Hi Aaron. I have now tested all my programs and most of them worked as they did in previous versions, See below for comments and reports.
This is a nice diversion from the Deep Learning course I take now. I'm much more comfortable with GOFAI. :-) On Wed, Jan 1, 2020 at 12:31 PM Aaron Sloman <[email protected]> wrote: > Hi Hakan, > > Thanks very much for testing the getpoplog.sh script so soon! > > Your success is a great reward for my attempts to understand what I was > trying to do -- with several intermediate messes. > > > What a great start of the new year, Now I got Poplog 16 to work! > > If I had known several months ago that it would take till now to reach this > stage I might have suggested calling it Poplog 20! > > But there are now too many references to V16. > > > The magic that got it to work on my Linux Ubuntu 18.04 box was the -nopie > > flag to getpoplog.sh . Excellent! > > It was all due to Waldek, who mentioned the problem AND the solution in > June. > > Thanks, Waldek! I saw the discussions earlier, but it wasn't you packaged the getpoplog.sh (and instructions) in this nice way, that I succeeded to install it properly. > Six months later, I think I understand what he had written, well enough to > use the information in the new download script, after several false starts! > > (I am not really a linux programmer.) > > > There are still some issues that I will check out in more detail, but > > running all the first 50 Project Euler programs is now 2s faster than the > > earlier 15.65 version. (See http://hakank.org/poplog/ for the programs.) > > I had previously referenced your pages in > http://www.cs.bham.ac.uk/research/projects/poplog/freepoplog.html > > I thought I should also include something in the new (still draft) V16 > web site: > http://www.cs.bham.ac.uk/research/projects/poplog/V16/AREADME.html > > Thanks for that. You can also add that Poplog16 also works on Ubuntu 18.04. Also, I've updated the files at http://hakank.org/poplog/ (e.g. some of the Euler programs). The public Pop-11 files are also available at my GitHub page: https://github.com/hakank/hakank/tree/master/poplog However I have a problem, which others may have also. If a file name ends > in '.p' Firefox refuses to display it, and merely asks where to save a > downloaded version. (It also does this with .sh files.) > > Does anyone know how to disable that behaviour? > > Curiously opera displays .p files, but refuses to display .sh files, and > only offers to download them. > > Yes. I had the same problem. There might be alternative ways to fix this. I added .p as a text/plain in my web server configuration, and now it works for me, i.e. the Pop11 program is shown inline instead of via download. > The universe has got too complicated for my simple mind, with too many > paths being traversed in parallel... > > Best wishes > Aaron > http://www.cs.bham.ac.uk/~axs > Here are some further comments: - super_test.p (test of super.p) http://hakank.org/poplog/super_test.p The file poplog_base/pop/lib/lib/super.p is the old one from 1989. In 2009 I reported error in this file and you fixed it in September 2009 and it is included in the versions after that, e.g. v 15.65. The new file is available here: http://hakank.org/poplog/super.p - Problem with interactive expert systems Note: this is not a new problem in v16. It has been like this since I can remember but it always annoy me to it's time to report it. Here are the programs that's affected: * http://hakank.org/poplog/emycin_test.p * http://hakank.org/poplog/eprospect_test.* p * http://hakank.org/poplog/eshell_test.p * http://hakank.org/poplog/id3_test.p * http://hakank.org/poplog/poprulebase_medrules.p * http://hakank.org/poplog/poprulebase_toy_eliza.p They are mostly tests of examples from the nice TEACH files. These interactive expert system programs has all the same problem, namely that running them direct via $ pop11 program.p they either give some strange errors or just prints in an infinite loop. For example $ pop11 emycin_test.p throws this error: ;;; MISHAP - NON MATCHING ARGUMENTS FOR --> ;;; INVOLVING: <termin> [? val] ;;; FILE : /home/hakank/poplog/me/emycin_test.p LINE NUMBER: 40 ;;; PRINT DOING ;;; DOING : --> ask findvalueof(*2) emycin trycompile However, it works if the program is loaded from the pop11 shell: $ rlwrap pop11 $ load 'emycin_test.p' ;;; The same error as above But now it works to run: emycin("species",system); It seems that the problem is that pop11 don't handle termin etc correctly when loading the program via command line. In some cases, the program prints in an infinite loop that one first has to stop with Ctrl-C. This is the case for these programs - eshell_test.p - poprulebase_medrules.p - poprulebase_toy_eliza.p Note: Some of the program might rely on some GOSPL function, which I load like this compile( '/home/hakank/poplog/gospl/init.p' ) in my ~/Poplog/init.p file (see http://hakank.org/poplog/init.p for the full file) - ved: Unknown terminal type: xterm-256color In poplog16, when ved is started, the following error is thrown: """ Unknown terminal type: xterm-256color Enter an alternative terminal type, or <RETURN> to abandon VED: """ Typing xterm<RETURN> then start ved without any problem. Ah, I just found out a solution of this: * Add the following line to ~/Poplib/init.p and it works:. 'xterm'->systranslate('TERM'); Best regards, Hakan -- Hakan Kjellerstrand http://www.hakank.org/ http://www.hakank.org/webblogg/ http://www.hakank.org/constraint_programming_blog/ http://twitter.com/hakankj https://www.facebook.com/hakankj
