Chinese Text Analyzer vs DuShu App

[Jan Finster]

Is there anyone, who has used both Imron's Chinese Text Analyzer and the DuShu App (https://play.google.com/store/apps/details?id=com.trescher.dushu&hl=de).

 

I am currently using DuShu and it seems to do a lot of things really well:

a) importing texts and extracting sentences. Letting you study sentence by sentence. 

b) assigning HSK levels to words

c) providing pinyin & translation 

d) saving vocabulary, creating flashcard and several exam functions...

 

How do both programs compare?

[mirgcire]

You can find a detailed review of Imron's Chinese Text Analyzer here written 5 years ago.

[furiop]

5 hours ago, Jan Finster said:

Chinese Text Analyzer and the DuShu App

 

I'm using DuShu (only Android) since the first version, and I am very satisfied.

Very different apps.

 

In DusShu the interlinear pinyin of unknown word is very useful and there are a loto of tools (in full version): flashcards from text, text to speech, translation via Google,  exercices, vocabulary (only in english), characters strokes and so on...

No windows version (but using scrcpy with phone in debug mode, you "copy the screen" of the smatphone and interact using mouse or touch).

 

Chinese Text Analyzer is interesting, for windows or mac, but in my opinion is only a draft version. Anyway is written in Lua scripting language (also used in game programming), so I think that in future there will be a lot of improvements.

 

 

[Shelley]

1 hour ago, mirgcire said:

You can find a detailed review of Imron's Chinese Text Analyzer here written 5 years ago

I think that will be a bit out of date, from what I have followed there have been many improvements and changes. 

 

I would also agree with @furiop that they completely different tools.

 

There is a very busy thread here https://www.chinese-forums.com/forums/topic/44383-introducing-chinese-text-analyser/

 

 

 

[ZhuoMing]

I have never used this DuShu app, but it sounds like a reader app, perhaps like the pleco reader? I could be wrong.  Regardless, I have always seen CTA as a program for analyzing the text - the value of it is the stats on the right side of the window.  I have always used CTA solely to find suitable material, not for doing the actual reading on, which appears to be the purpose of DuShu.  

[imron]

5 hours ago, ZhuoMing said:

I have always used CTA solely to find suitable material, not for doing the actual reading on

And this was very much the guiding purpose when I designed CTA.

 

Other people might use it for other things (and that's a good thing!), but for me I designed it so that I could determine whether something was appropriate to read (or which of several texts would be the most appropriate) based on my current level, and then also so that I could extract high-frequency unknown vocabulary from those texts, because learning the high-frequency words will provide the biggest payoff in terms of understanding.

 

It was not primarily designed as a reader for text (and it shows), because I believe that electronic readers with popup aids for pronunciation and translation are detrimental to developing the skills necessary to read without those aids.

 

6 hours ago, furiop said:

Anyway is written in Lua scripting language

It's written mostly in C++ (with some Objective-C on macOS), with Lua scripting available to end users who want to do custom text processing but don't want to have to wait for me to get around to adding new features

[Jan Finster]

This really helped me. I guess they are very different programs.

 

Does CTA do any translation or automatic pinyin assignment?

[furiop]

9 hours ago, ZhuoMing said:

DuShu app, but it sounds like a reader app

 

Not only, it is a tool for learning written / spoken chinese TROUGH texts.

 

During reading activity you can also test your knowledge of the text (till a specific point), learning unknown words (definition, writing, pronuciation, frequency, level,...).

It is not compulsive, but you can use flashcards text-driven and multiple answer test.

There are also statistics about the text and a classification of the level of difficulty.

 

Where are texts? DuShu start with a long list of sentences of HSK1 to HSK6 levels and links to sites like iMandarinPod, JustLearnChines

 

You can test the free app, but, in my opinion, the pro version is not expensive, and worth the cost.

[imron]

2 hours ago, Jan Finster said:

Does CTA do any translation or automatic pinyin assignment?

No, but it does allow you to look up the meaning of words.  Although one of it's more controversial features is that whenever you do that CTA automatically marks the word as "unknown", based on the premise that if you knew it, you wouldn't have needed to look it up in the dictionary, and CTA wants you to confront the words you don't know rather then let you pretend you know them when you don't.

[NanJingDongLu]

On 5/18/2019 at 5:25 AM, imron said:

[CTA] was not primarily designed as a reader for text

Sorry to revive a dead thread but @imron if CTA is not intended to be used as a reader, could you comment on how a user is supposed to update their list of "known" words? For myself, and I assume many learners, my vocab does not overlap very well with HSK or TOCFL levels as I know many words that aren't in these lists, but also haven't learned many of the words that are in these lists, even for my HSK level (especially since your criteria is "100% certain without needing to check"), and so the only way that makes sense to me is to read in CTA and update the list of words as I go.

 

Alternatively, I suppose one could just use the list of unknown words and mark the ones you do already know, but without context? What do other people do?

 

It seems like the more accurate you can make your list of "known" words, the more useful CTA is at assessing how easy you will find a text.

[imron]

8 hours ago, NanJingDongLu said:

could you comment on how a user is supposed to update their list of "known" words?

“Supposed” is a loaded word, because if something works for you, then I don’t want to discourage that, however if you wanted to use CTA for mining vocabulary, then when you export words you can click the checkbox that says ‘mark exported words as known’.

 

The expectation being that you’d then go and learn the exported words.  You can do this based on unknown words or looked up words or a bunch of other options.

 

Used regularly this would cause your known words to increase. 

[NanJingDongLu]

3 hours ago, imron said:

“Supposed” is a loaded word, because if something works for you, then I don’t want to discourage that

I get that you're not telling people what not to do; I was just curious as to your intended use case.

 

3 hours ago, imron said:

when you export words you can click the checkbox that says ‘mark exported words as known’. [...] then go and learn the exported words.

I can see how this is a good alternative to what I'm doing.

 

Personally the downside for this is that since I'm new to reading novels, when I look at 'unknown words' for 活着 it shows 3,731 unique words. maybe 1,000 of these are false positives, for example I know 听 and -到, but 听到 is marked as unknown because it's not directly in the word list. but that leaves ~2500 unique words that I will want to look up when reading. If I export and add to my flashcard rotation and am able to master 10 new words per day, it's going to take over 8 months to learn them all, which means I won't be able to read 2 full novels this year if I want to ensure my known words is accurate. This means I need to be cautious about which words I look up in CTA and which I just leave as unknown words since not all words are important enough to warrant studying and moving to 'known'.

 

Personally, I am perfectly happy with CTA as a text reader (especially once you pick the right font) and am happy to read with it, updating the word list as I see fit while I go, but I'm curious what other people are doing.

 

My main complaint with CTA is how the characters are paired into words, but I realise this is a very difficult problem to solve with code. 看着他们将提上的井水泼在地上 should clearly be 在  地上 but it is paired as 在地  上 and then marks "在地 - local" as an unknown word.

[Jan Finster]

42 minutes ago, NanJingDongLu said:

If I export and add to my flashcard rotation and am able to master 10 new words per day, it's going to take over 8 months to learn them all, which means I won't be able to read 2 full novels this year if I want to ensure my known words is accurate. This means I need to be cautious about which words I look up in CTA and which I just leave as unknown words since not all words are important enough to warrant studying and moving to 'known'.

 

Personally, I am perfectly happy with CTA as a text reader (especially once you pick the right font) and am happy to read with it, updating the word list as I see fit while I go, but I'm curious what other people are doing.

 

My main complaint with CTA is how the characters are paired into words, but I realise this is a very difficult problem to solve with code.

 

You might want to give Lingq.com a try. It can do all the things you mention above. You can mark words not as known vs unknown, but on a scale from 1-5 ("new", "recognized", "familiar", "learnt" and "known"). You can get a life-time premium account for 199$.

[Sobria-Ebritas]

On 5/20/2021 at 5:53 PM, NanJingDongLu said:

 

My main complaint with CTA is how the characters are paired into words, but I realise this is a very difficult problem to solve with code. 看着他们将提上的井水泼在地上 should clearly be 在  地上 but it is paired as 在地  上 and then marks "在地 - local" as an unknown word.

Thank you for "reviving this dead thread".