time, but it was worth a try. If XSB Prolog won't work on OSX's kernel, do you know if a port for XSB to OSX is planned for the future? I'm sure it wouldn't be that hard :) Thanks for your time, Ben Voigt |
From: <ts...@cs...> - 2001-03-22 18:47:11
|
Josh: > I've rewritten my code to skip the check for large expressions, but I'm > concerned that I may run into problems with big terms later. Have these > predicates moved elsewhere? Or can I now trust the compiler to compile > arbitrarily large terms. David: > No, the compiler hasn't changed (much).. Well, its still pretty lousy for big terms, but last September I integrated a new flattening algorithm of Kostis into the compiler. So we're a little better, though our register allocation is still quadratic for large lists (including strings). |
From: David S. W. <wa...@cs...> - 2001-03-22 17:57:57
|
Joshua Engel writes: > Going through some of my old code, I found the line: > > :- import termsize/2 from copyterm. I don't know what happened to termsize. I guess it left with copyterm, which left when copy_term was moved down into C for performance reasons. If you still need a termsize, it's easy to write in Prolog. > termsize/2 (and copyterm) no longer seem to exist. I had used it to get > around a limitation in the XSB compiler. When the termsize was large, I > used bts/2 to evaluate the complicated term. (bts/2 doesn't seem to exist > any more, either.) These changes had originally been suggested to us by > Dr. Warren, for an older version of XSB (possibly as old as 1.8). I don't recall what bts/2 did. (bts for Big_Term_Size? Was it a Prolog predicate that used assert instead of the compiler somehow?) > I've rewritten my code to skip the check for large expressions, but I'm > concerned that I may run into problems with big terms later. Have these > predicates moved elsewhere? Or can I now trust the compiler to compile > arbitrarily large terms. No, the compiler hasn't changed (much).. > [Context: the code here automatically generates auxilliary rules for > queries involving complicated tnot/1 bodies. The complex subquery is > asserted as a rule, allowing me to use an arbitrarily complicated query > with a tnot/1. But really complex assertions were giving me grief at some > point in the past, so Dr. Warren suggested this solution involving checking > the term size and using bts/2.] Oh, I guess you are talking about assert (which I don't usually call ``compiling'', which threw me off.) Assert used to run out of registers because it didn't reclaim them very well. So, for example, asserting a long list would run out of registers. But Bart changed the way registers were released and so now we don't run out of registers in assert anywhere near as much as we used to. So, if this was the problem, I think you're pretty safe. -David |
From: Joshua E. <en...@er...> - 2001-03-22 16:20:35
|
Going through some of my old code, I found the line: :- import termsize/2 from copyterm. termsize/2 (and copyterm) no longer seem to exist. I had used it to get around a limitation in the XSB compiler. When the termsize was large, I used bts/2 to evaluate the complicated term. (bts/2 doesn't seem to exist any more, either.) These changes had originally been suggested to us by Dr. Warren, for an older version of XSB (possibly as old as 1.8). I've rewritten my code to skip the check for large expressions, but I'm concerned that I may run into problems with big terms later. Have these predicates moved elsewhere? Or can I now trust the compiler to compile arbitrarily large terms. [Context: the code here automatically generates auxilliary rules for queries involving complicated tnot/1 bodies. The complex subquery is asserted as a rule, allowing me to use an arbitrarily complicated query with a tnot/1. But really complex assertions were giving me grief at some point in the past, so Dr. Warren suggested this solution involving checking the term size and using bts/2.] -- Joshua |
From: <ts...@cs...> - 2001-03-22 15:51:35
|
Hi Michael: Glad to see you are working with XSB. I think that a partial evaluator that worked with tabled code would be a great benefit to the XSB community. I'm not sure if you're thinking along these lines, but I thought I'd put in my two cents :-) The semantics of tnot is restrictive in several ways. First, as you note, only a single tabled goal can be used as an argument to tnot; second the goal must be ground. Thus, tnot does not yet handle goals analagous to the Prolog goal not(p(X),g(X)) which can be perfectly sound if X is not shared with any variable outside the not. sk_not/1 solves most of this by allowing users to specify atomic, non-ground goals that may or may not be tabled -- but it doesnt handle multiple goals as above -- though perhaps we should extend it to do this. (Let me know if handling non-atomic goals is important to you). The bigger problem is that we dont currently allow cuts over a tabled subgoal, so you cant call a tabled predicate in the scope of a Prolog not or \+. The reason for this is that solutions for a tabled subgoal must be global to a computation, while the cut is a local mechanism and our engine is not currently set up to determine that answers can be sent back to certain calling contexts but not to others (right now, this is partially addressed by the predicate table_once/1, but deeper engine changes are underway). In the case where you could use neither not nor tnot, it may have been where a Prolog goal was calling a tabled subgoal? Thus you couldnt use tnot because the goal wasnt tabled and couldnt use not because it cut over a tabled predicate? In principle, sk_not should be general for atomic goals. You can then think of tnot as an optimization for sk_not where the goal is already tabled and ground; Prolog not is an optimization for when neither the goal nor any subgoals it calls is tabled. I couldn't reproduce the problem with sk_not being undefined that you describe (as shown below). Perhaps the bug was fixed in the new version? (now available from the web site) Best regards Terry Swift ------------------------------------------------------------------- xsbcommon [xsb_configuration loaded] [sysinitrc loaded] [packaging loaded] [xsbrc loaded] XSB Version 2.3 (Zombie) of February 16, 2001 [i686-pc-linux-gnu; mode: optimal; engine: chat; gc: copy; scheduling: batched] | ?- assert((p(X):- fail)). X = _h94 yes | ?- sk_not(p(X)). X = _h90 yes ------------------------------------------------------------------- |
From: Michael L. <ma...@ec...> - 2001-03-22 14:29:19
|
Hi, I am currently using XSB Prolog and have a few problems with negation. The manual mentions a predicate sk_not which is like tnot but does not do the floundering check. However, XSB 2.2 tells me that sk_not is an undefined predicate. What do I have to do ? sk_not is what I need as I want to run XSB on automatically generated programs (generated by partial evaluation which can sometimes introduce extra arguments). Actually, it would be even better if there was a universal not which works on both tabled and untabled predicates (as sometimes unfolding can change a tabled into an untabled call and vice-versa !). In fact, I also had a case where I could neither use not nor tnot: tnot complained because the predicate was not tabled and not complained because the predicate depended on a tabled predicate (the original program worked fine though). Does a universal not exist, i.e., a not that will always work ? Thanks, Best, Michael Leuschel ============================================================ Michael Leuschel (http://www.ecs.soton.ac.uk/~mal) Declarative Systems and Software Engineering Department of Electronics and Computer Science University of Southampton Highfield, Southampton, SO17 1BJ, U.K. Fax: +44 2380 59 3045 Tel: +44 2380 59 3377 Mobile: +44 7879 497157 E-mail: ma...@ec... ============================================================ |