IMHO we should pay much more attention to links in general.

There is no information when someone deletes a link, there is no “Recent (Un)linking” page to see them… And no way to see the number of links to the sentences in certain language (it can be a way to motivate people to create new links to promote their language, not create endless duplicates).

IMHO, links are also very important.

“This software” is Google language API. :)

If nothing has changed, Tatoeba uses it for language detection.

boracasli put it there to show sysko on the IRC a week or two ago.

I don’t think it’s a big problem.

Since no one added a translation to these, maybe they're just too hard to translate. :o

> I don't think Demetrius uses g-mail
In fact, my clone account uses gmail ;)

Please see my comment to the post below. ^^

http://tatoeba.org/eng/wall/sho...7#message_3187

Please read:
http://tatoeba.org/eng/faq#new-language

It may take some time to add a new flag.

You can start adding Interlingua sentences right away, but please say in comments it's Interlingua.

Well, in fact I had to use an automatic Esperanto->English translator to understand what your post means, so probably not. ^^

Ну, на самом деле мне пришлось использовать автомататический эсперанто-английский переводчик, чтобы понять, о чём Ваше сообщение, так что, наверное, нет. ^^

Fakte mi devis uzi automata tradukilo por kompreni vin. Tial mi pensas, ne. ^^

> Yes, it's called Globish.
I think you're the only one who calls it like that.

I think the problem is that we need rules about what sentences can be put under what level.

I think we've decided to use established systems from tests for foreign learners...

> The problem in fact is not one word sentence
> itself, it's that some user will abuse, and I
> want simple rule for moderators to play their
> role without waiting for me or Trang to say
> "yes" or "no".
Hmm... That may be useful indeed.

But please make it clear that this guidelines are only for adding new sentences, not for translating the existing, and that keeping sentence natural is more important than making the sentence limit.

> It's still a beginning, but as you can see
> they handle conjugated forms, common sayings
> etc. and translations.
That is an interesting project. :)
I'm looking forward to its opening.

> and by the way, I warn you that
> many French audios on Tatoeba don't
> actually feature proper liaisons, so
> are actually useless in terms of
> education for foreign-learners,
> as most French people actually
> pronounce these liaisons...
Unfortunately, I currently can’t listen to audio on Tatoeba, but it seems to be a great problem.

Unlike sentences, audio seems to be a “bonus” feature that most people don’t use, so it’s immature and prone to errors and undetected bugs.

What is more, while we have an established way of fixing errors in a sentence, only administrators can fix the audio... :o

Maybe you can give an example to attract some attention to this problem?


> Since I joined Tatoeba, I haven't seen
> the deduplication script run once!
> I ended up believing it is a myth!
Well, I did see those mythical times. :)
But I hope the script will be fixed soon.

> Actually, it should be new Globish phrases
> altogether that should be banned !
Well, what concerns non-standard phrases, I believe that tags should be keep track of these.

If some word is a new coinage, or colloquial, it should be tagged as such, so that people who want to re-use Tatoeba data will be able to filter such sentences if they want to stick to the established word usage.

Additional tags can be created for some languages. For example, in Ukrainian we have tag “Russism” for those who don’t want to use the language constructions influenced by Russian.

Once more on 1-word sentences
-------------------------------------------------

### What’s wrong with them? ###

We don’t like one-word sentences not because they can be found in a dictionary.

In fact, lots of things can be found in a dictionary. There are lots of proverb dictionaries out there.[1] Does it mean Tatoeba is duplicating things instead of collaborating? *Yes.* But if it does it better, it’s OK.[2]

We don’t like one-word sentences because it’s an unneccesary duplication. In fact, 90% of them are duplicales of the small subset of grammatic constructions:

* A nominal sentence
* A sentence with the predicate omitted
* An imperative sentence

They have no valuable to learners.


### Why banning is not an option ###

BUT! There are good useful one-word sentences:

* Sometimes we have some etiquette phrases that have several words in one language but one word in other. E.g. «Thank you» is «Спасибо» in Russian.

* Most languages allow omission of some parts of sentence; in some cases the sentence with the omission is natural while the ‘full’ sentence is not (e.g. pronouns like ‘yo’ in Spanish[3]).

* There are polysynthetic languages where a word can mean a lot of things



If you ban one-word sentences:

* Tatoeba will never cover all the grammatic structures

* Users will be unable to add translations to some sentences, or, what is worse, will have to add unnatural 2-word translations, or translations that do not really match


### What is really neccessary ###

What we need to do are some rules against unneccesary duplication (e.g. we have «Пожалуйста!» and «Пожалуйста.» with an exclamation mark and with a dot — it is surely superfluous), not rules against one-word sentences.

As CK once noticed, sentence pattern “What is happiness?” can be used to cover a lion’s share of the vocabulary: from “What is aali?” to “What is zytel poliamide?”. Is such duplication less dangerous than one-word sentences?


### Some more criticism ###

Why was ‘One, two, three…’ deleted? In Russian, ‘one’ is ‘один’, but when counting something we use ‘раз’ (a time): “Раз, два, три…”.

It’s never really used outside of this context, only when counting something without specifying a thing, so you’ve deleted a valuable example.


### Footnotes ###
[1] In fact, even [Wiktionary](http://en.wiktionary.org/) lists proverbs.
[2] It does it better because it’s free, multilingual... and because it works well.
[3] Please correct me if I’m wrong.

People debate because they believe sth is better...

But do as you wish. I'm out.

> The goal of tatoeba is not to cover
> all the panel of possible things one
> can say or write. If we cover 90% of
> possible structure, without having to
> debate during days about whether an
> entry has to be or not in tatoeba.
I don't understand in what way "Buy!" contradicts to the goals of the Tatoeba.

I don't understand why you've started this debate in the first place.

> and yep anyway tatoeba is a
> dictionnary of example sentences. I
> mean a "word dictionnary"
These 2 sentences seem self-contradictory to me.

> seems you still haven't read
> the "it will all be explained" and
> stated in the blog post.
Blog post is the final decree.

What I question is the neccesity of such a decree.

Suppose we didn't have a "Sorry" in Tatar.

Someone would add "Ğafu itegez". It will pass the the rule. BUT no one will be able to translate it, since the translations won't pass the rule.

That's a pity.

What is a 'dictionary entry'?

I have several dictionaries of proverbs on my bookshelf. In fact, **anything** can turn out to be a dictionary entry.

> for the precise decision I've said it will
> be done in the blog post
Ah, yeah, we know how precise the blog posts are.
If we don’t yet have a definition of what a “sentence” in Tatoebian sense is (and I believe any attempt to give it would require deleting something from the database), so how can you give a definition of what an one-word sentence is?

> but do we have time to debate on all sentences ?
Well, since you did have time to delete boracasli’s harmless one-word sentences, I presume you do have spare time… -_-

> So my opinion is rather
> a little strict rule, which
> will maybe avoid some
> sentences, but easy to remember
> apply, rather than weak definition.
Strict rules simply don’t work when dealing with languages.

> because "An apple." can be a nominal
> sentence, answering to "what are
> you eating".
Note you can't find it in a dictionary, since dictionary doesn't list articles and article is not an inflection, so it passes the defintion you've given before.

IMHO what makes Tatoeba so nice is that the common sense prevails over the rules here. But everything does change... >_<