this string needs to suit other languages better

[Maxxius]

hi, here's one instance where you can't translate accurately from english to my language (or I assume to some other languages too).

There is this string user_own_activity_item which sets this string %2$s posted %1$s. In english its all good. you can post an image you can post a forum post, you can post a comment.

But in my language you can only:
you can post (patalpinti) an image
you can post (parašyti) a forum post
you can (pakomentuoti) post a comment.

each activity has its own word best suited for it.

something similar like that was attempted since there is also this string

album_user_own_activity_item - %2$s added images to %1$s in

but its not enough. The same goes for the dreaded create a database where you need to enter:

Singular Lowercase Noun
Plural Lowercase Noun
Singular Capitalized Noun
Plural Capitalized Noun
Indefinite Article

which is a good attempt at letting us localize it but still its not enough first thing that stands out is that when you view somebody's activity on their profile page and you get to Articles. Word strings look like that:

module__cms_records1 (good)

content_db_lang_su_1_pl (good)
content_record_comments_title_1_pl (bad) because it takes Singular Capitalized Noun + adds the translation of word comments. Thing is that Singular Capitalized Noun does not go well here while it goes good in other areas. And none of the 5 nouns that we can input when creating a database matches the word which should be in content_record_comments_title_1_pl we need a 6th string.

I will post back in this topic when I see or remember anything else related to the difficulties translating IPS.

[Sonya*]

1 hour ago, Maxxius said:

Singular Lowercase Noun
Plural Lowercase Noun
Singular Capitalized Noun
Plural Capitalized Noun
Indefinite Article

I can confirm, using above strings is awkward in German and grammatically wrong in Russian 

This has been already reported in the past with very good explanation why this concept is not appropriate for other languages:

 

 

[Sonya*]

I still think the best solution would be to use a case matrix as described here and then pick the right word form from the matrix by case and number

 

[Meddysong]

3 hours ago, Maxxius said:

Singular Lowercase Noun
Plural Lowercase Noun
Singular Capitalized Noun
Plural Capitalized Noun
Indefinite Article

which is a good attempt at letting us localize

That's not actually an attempt to facilitate internationalisation but something required for English! You can't just add an s to create a plural in English (sheep > sheep, child > children, fly > flies), and the entry for indefinite article is there so that we can indicate whether the word should be preceded by a or an. 

[Meddysong]

54 minutes ago, Sonya* said:

I still think the best solution would be to use a case matrix as described here and then pick the right word form from the matrix by case and number

I initially thought to myself that this couldn't work (we have different numbers of cases and don't all use the same one in the same construction) but having reread your comment, I see now that it doesn't need to be complicated.

In your design, each user can add rows and columns as required. I'm fortunate in that I only have two cases, so I would have a small matrix, whereas Russian has several so your grid would be larger. And then we adjust language strings to be something like "User added [x,y]", which would pick out from the grid, in my instance, a singular accusative, and English "a record/an entry" etc.

Yes, I really think that could work.

[Sonya*]

Yes, it is like a permission Grid where we have cases at the left side and numbers at the top. Picking the right word form we could ensure the string is grammatically correct. And yes, we can use the same word form in the matrix if there is no difference in one specific language. In German we have "only" 4 cases and a lot of words do not really have different forms in all these cases. In Russian we have 6 cases and they are heavily used to build grammatically correct sentence as we do not have firm word order in Russian. The only possibility to distinguish between subject and object is to apply a correct case. 

[Meddysong]

I know that @Mark is extremely keen on internationalization and that @Matt is the Pages guru, so it probably makes sense to draw their eyes to your suggestion. I think it could solve a lot of persistent problems faced by those of us using different languages.

[Maxxius]

Another one:  __indefart_blog_entry

In one instance it should be translated as blogo įrašą and it sounds good when the sentence is like Somebody reacted to blog entry.
But when you are mentioned in the comments of the blog entry when you get a notification it should say blogo įraše. and instead you can only have blogo įrašą and in that context it makes you look like you made a completely childish mistake when you made the translation 😄

[Meddysong]

I know what you're saying but it's not going to work if you just list individual language strings every time you find one. It needs to be something much more systematic.

The problem you're describing there is a result of our having to define individual words for things like a blog entry, a post, a topic, a file. Those words are then dropped into other strings such as "%s added %s", "%s replied to %s", "%s reacted to %s", and so on. In English, you can just insert those words, which is why the system works. But for our languages, the grammatical role of the thing being inserted changes depending on the construction, and so the word is modified.

Several of us have raised these thing in the past but realistically, we're such a small part of the market and the changes would be so intensive, that I think we're not likely to get them in 4.x. Our best bet is probably to hope that IPS in 5.x is open to dropping the approach of inserting a string into various scenarios if English permits this, even though it makes it look inefficient to people looking at the code and thinking that the other approach is better, because they only have English as a reference point.

[Maxxius]

Yeah.. I have little hope for IPS 4, I'm just trying to get the attention so it would be addressed in IPS 5 at least. Because if you'd bring it up when IPS 5 is released then you can wait for many more versions. At the time I'm strongly thinking about using abbreviations for such words and be done with it.

How did you worked around these limitations? Anything out of the ordinary that gave you good results @Meddysong? I see you know a bit about my language. Are you also from the same country?

[Meddysong]

4 hours ago, Maxxius said:

I see you know a bit about my language. Are you also from the same country?

It's not that I know much about your language (I'm afraid I don't know anything, although I've been there before), it's just that I'm familiar with languages in general.

I haven't often found solutions for my own language, so the mistakes are still in there on my site. I did have the thought that what I could do was to write a plugin to change some of the templates into using new language strings (word_acc, word_gen, word_dat etc) for the various cases. I suppose that might work but it would take a long time to do, especially for someone like me who isn't an experienced coder.

[Sonya*]

1 hour ago, Meddysong said:

although I've been there before

Me too. Back in my teens in Ignalina learning Lithuanian. Unfortunately the only things I can remember are: Labas and Tu turi cigarettes?  Ten years ago then the second time with boot in Klaipeda and Nida. Beautiful country and very unusual language. 

[Rikki]

The 'grid' idea is interesting. Know of any other product that does localization with that method?

[Maxxius]

@Meddysong oh you traveled a lot , nice 🙂 I've read your site some time ago , now I remember.

@Sonya* wow you know the right thing to ask 😄

[Sonya*]

9 hours ago, Rikki said:

The 'grid' idea is interesting. Know of any other product that does localization with that method?

No other product. This was entirely my idea and I am giving it to you as a gift