menu
Tatoeba
language
Register Log in
language English
menu
Tatoeba

chevron_right Register

chevron_right Log in

Browse

chevron_right Show random sentence

chevron_right Browse by language

chevron_right Browse by list

chevron_right Browse by tag

chevron_right Browse audio

Community

chevron_right Wall

chevron_right List of all members

chevron_right Languages of members

chevron_right Native speakers

search
clear
swap_horiz
search
sysko sysko May 24, 2013 May 24, 2013 at 7:38:37 AM UTC link Permalink

About sentence coming from the Tanaka corpus and still orphaned
(I'm posting as a new thread so that it will be easier to see)


What about this

1. if you're a native speaker and you see a orphaned sentence that sounds unatural, you had a natural one
2 - you post a command and tag asking for the deletion of the orphaned one.

At the same time I'm going to see to not permit the translation of orphaned sentence, and to remove them from the search engine (which will have the consequence to also remove them from the 'random', as the random id is given by the search engine)

What do you think?

I know the easiest way would be to remove them directly and start again, but as Tatoeba is also used for editing the index of the www-jdict.

{{vm.hiddenReplies[16635] ? 'expand_more' : 'expand_less'}} hide replies show replies
tommy_san tommy_san May 24, 2013 May 24, 2013 at 2:02:18 PM UTC link Permalink

> 1. if you're a native speaker and you see a orphaned sentence that sounds unatural, you had a natural one (You mean "you *add* a natural one," right?)
> 2 - you post a command and tag asking for the deletion of the orphaned one.

I think there's no problem with this, although I'm quite sure that we won't actually delete many sentences because it's hard to decide that a sentence is absolutely unnatural.

> At the same time I'm going to see to not permit the translation of orphaned sentence, and to remove them from the search engine

I find this great!


There will still be at least two problems.

First, when a sentence has several translations, both owned and unowned ones, you'll need to visit each pages to see if the sentence is owned.

Second, even if you don't show orphan sentences as a random sentence, they will still show up frequently as a translation of the random sentence, which will likely to give native Japanese speakers the impression that this must be an unreliable website.

I think you can solve them both by making a new option "Display orphan sentences" in the settings page and set it "off" by default.

{{vm.hiddenReplies[16636] ? 'expand_more' : 'expand_less'}} hide replies show replies
AlanF_US AlanF_US May 24, 2013 May 24, 2013 at 7:40:49 PM UTC link Permalink

It's true that today you must visit each translation to see whether the sentence is owned. However, I imagine that it wouldn't be too hard to change the interface to display the owner (if any) next to each translation. That would be useful for several reasons.

I agree with everything else you said.

{{vm.hiddenReplies[16640] ? 'expand_more' : 'expand_less'}} hide replies show replies
tommy_san tommy_san May 25, 2013 May 25, 2013 at 12:45:42 AM UTC link Permalink

Ah, yes. That would be really desirable, too!

tommy_san tommy_san May 25, 2013 May 25, 2013 at 1:08:11 AM UTC link Permalink

I'd like to ask you if you really want us to spend our time deleting bad orphan sentences. I feel that adding new good sentences would be much more rewarding and profitable. There are so few native Japanese speakers here, and it takes much more time to judge whether a sentence is bad enough to be deleted than to make a new sentence (at least for me).

I agree that owned unnatural sentences must be changed or deleted, but I'm reluctant to approve the idea that we delete bad orphan sentences. They will go to the "low reliability database" in the near future anyway, won't they?