Weekly Incubator Update: Tracking Progress from November 27th to December 4th
Several eggs ready to hatch
Hopefully we will see the top three courses go to beta within a month.
The "Related discussions" feature now brings some of the past WIUs next to each week's WIU. Cool! I am assuming this is available to most if not all of the users.
PHASE 1 Progress: Total 22 courses
Portuguese for French - 99% | 99% | 99% | 99% (+0) 31-Dec-2016
English for Thai - 99% | 99% | 99% | 99% (+0)
German for Italian - 99% | 99% | 99% | 99% (+0)
Italian for Portuguese - 73% | 75% | 78% | 81% (+3) 25-Dec-2016 *
Swahili for English - 70% | 70% | 80% | 80% (+0) 31-Jan-2017 ^
Czech for English - 72% | 73% | 74% | 75% (+1) 7-Jul-2017 ^
Spanish for Italian - 70% | 70% | 70% | 70% (+0)
Swedish for Russian - 69% | 69% | 70% | 70% (+0)
French for Chinese - 49% | 49% | 49% | 49% (+0)
Russian for Turkish - 47% | 47% | 47% | 47% (+0) 31-Dec-2016
French for Turkish - 45% | 45% | 45% | 45% (+0)
Korean for English - 35% | 36% | 37% | 38% (+1) 31-Dec-9999
Klingon for English - 33% | 33% | 33% | 33% (+0) 31-Dec-2016
Hindi for English - 28% | 28% | 28% | 28% (+0) 26-Jan-2017 ^
Indonesian for English - 28% | 27% | 27% | 28% (+1)
English for Tagalog - 25% | 26% | 26% | 26% (+0) 24-Dec-2016
English for Bengali - 23% | 23% | 23% | 23% (+0) 20-Feb-2017
English for Tamil - 22% | 14% | 15% | 17% (+2) 28-Feb-2017
Spanish for Arabic - 9% | 10% | 10% | 11% (+1) 31-Dec-2017 ^
Yiddish for English - 9% | 9% | 9% | 9% (+0) 7-Jul-2018
English for Telugu - 6% | 6% | 6% | 6% (+0)
High Valyrian for English - 1% | 2% | 3% | 4% (+1) 1-Jun-2017
Course - 3 weeks ago | 2 weeks ago | a week ago | Now (Progress delta); Estimated Launch Date (provided by contributors) (Date delta)
'Estimated Launch Date' only when provided by the course contributors
Mean - 0.35% | 0.29% | 0.82% | 0.41% (-0.41)
Median - 0% | 0% | 0% | 0% (+0)
* This week's Leader Extraordinaire!
^ The Hindi, Czech, Swahili, Spanish (for Arabic), & Tamil teams' progress is as per their own calculation
Here's what the contributing teams have said during the last week:
(For Phase-1 and Phase-2 courses by default, and for Phase-3 courses per request).
Trees and con-tree-butors
Just a quick update: we've taken on our new contributors as planned, and are in the process of getting everyone oriented to the work ahead. They'll be building lessons soon!
In the meantime, here is a screenshot of the first 70% or so of the eventual tree-to-be:
Of course, other areas of topical or functional vocabulary will be embedded within the various modules, but the above gives at least a sense of the overall contour.
You may notice the presence of lessons based around particular holidays. These are designed to address culturally important sets of vocabulary (in Yiddish-speaking communities both religious and secular), as well as to give another opportunity for contextualized language practice and review.
Zayt gezunt! Mendl
purtroppo siamo costretti a posticipare la data di rilascio del corso in Beta. Non ci è stata ancora comunicata dal team di Duolingo una stima dei tempi che ci vorranno, quindi mettiamo arbitrariamente la data avanti di un mese.
Nel frattempo il nostro team sta lavorando per controllare e migliorare il corso e aggiungere i suggerimenti nelle varie lezioni.
Nel frattempo vorrei ringraziare tutti quelli che si sono prenotati per il corso e che sono ansiosi di imparare il tedesco! Sappiamo che l'attesa è stata lunga e ci dispiace doverla estendere ancora:(
- 887/2408 words
- 100/105 images
Also, I would like to use this moment to pay tribute to Chapecoense and the victims of the LaMia Airlines Flight 2933. Our hearts goes out to the victims of the tragic incident. May the heroes who lost their lives rest in peace.
Some people use vague translations from tagalog to english and vice versa. However, we shouldn't do that. We must make our translations as clear and as specific as possible in order to avoid confusion.
Take the word "take" for example: e.g. Take it. Take food and water.
There are some cases when the word "take" could be translated to the word "dalhin", which is the tagalog word for bring...instead of going with this translation, we could use a more specific translation..."kuhanin" which is the tagalog word for "take"...
If you translate the sentences above...it would look like this: Kunin mo ito. Kumuha ng pagkain at tubig.
This is more specific than using the word "dalhin": Dalhin mo ito. (bring it) Magdala ng pagkain at tubig. (bring food and water)
However, there are some english words that doesn't have tagalog translations...and there are some english words that has vague tagalog translations...let's all do our best! ^^
1490 words, 75%.
Looking forward to the December Holidays
As always it is overwhelming to see the high level of interest in our upcoming course. We are looking forward to a major progress achievement during December.
Yiddish update this week.
Previous Update 20-Nov to 27-Nov
Last updates from:
2016: Nov, Oct, Sep, Aug, Jul, Jun, May, Apr, Mar, Feb, Jan
2015: Dec, Nov, Oct, Sep, Aug, Jul, Jun, May, Apr, Mar, Feb, Jan
2014: Dec, Nov, Oct, Sep, Aug, Jul, Jun, May, Apr, Mar, Feb, Jan, 1st.
Thank you as always for your work!
Has anyone been in touch with the Portuguese for French speakers group? It seems like they've been at 99% for the longest time, so it would be nice to know what items remain to be completed and if they still believe that they can hit the launch date of the end of this month.
Same thing with English for Thai, but at least they do (some) updates.
Languages I would love to see in Duolingo: Cantonese, Mandarin and Japanese.
Very Asian XD
Languages I would love to see in Duolingo: literally anything people have asked for. Today is the 10 month anniversary (mensiversary?) of the Korean course.
Very exciting Yiddish update from Mendl! Can't wait to see how the tree develops and, of course, for its eventual release.
Yiddish actually went down to 8% and then back up to 9% in the last week. So in addition to their update, there is also movement, even if you couldn't see it.
Switching to the two-digit format, e.g., "76" next time, to avoid burying the main thing in fluff. ;-)
I wasn't planning on doing Yiddish, however seeing all the cultural lessons in there makes it pretty tempting.
Happy about Yiddish :D Also excited for the Swahili course to come out!
I hope to see updates of "English for Persian" in "weekly incubator update" series soon! :)
Wishing for the realization of the hope :)
And I suppose, there are many many more out there wishing so:)
Did you receive any feeler from Team Duolingo :)
Thank you! And it's lovely to see the sidebar related discussions.
eta: Oh, wow. I just clocked the "con-tree-butors." Well done, pun patrol.
Yes, this week's post is delayed by more than a day...
Please check after 12 hours from now.
On the same note there will be no update on the day of Christmas. Please expect that edition on the 26th.
They have been, but they don't use the automatic completion that Duolingo has. They make their own updates for when they progress, but if they did that every time they did something, the incubator page would be cluttered!
"major progress achievement" is the new Soon^tm
LOL, jk Jiten! :D
That's GREAT news!!!! Can't wait for Hindi. I am sure I'll just love it and try to master it for a few years then go to India.
Assuming that the alpha-test for the alpha-syllabary level is over by now, a number of discerning learners and some Hindi natives who like to be participators in the upcoming course's forum, would like to know whether Devanagari alpha-syllabaries will be taught through some Roman transliteration standard? If yes, which one the following standard will it be: IAST / Harward-Kyoto / ITRANS / Velthuis ?
Thank you, and wish you best of luck for the major progress achievement :)
Really? Are they off by that much? 74% vs 99% ? I thought they are the number that's in the WIU.
The "incubator" numbers can be worse than useless for courses that start out from an empty tree (Czech) or from an unrelated hand-me-down tree (HV). This is because the official % progress is (very nearly) 100*[words with 3 or more sentences entered]/[total words in the tree, including those with less than 3 sentences]. So obviously the denominator can be a problem. For example, the Czech 99% is so because I always add only about the next row of empty skills by the time of completion of any given row, and that empty row of skills may even be only partly populated with empty words when I get to working on it. I am always "nearly done" because I do not structure too far ahead. The simple solution would be for Duo to allow teams to specify their total target word count to use as the denominator, much like we are able to estimate our own custom date of completion in lieu of the bogus auto date. Not happening, not even "soon".
Czech team said 5 months ago progress bar is misleading, and giving very wrong information.
I just realized something about the High Valyrian for English course. One of the contributors is the guy who constructed this language in the first place, right?
He could construct more of the language according to what works best in Duolingo as it already is...
...or construct more of the language specifically to test Duolingo's limits.
For example, if Duolingo releases a course teaching Chinese or Japanese and has problems with the lack of spaces between words, that's going to look really bad. :/
If Peterson invents a script for High Valyrian that doesn't put spaces between words, and then Duolingo releases the course and has problems with that lack of spaces, it's not going to look as bad. It would also give Duolingo some more experience at dealing with the problems so they're more prepared to make courses teaching Chinese and Japanese. ;)
If Peterson invents a script for High Valyrian that doesn't put spaces between words
It would be annoying to type.
Honestly I don't understand the no spaces between words being a problem at all, we already have Vietnamese, which puts a space between all the 'characters', why can't Chinese do the same thing?
M a y b e i t i s h a r d e r f o r t h e s o f t w a r e t o t e l l w h e r e o n e w o r d y o u t y p e d b e g i n s a n d a n o t h e r e n d s ?
Why am I being downvoted? Vietnamese ba is functionally identical to Chinese 三. Vietnamese co ay (no keyboard on this machine) is more difficult than Chinese 她. Vietnamese is fine with having a space between everything that in chu nom was a character, Chinese people read Chinese without encountering difficulties. I do not see any problem. Sarcastically putting a space between every letter is an entirely different kettle of fish. M means nothing, 三 means three
Seriously, even in pinyin, ren shi nin wo zhen gao xing makes precisely as much sense as renshi nin wo zhen gaoxing, only convention says differently.
Putting a space between every character would be just as bad as having no spaces. For all the languages currently taught, the space lets the system know where one word stops and another ends. For agglutinative languages that add prefixes or suffixes, the system treats it as a different word each time a new letter is added.
Maybe it would be easier if Chinese had spaces like you added in your pinyin for "renshi nin wo hen gaoxing," but the course contributors would have to make up their own rules for when to add spaces, and no real Chinese is written like that, leaving all the users at a disadvantage.
Plus, the spacing issue is just one problem. Chinese would be the first course not to use some sort of alphabet, and Duolingo is still figuring out how to do that. Would people just type in pinyin and have the characters show up, or would there be some sort of character practice?
Chinese people read Chinese without encountering difficulties, and they're people. Computer software isn't as good at all the pattern recognition that people have.
Naturally, but the future learners of Chinese on Duolingo are people, the sentences those people will translate will be written by people, and Vietnamese already provides us with an example of a language that has a grammar system effectively identical to Chinese, actually works by having a space between each character, and like Chinese multiple characters when combined make 'words', like 'ren shi' meet and 'co ay' she/her.
The Chinese course faces a problem in Chinese characters, and also in teaching method, but these are the issues. Spaces between characters are not (as shown by Vietnamese), and levels of formality, which has also been mentioned, is also not the problem as Vietnamese has a very similar system which isn't taught at all by the course.
Vietnamese still uses the Roman alphabet and still uses spaces between pieces of sentences. Think about the last time you saw written Vietnamese: did it look more like "Tôi gửi ít đồ cho gia đình" or more like "Tôigửiítđồchogiađình"?
Also, I agree with you that the future learners of Chinese on Duolingo are people and the sentences those people will translate will be written by people! :)
At the same time, when you take a Duolingo course and translate a given sentence from one language to another, the sentence you type is evaluated at that time by the software itself. You don't have to wait for a person to read what you wrote before you get an answer from Duolingo.
When I type "nova" in Spanish, it's the software itself that needs to tell the difference between "nova" and "no va" in order to let me know if I translated it right or wrong ("nova" is a correct translation for "dead star" and an incorrect translation for "doesn't go"). Duolingo doesn't send what I typed over to a Spanish-speaking person to read and reply before it gives me an answer.
Vietnamese uses punctuation to separate pieces of text, I've never seen anyone write Vietnamese like Tôigửiítđồchogiađình, in fact Vietnamese keyboards don't even let you do it. It would be so confusing. Chinese doesn't have spaces, but all the characters take up the same amount of room, and are clearly distinct from each other because each Vietnamese 'word' in Chinese is written with a separate character. This makes it a non-issue. (Let me reiterate again: what I suggest Chinese does, and what Chinese is doing on the reverse course, is exactly what Vietnamese actually does in real life, except with characters instead of pinyin)
As I understand it, the software looks in a database for translations that are to be accepted, and each possible translation has to be entered separately by humans. This was, I heard, one of the large issues with Japanese. There are so many possible translations that it would be too hard to enter them all unless the limit of possible translations for each sentence (currently around 3000) were increased. This is also the reason for the 'slow' progress of the Czech team, despite them being perhaps the hardest working team this site has seen (and there's a lot of really hard-working teams). For so many sentences they are pushing that limit of possible translations, which includes typos. There is automation used to generate the vast databases of translations, but this is done by the humans building the course, not by internal software. It's possible that I entirely misunderstand, but that's the impression I've been left with.
The issue is not that humans can't read the mushed together text, it's that a COMPUTER can't read it.
I'm gonna assume you have only a little if any programming experience here so sorry if that proves otherwise. If I told you that I had a file containing some sort of text, that that text had meaning and could be read by a human, but there were no spaces or characters to separate words, how would you write a program that could then separate the words for you?
A computer can only read something one character at a time, so it can't tell where a word starts and where a word ends.
Yes, the program works by checking input against a database, but it can't do that if it can't even figure out the input against which to check because it can't figure out what words the user typed in the first place.
Kippis is working on the English for Japanese speakers course, and his post and comments at https://www.duolingo.com/comment/7862207 are really worth reading. The spaces issue is in the first comment in the comment section below the main discussion post.
Meanwhile, http://stackoverflow.com/questions/1605353/word-break-in-languages-without-spaces-between-words-e-g-asian is worth a look.
I read through this, but I really don't see the issue still. I can only reiterate that the Vietnamese course works fine.
I'm afraid I still don't understand. When you click submit, your input is sent to the computer, where it's compared to a file full of entries. Many of these entries can be automatically generated by a program based on common mistakes. Midtakes, miatakes, miwtakes, mietakes, etc are all easy typos, miptakes is not. It's been a long time since I've studied computing, so I don't remember the term for that, but it's common and easy to do.
Duolingo itself might be able to work out some of these, although I have had a lot of my typos marked wrong, because the software has limits. I just don't see how this is an issue with characters though. The input either matches exactly, or it doesn't. A Chinese course would teach Chinese, not pinyin. Even if the course taught pinyin, Chinese is so regular, each character has an initial and a final, that is all. Each character has a tone, and only one tone. Just as with Vietnamese, I don't see an issue.