Creative Ways to Use FSI

[Speakeasy]

I have appended below a link to the HTLAL discussion thread "Creative Ways to Use FSI" as forum member "hulk_smash" has added a rather interesting post. I have offered to change the title of this discussion thread to something more evocative of the "Revived FSI" project. Hopefully, the discussion that Darren has initiated on the HTLAL will continue on this forum.

http://how-to-learn-any-language.com/forum/forum_posts.asp?TID=40957&PN=1&TPN=2

[hulk___smash]

Thanks Speakeasy,

I'm happy to continue any discussion on this forum over the old one.

[James29]

Well, Darren, I'd like to give you a big "thank you" for that old website. I was there every day for a very long time learning Spanish.

I'd also like to say that I did a ton of wonderful things with the money that I saved by learning Spanish through your free website. I spent much of that money on my trips to practice Spanish in various countries.

In terms of what could be done with FSI, I've always thought it could become a super-comprehensive course with a bit of work. In addition to the way it is now (I like FSI better than the "improved" Platiquemos version), I think the lessons could very easily be Assimil-ized. This would require recording the actual dialogues in a real dialogue way like in Assimil and putting the text together in a parallel format. That's pretty much all that would take to turn FSI into an Assimil course.

It would be almost idiot proof to go through FSI in an Assimil like format and then do the traditional drilling version. The learner would already be comfortable with the vocabulary and grammar thanks to the "Assimil" type wave. Or, the learner could to the "Assimil" format of the lesson and then, the next day, do the drilling section of the same lesson. The flexibility of the course would be great.

Although I don't think you are really asking much about the text, I have to say that I thought the reading part of the course was the least useful. It probably would not be too hard to use the FSI content to create a resource similar to Sandberg's Spanish for Reading. That would add another level of usefulness and flexibility.

Anyway, good luck with your endeavors and thanks again for making this amazing resource available to so many of us.

[luke]

There is Hard Core FSI technique:

I discovered that Audacity, the free sound editing software has a "Compress Silence" under the "Truncate Silence" option. This is found under the Effects menu. The default "compress silence" is to remove 50% of it, so a 10 second "silence" becomes 5 seconds. "Truncate silence", on the other hand, completely eliminates any silence over the threshold. So, in this example, a 10 second silence becomes 1/2 second.

In doing FSI drills over the years, I've noticed that when I really have a drill down, I can repeat the response twice in the pause. When I really felt on top of things, I could also shadow the response and repeat it again two more times in the following pause. That's overkill and hard on the voice.

This new discovery is a middle ground between repeating a drill that you have down pretty well, which takes some time that might could be used more productively. Also, since you've upped the ante, it's more challenging.

The theory behind this is that FSI drills are designed to promote automaticity and proper habit. By reducing the think time, automaticity is fostered. You really have to have it down cold. Also, by making the drills shorter, you have more study time.

[Ericounet]

hi,

some weeks ago, I put on a gitlab server (https://git.yojik.net) to centralize my work (and others) around digitizing. I put there all my Russian FSI/DLI work as well as a little part of German FSI and a lot more of Chinese FSI/DLI.

If you don't want to register, just hit the "explore" button at the bottom of the home page. (yes, it needs to be more visible )

I recently put CSV files , from the first Chinese FSI modules, and csv file from the DLI Chinese glossary. I added the hanzi part which is not in the FSI courses.

I import them in Anki, but every flash program can import csv files. If you want, I can put my Anki decks on the git server.
I added audio from http://shtooka.net to the HSK1 word list, so, if someone is interested, I'll put on the git server too.

It's a free git website, but on one of my home servers (upload speed (so download for you) is not very good, but usable).

Feel free to use it for your own works.

I just found out that https://chinese.remembr.it/ do the same work as me (for Chinese) ... so much waste of resources ...

About funding the digitizing work: It will be a very huge (and expensive) work: the OCR programs just don't recognize the stress marks for Russian, for example, and I had to enter them by hand .

In Chinese, the FineReader (OCR) program doesn't know the pinyin language, so, everything has to be entered by hand .

And all the texts have to be checked by native readers .... my wife is Russian native, so it was easier for me. (for Russian)

For Chinese, I hope I can get my teacher (payed teacher) to check it .. we'll see.

I'm not sure that the funding program could work: people interesting in learning have the courses already ... and use them. Only people interesting in the future use of the courses would be interested, and I don't think there are many. And most of them want to use them commercially or as a introduction to payed courses...

I asked for help (just proof checking of my work) and didn't receive any help: no one remark, correction.

So, yes, I'm not very optimistic ... maybe because I become old ....

I'll though continue my work ...

Best regards,

Eric

ps: some people like SpeakEasy, Abadzi and others helped me to put new materials on the website, correct some bad links , and I want to thank them again.

ps2: I just checked the site https://chinese.remembr.it/ and they do exactly what I wrote above: the course is a introduction/ad to a payed/commercial part ....

[Tomás]

What sort of human labor would be involved in digitizing the course books?

[Ericounet]

hi

1) using a ORC software: I use FineReader. Finding an OCR program which could deal with mixed language texts was not easy. It's the best I could find many years ago (I tested some of them, when they offered free trial period). Maybe, there is another one (and affordable for my purse) ....

2) converting the result in whatever editable format you want, better an inter operable format : so no Microsoft stuff.
I use simple text, csv and Docbook format: the Docbook format is an xml format: you can add the language used for parts like this (this is a part of the DLI Chinese-glossary https://git.yojik.net):

<row>
<entry><foreignphrase xml:lang="cmn-Latn-pinyin">Alasīyā</foreignphrase></entry>
<entry><foreignphrase xml:lang="cmn-Hani">阿拉斯加</foreignphrase></entry>
<entry>N:</entry>
<entry>Alaska</entry>
<entry>L27</entry>
</row>

You can see the "foreignphrase part".
It's structured (uses tags like html).
Xml is easy to use and to transform; you can extract information (here, for example, all foreign parts), transform it (for example -> pdf, LaTeX, doc, simple text, html or xhtml , etc ...); it's easy to edit: every text editor is suitable (and there are many specialized editors/ide on every OS).

3) complete what was not done by the OCR software: stressing marks (not part of the language (Russian for example: Ка́ма ), or romanization (like pinyin). Add missing parts (some pages are missing). Add useful parts, like Hanzi in the FSI Chinese course (which is missing):

1. Qīnwèn, nǐ zhù zai nǎr? May I ask, where are you staying?
to
1. Qīnwèn, nǐ zhù zai nǎr? 亲吻，你朱在那儿？ May I ask, where are you staying?


4) check the result: you need to know the target language very well .. or find someone to do it.

5) transform the docbook file into what you want. Generate csv files (for vocabulary for example, or sentences)

6 publish and it share the result .. the hard part: people having worked hard on all the precedent parts want to make money for it. So many people wrote to me that they "already" done this job for a language they are learning, usualy parts of documents, but never mind .. never shared anything. When I asked them why they don't do it (I offer the server space for it), it's like a black hole: no answer.

7) what would be perfect: correct the document: language evolved since the building of the courses: I learned today, that Xiăojiě, which meant "Mademoiselle" is now a word to say "prostitute" .. so the documents need actualization . It would not be correct to let people learn words, sentences which could put them in funny or weird situations.

8) what would be the paradise ... make new recordings with high sound quality and tagging (we did this for vocabulary in the shtooka project : http://shtooka.net, based on frequency lists)

Conclusion: so, you see, a LOT of work ....

[hulk___smash]

Hi Eric,

Thanks again for your great contribution to the FSI courses. I'm checking out your server now.

I completely understand your skepticism. There is a huge amount of work to do and people are just not offering to chip in. I understand that, i've offered to pay people from my own pocket out here (in Colombia and Brazil) to do some of the courses digitization for me...they just gave up without asking for payment. It is very tedious...but that doesn't leave me undeterred in the slightest.

I have almost completed one of the smaller courses, From Spanish to Portuguese, which is the most in demand course on my site.

Once this is complete i'll put up the prototype of the application to show what is possible.

Now, the ability to digitize every single course is obviously humongous. It'll take a good couple of years, no doubt. But I am to go for the low hanging fruit to start off with. A handful of in-demand courses, as well as less in-demand smaller courses.

I have to monetize it somehow to get paid help for the other courses. Whether I do that via KickStarter, Patreon or just paid services...I don't know. I don't want to charge for anything.

I've a lot of development work to do. Only last week I took a week off my paid job (I'm a freelancer) to soldier on developing the application.

I can easily bring in words from Forvo's API. I can easily export the data for use in more customized applications. Customising output of Ankidecks etc. There is a lot that can be done.

I will make this work.

Darren

[hulk___smash]

Guys,

I've put a video/screencast up on the homepage of my website to show you how the application is looking thus far:
http://www.fsi-language-courses.net

There is still a lot of work to do in terms of development. Development time is all I worried about at the moment, the digitization can be streamlined by a team once I find funding somehow.

I hope you like it and let me know what additions you'd like to see.

Darren

[Elenia]

Guys,

I've put a video/screencast up on the homepage of my website to show you how the application is looking thus far:
http://www.fsi-language-courses.net

There is still a lot of work to do in terms of development. Development time is all I worried about at the moment, the digitization can be streamlined by a team once I find funding somehow.

I hope you like it and let me know what additions you'd like to see.

Darren

Looks good so far. It would be great to have some kind of search functionality, with audio. So, for example, you could find all instances of the word 'Tag' and play the sentences without having to navigate away from that screen. But that would be a bauble, really.Thanks for the hard work!