Toki Pona and CopyRight : an IRC geek discussion about its possible implications
Time Nick Message
00:13 jeffreykegler muvets: good luck. til Monday is not much time.
00:14 jeffreykegler I've put together a youtube playlist of Marpa videos -- there's 4, but one is an interview and one is in Russian.
00:15 jeffreykegler Of the talks, the one by Stephen Flitman is most recent, https://www.youtube.com/watch?v=t1SjsfAQiSg&list=PLrSLalZGd6yLq_CqWkbti8m18iL0d4hqE&index=2
00:15 jeffreykegler (most recent in English, that is)
00:16 jeffreykegler I've yet to listen to it through a 2nd time, but Dr. Flitman goes throw things very slowly, and at a very basic level --
00:21 jeffreykegler BNF-driven parsing, once better known than regular expressions, has fallen off the radar, so explaining Marpa become much like, say, explaining a checkbook balancing app to an audience that does not know how to balance a checkbook or why one would want to.
00:22 jeffreykegler Much as I *love* to see folks give talks on Marpa, if you are starting from scratch, you might want to talk about your project, about toki pona, and say about Marpa ...
00:24 jeffreykegler "and I'm planning to use Marpa which looks really cool" and a few remarks about what you hope to get out of it, and some random thoughts from what you've learned by Monday.
06:07 mvuets jeffreykegler: thanks for the playlist! i was able to find these videos individually before, but now i see it's all that we have so far, too sad
06:08 mvuets thanks for the advise, i'll follow it. technically my talk isn't on Marpa. Marpa is an excuse to give a non-tech talk in a tech meetup (-:
06:09 mvuets if everything works out well, i'll be giving a similar talk in a workshop, but using Perl 6 built-in grammars instead of Marpa
16:17 jeffreykegler muvets: note that perl6 built-in grammars are essentially top-down, in different packaging
16:17 jeffreykegler That is, they're a kind of semi-automated recursive descent.
16:18 jeffreykegler Typically, top-down cannot handle NLP grammars.
16:23 jeffreykegler Also, is Toki Pona not "open source"?
16:25 jeffreykegler The only thing mentioned on their website is a book which you have to buy: http://tokipona.org/
16:26 jeffreykegler It's OK for the full book to be behind a paywall, but there seems to be no "open source" reference doc.
17:55 mvuets jeffreykegler: the website had been rather abandoned for a long time. there are many good open resources elsewhere. wikibooks has a decent bunch of lessons
17:56 mvuets here are some references - https://www.reddit.com/r/tokipona/comments/3dfwes/duolingo_toki_pona_or_some_place_where_i_can/
17:56 jeffreykegler Well, in particular, I was looking for an authoritative grammar
17:57 jeffreykegler I can find on-line lessons about Oracle, but that doesn't make it open-source.
17:58 mvuets i might be wrong, but i am afraid authoritative grammar is only available in that book, yes
17:58 mvuets not sure what was the state on a moment when the language was release into the public though
17:58 mvuets perhaps the site had official grammar and dictionary
17:59 mvuets and here is a paper on the formal grammar that i mentioned earlier - http://www2.hawaii.edu/~chin/661F12/Projects/ztomaszewski.pdf
18:01 jeffreykegler Just to be clear, Sonja has every right to sell a book about Toki Pona, and TP would still be open source ...
18:02 jeffreykegler Roberto, for example, has a Lua book which is not free and not on-line.
18:02 mvuets sure, i see your point. i think you are saying that so far she is _only_ selling a book
18:03 jeffreykegler But his *reference docs* are on-line and open source and this would be a problem if it is not the case with Toki Pona
18:03 mvuets you raised a good issue, i'll keep it in mind and try to investigate it when have an opportunity
18:04 jeffreykegler The Tomaszewski article or its references may contain a grammar. I'll look into it later.
18:06 mvuets it does, it's what i'll by trying to run with Marpa
18:07 jeffreykegler I note in skimming that the Toki Pona grammar is ambiguous, which means Marpa may be just about the only thing out there that will handle TP
18:08 mvuets hehe, that's even better!
18:09 mvuets looking into the web.archive.org i reckon TP just has never been properly "open sourced"
18:09 mvuets http://web.archive.org/web/20110813111549/http://en.tokipona.org/wiki/Category:Grammar
18:09 mvuets http://web.archive.org/web/20110728120132/http://en.tokipona.org/wiki/Category:Toki_Pona_words
18:10 mvuets some attempts, but not complete and not comprehensive
18:10 jeffreykegler I was into a language called Loglan a while ago, and the creator retroactively slapped a copyright on it and closed it.
18:11 jeffreykegler Loglan had actually gotten a Scientific American article, back when SciAm was very prestigious.
18:11 jeffreykegler Anyway, I don't believe he ever made a dime from this move, but it meant a lot of folks had wasted their time.
18:12 jeffreykegler That experience makes me cautious here.
18:12 mvuets oh. and what does it mean to copyright *a language*?
18:13 jeffreykegler In fact, I believe under common law, languages are one of the things that explicitly cannot be copyrighted ...
18:14 jeffreykegler but the practical effect if that you'd be open to expensive litigation.
18:15 jeffreykegler And I can't imagine anyone wanting to learn a copyrighted language.
18:16 jeffreykegler The Loglan effort forked over that by the way, and Lobjan is the open source fork.
18:17 mvuets i see! now i get the Loglan vs. Lobjan difference
18:21 mvuets i think i got a partial answer to your question. i own a hard copy of the official book and it actually has the following copyright notice
18:22 mvuets pp. 1--22 and pp. 91--124 are under CC-BY-NC 4.0
18:23 mvuets weird that first range covers 3 of 19 lessons only
18:24 mvuets second range is misc. dictionaries (phrase book, sign language, etc.)
18:24 jeffreykegler I guess that means that those pages could be scanned and put on-line
18:24 mvuets and the Official Toki Pona dictionary is "in the public domain".
18:25 jeffreykegler The "non-commercial" is a problem also, but that's another topic.
18:26 mvuets i presume the rest that's not mentioned is under "regular copyright law"? there is a "© 2014 Sonja Lang"
18:26 jeffreykegler All that she has to open source is the material necessary for reference.
18:28 jeffreykegler Re the non-commercial -- it is unlikely to stop the mega-corporations from doing whatever they like but it will stop small businesses, consultants, etc. ...
18:28 jeffreykegler so the effect is likely to be the opposite of the real intention.
18:28 mvuets sigh
18:28 jeffreykegler And it is also a fact that useful stuff comes via for-profit megacorporations.
18:29 jeffreykegler I don't know how I'd live without Google search
18:29 mvuets i am glad to find out that the dictionary is in public domain, yet it does not suffice as there is no "open" grammar reference, as you also pointed out
18:29 jeffreykegler and I certainly find Github useful.
18:30 jeffreykegler Sonja may not be well versed in these issues -- those of us in the software realm have been forced to learn this stuff the hard way ...
18:30 jeffreykegler and are still learning.
18:31 jeffreykegler For example, I've been over 4 decades in the computer business, but was just educated by the Marpa community re the MIT license ...
18:32 jeffreykegler which Marpa is changing over to.
18:32 mvuets interesting! i don't know much about legal issues like this. sadly for me
18:32 jeffreykegler or happily for you :-)
18:33 mvuets could be (-: how is Marpa licenced now?
18:34 jeffreykegler A mix, with the trend towards MIT. For details see the docs.
20:02 mvuets does the position of a pseudorule :discard within a grammar description matter?
20:16 pczarn mvuets: the position in relation to other rules? no, doesn't matter
20:21 pczarn mvuets: The set of rules is unordered. Choose any ordering, but :discard is usually placed near the end
20:24 mvuets pczarn: makes sense, thanks
21:00 pczarn I have an idea: complete Earley sets can be sorted by postdot nonterminal symbols or their absence (before memoization)
21:01 pczarn I'd like to ask if this is worth implementing and optimizing
23:59 jeffreykegler pczarn: re http://irclog.perlgeek.de/marpa/2015-07-21#i_10933923 -- almost everyone does index entries by postdot symbol -- Marpa certainly does.
23:59 jeffreykegler And IIRC Earley suggests it in his original paper.
https://irclog.perlgeek.de/marpa/2015-07-21/text
#TokiPona #mention #sona #anno2015
Time Nick Message
00:13 jeffreykegler muvets: good luck. til Monday is not much time.
00:14 jeffreykegler I've put together a youtube playlist of Marpa videos -- there's 4, but one is an interview and one is in Russian.
00:15 jeffreykegler Of the talks, the one by Stephen Flitman is most recent, https://www.youtube.com/watch?v=t1SjsfAQiSg&list=PLrSLalZGd6yLq_CqWkbti8m18iL0d4hqE&index=2
00:15 jeffreykegler (most recent in English, that is)
00:16 jeffreykegler I've yet to listen to it through a 2nd time, but Dr. Flitman goes throw things very slowly, and at a very basic level --
00:21 jeffreykegler BNF-driven parsing, once better known than regular expressions, has fallen off the radar, so explaining Marpa become much like, say, explaining a checkbook balancing app to an audience that does not know how to balance a checkbook or why one would want to.
00:22 jeffreykegler Much as I *love* to see folks give talks on Marpa, if you are starting from scratch, you might want to talk about your project, about toki pona, and say about Marpa ...
00:24 jeffreykegler "and I'm planning to use Marpa which looks really cool" and a few remarks about what you hope to get out of it, and some random thoughts from what you've learned by Monday.
06:07 mvuets jeffreykegler: thanks for the playlist! i was able to find these videos individually before, but now i see it's all that we have so far, too sad
06:08 mvuets thanks for the advise, i'll follow it. technically my talk isn't on Marpa. Marpa is an excuse to give a non-tech talk in a tech meetup (-:
06:09 mvuets if everything works out well, i'll be giving a similar talk in a workshop, but using Perl 6 built-in grammars instead of Marpa
16:17 jeffreykegler muvets: note that perl6 built-in grammars are essentially top-down, in different packaging
16:17 jeffreykegler That is, they're a kind of semi-automated recursive descent.
16:18 jeffreykegler Typically, top-down cannot handle NLP grammars.
16:23 jeffreykegler Also, is Toki Pona not "open source"?
16:25 jeffreykegler The only thing mentioned on their website is a book which you have to buy: http://tokipona.org/
16:26 jeffreykegler It's OK for the full book to be behind a paywall, but there seems to be no "open source" reference doc.
17:55 mvuets jeffreykegler: the website had been rather abandoned for a long time. there are many good open resources elsewhere. wikibooks has a decent bunch of lessons
17:56 mvuets here are some references - https://www.reddit.com/r/tokipona/comments/3dfwes/duolingo_toki_pona_or_some_place_where_i_can/
17:56 jeffreykegler Well, in particular, I was looking for an authoritative grammar
17:57 jeffreykegler I can find on-line lessons about Oracle, but that doesn't make it open-source.
17:58 mvuets i might be wrong, but i am afraid authoritative grammar is only available in that book, yes
17:58 mvuets not sure what was the state on a moment when the language was release into the public though
17:58 mvuets perhaps the site had official grammar and dictionary
17:59 mvuets and here is a paper on the formal grammar that i mentioned earlier - http://www2.hawaii.edu/~chin/661F12/Projects/ztomaszewski.pdf
18:01 jeffreykegler Just to be clear, Sonja has every right to sell a book about Toki Pona, and TP would still be open source ...
18:02 jeffreykegler Roberto, for example, has a Lua book which is not free and not on-line.
18:02 mvuets sure, i see your point. i think you are saying that so far she is _only_ selling a book
18:03 jeffreykegler But his *reference docs* are on-line and open source and this would be a problem if it is not the case with Toki Pona
18:03 mvuets you raised a good issue, i'll keep it in mind and try to investigate it when have an opportunity
18:04 jeffreykegler The Tomaszewski article or its references may contain a grammar. I'll look into it later.
18:06 mvuets it does, it's what i'll by trying to run with Marpa
18:07 jeffreykegler I note in skimming that the Toki Pona grammar is ambiguous, which means Marpa may be just about the only thing out there that will handle TP
18:08 mvuets hehe, that's even better!
18:09 mvuets looking into the web.archive.org i reckon TP just has never been properly "open sourced"
18:09 mvuets http://web.archive.org/web/20110813111549/http://en.tokipona.org/wiki/Category:Grammar
18:09 mvuets http://web.archive.org/web/20110728120132/http://en.tokipona.org/wiki/Category:Toki_Pona_words
18:10 mvuets some attempts, but not complete and not comprehensive
18:10 jeffreykegler I was into a language called Loglan a while ago, and the creator retroactively slapped a copyright on it and closed it.
18:11 jeffreykegler Loglan had actually gotten a Scientific American article, back when SciAm was very prestigious.
18:11 jeffreykegler Anyway, I don't believe he ever made a dime from this move, but it meant a lot of folks had wasted their time.
18:12 jeffreykegler That experience makes me cautious here.
18:12 mvuets oh. and what does it mean to copyright *a language*?
18:13 jeffreykegler In fact, I believe under common law, languages are one of the things that explicitly cannot be copyrighted ...
18:14 jeffreykegler but the practical effect if that you'd be open to expensive litigation.
18:15 jeffreykegler And I can't imagine anyone wanting to learn a copyrighted language.
18:16 jeffreykegler The Loglan effort forked over that by the way, and Lobjan is the open source fork.
18:17 mvuets i see! now i get the Loglan vs. Lobjan difference
18:21 mvuets i think i got a partial answer to your question. i own a hard copy of the official book and it actually has the following copyright notice
18:22 mvuets pp. 1--22 and pp. 91--124 are under CC-BY-NC 4.0
18:23 mvuets weird that first range covers 3 of 19 lessons only
18:24 mvuets second range is misc. dictionaries (phrase book, sign language, etc.)
18:24 jeffreykegler I guess that means that those pages could be scanned and put on-line
18:24 mvuets and the Official Toki Pona dictionary is "in the public domain".
18:25 jeffreykegler The "non-commercial" is a problem also, but that's another topic.
18:26 mvuets i presume the rest that's not mentioned is under "regular copyright law"? there is a "© 2014 Sonja Lang"
18:26 jeffreykegler All that she has to open source is the material necessary for reference.
18:28 jeffreykegler Re the non-commercial -- it is unlikely to stop the mega-corporations from doing whatever they like but it will stop small businesses, consultants, etc. ...
18:28 jeffreykegler so the effect is likely to be the opposite of the real intention.
18:28 mvuets sigh
18:28 jeffreykegler And it is also a fact that useful stuff comes via for-profit megacorporations.
18:29 jeffreykegler I don't know how I'd live without Google search
18:29 mvuets i am glad to find out that the dictionary is in public domain, yet it does not suffice as there is no "open" grammar reference, as you also pointed out
18:29 jeffreykegler and I certainly find Github useful.
18:30 jeffreykegler Sonja may not be well versed in these issues -- those of us in the software realm have been forced to learn this stuff the hard way ...
18:30 jeffreykegler and are still learning.
18:31 jeffreykegler For example, I've been over 4 decades in the computer business, but was just educated by the Marpa community re the MIT license ...
18:32 jeffreykegler which Marpa is changing over to.
18:32 mvuets interesting! i don't know much about legal issues like this. sadly for me
18:32 jeffreykegler or happily for you :-)
18:33 mvuets could be (-: how is Marpa licenced now?
18:34 jeffreykegler A mix, with the trend towards MIT. For details see the docs.
20:02 mvuets does the position of a pseudorule :discard within a grammar description matter?
20:16 pczarn mvuets: the position in relation to other rules? no, doesn't matter
20:21 pczarn mvuets: The set of rules is unordered. Choose any ordering, but :discard is usually placed near the end
20:24 mvuets pczarn: makes sense, thanks
21:00 pczarn I have an idea: complete Earley sets can be sorted by postdot nonterminal symbols or their absence (before memoization)
21:01 pczarn I'd like to ask if this is worth implementing and optimizing
23:59 jeffreykegler pczarn: re http://irclog.perlgeek.de/marpa/2015-07-21#i_10933923 -- almost everyone does index entries by postdot symbol -- Marpa certainly does.
23:59 jeffreykegler And IIRC Earley suggests it in his original paper.
https://irclog.perlgeek.de/marpa/2015-07-21/text
#TokiPona #mention #sona #anno2015