Sentence collector copyright issues

mkohler · July 6, 2021, 11:27pm

Thanks for reporting, these sentences have been removed.

phopo2 · August 7, 2021, 2:27am

Chinese - China (zh-cn):
Many sentences are from https://zh.wikisource.org/wiki/国家监委调查组负责人答记者问, which is from a Chinese government press conference transcript. It was removed from Wikisource for copyright violation.

Many sentences are also from 《毛泽东选集》, which is the Selected Works of Mao Zedong. These are not in public domain.

mkohler · August 7, 2021, 12:18pm

Thanks for reporting, this will be taken care of with the next deployment of the Sentence Collector.

bozden · September 11, 2021, 10:31am

@mkohler @ftyers

There are some ~590 sentences waiting in Turkish. I scanned the first 10-20 or so, they are from:
https://tr.wikisource.org/wiki/En_Alttakiler

It is indicated as “public domain in Turkey” at the top, but at the bottom CC BY-SA…

Also many of them are incomplete sentences, only sentence parts divided at any punctuation, including commas. So many of them are grammatically incorrect anyway…

PS: I did not review all of them… Accepted a couple then reviewed the CC0 status.

Edit: Scanned other sources in the set, they are mostly poetry with similar copyright status. Such as
https://tr.wikisource.org/wiki/Takatım_Tak_Oldu_Bican_Olmuşum
https://tr.wikisource.org/wiki/Çocuklara
https://tr.wikisource.org/wiki/Bir_Roman_Kahramanı
https://tr.wikisource.org/wiki/Sayfa:Halk_Edebiyatı_Antolojisi.pdf/238

ftyers · September 11, 2021, 2:16pm

Hey there! The CC-BY-SA at the bottom is the copyright of the MediaWiki interface. But in any case it looks like a copyrighted book and published recently, so not public domain / CC-0. If you think the sentences are worthwhile, get in contact with the authors, if not, then I’d say we should just delete them. Could you file an issue on sentence collector ?

I wonder if there is a way to find out who uploaded them to help them work more effectively?

mkohler · September 11, 2021, 2:33pm

No, this is exactly what this thread here is for. I’ll take care of this.

bozden · September 11, 2021, 2:40pm

The first one is a report of a Turkish NGO’s EU project. In the report it is declared as “public domain”. On the other hand exact wording results that it is not CC0 (Google Translate) - in fact a bit eclectic.

This report is in the public domain. It is possible to quote from the report by showing the source. All or part of the report may be printed, reproduced, photocopied, copied to electronic media or distributed widely without permission.

Anyway, most of the sentences are not complete…

The others are in fact public domain (folkloric or end of protection period). But these are poems ! I don’t know any rules towards these, but they are not natural sentences of course. I had already examine
Orhan Veli Kanık’s work, some verses can be appropriate but one must pre-select them…

If we could build a Turkish community we could know each other under a sub-forum but now I don’t know who posted these.

I would say delete them.

Thank you btw…

ftyers · September 11, 2021, 2:45pm

If it is “public domain”, then it counts as CC0, so that is ok. But in any case if the sentences are poorly segmented, I think probably we’d want to delete them anyway and just re-import them.

As for poetry, I think this is not the ideal text… we want stuff that is dialoguey as far as possible. Think software with voice interaction. As much fun as it might be to talk in Turkish or Ottoman poetry with my GPS assistant, perhaps we should start with more day-to-day texts

Having a Turkish community subforum sounds like an excellent idea. We could do it with a Matrix chat, or perhaps ask for a subforum on Discourse?

ftyers · September 11, 2021, 2:46pm

Oops sorry ! Belay that suggestion!

bozden · September 11, 2021, 2:57pm

If you can delete the existing ones, I’ll be happy to re-add them correctly…

Matrix flows away, discourse is best. As many natives do not know English good enough and information is very scattered, I would like to collect guidelines in Turkish in there.

Sorry for hacking the thread…

mkohler · September 11, 2021, 3:06pm

So, to get it right, I should delete all sentences coming from the sources mentioned in your original post here? Sentence collector copyright issues

Anything else?

bozden · September 11, 2021, 5:36pm

Sorry for the late reply. I checked sources I could access and suggest deletion of the list below. Probably these are all of them. All are incomplete sentences or verses of poetry…

(the links got malformed in discourse thou)

Thank you @mkohler …

mkohler · September 12, 2021, 10:57am

Thanks for the list @bozden. This has now been taken care of!

bozden · September 12, 2021, 11:36am

Thank you @mkohler, I reposted allowable sentences from the first link.

takfox · November 6, 2021, 11:11pm

There are a lot of sentences from Tatoeba in the Norwegian Bokmal collection. As noted for Swedish, this collection is mostly CC-BY, and there are only two Bokmal sentences in the entire collection marked CC0. Bit of a shame since it looks like it might be most of the sentences

takfox · November 6, 2021, 11:30pm

Incidentally, https://www.nb.no/sprakbanken/en/resource-catalogue/oai-nb-no-sbr-47/ appears to be a high quality source of thousands of CC0 sentences. I’m happy to import them if appropriate

mkohler · November 7, 2021, 12:56am

Unfortunately quite a few sentences, but it is what it is. I’ve taken care of those.

“In total, the material consists of approximately 700,000 translation pairs/sentence pairs.” sounds very promising indeed. License seems fine as well. I think importing those would be great. You might want to have a look at the bulk import process though, as going through Sentence Collector with that many sentences is not really efficient. See “Bulk submission” at https://common-voice.github.io/community-playbook/sub_pages/text.html.

jakub.wrobel7 · November 10, 2021, 9:11pm

I did a quick scan through polish sentences and found some unwanted sources from opensubtitles:

“source”: “Open Subtitles: https://www.opensubtitles.org/pl/subtitles/7665545/z-nation-at-all-cost-pl”,
“source”: “https://www.opensubtitles.org/pl/subtitles/7716578/the-100-sanctum-pl”,
“source”: “https://www.opensubtitles.org/pl/subtitles/7719533/guava-island-pl”,
“source”: “https://www.opensubtitles.org/pl/subtitles/7723395/the-orville-sanctuary-pl”,
“source”: “https://www.opensubtitles.org/pl/subtitles/7724670/brooklyn-nine-nine-he-said-she-said-pl”,
“source”: “open subtitles”,
“source”: “opensubtitles and project gutenberg”,
“source”: “opensubtitles.org”,

mkohler · November 11, 2021, 7:27pm

Thanks. I’ve taken care of this.

Sobsz · February 17, 2022, 11:15pm

In Toki Pona (tok), there are some sentences credited to http://tokisoweli.blogspot.com/, which doesn’t mention any specific rights other than “mi pana e sitelen ali mi tawa jan ale.” (“I give all my writings to everyone.”).