Talk:LetsType: Difference between revisions

From OLPC
Jump to navigation Jump to search
(Keyboards and IMEs)
 
(11 intermediate revisions by 6 users not shown)
Line 13: Line 13:


There are 30 modern writing systems, mostly alphabetical, but some requiring IMEs, including Chinese hanzi, Japanese kana and kanji, Korean hangeul and hanja, and Ethiopic. I can assist with layouts and elementary exercises, and in some cases with fairly detailed information on the order of lessons.--[[User:Mokurai|Mokurai]] 05:17, 2 April 2008 (EDT)
There are 30 modern writing systems, mostly alphabetical, but some requiring IMEs, including Chinese hanzi, Japanese kana and kanji, Korean hangeul and hanja, and Ethiopic. I can assist with layouts and elementary exercises, and in some cases with fairly detailed information on the order of lessons.--[[User:Mokurai|Mokurai]] 05:17, 2 April 2008 (EDT)

== not porting is an advantage???? ==

Somebody might not understand Open Source development. You do way better if you can share code.

Tux Type is a program that could be ported. It's quite alive, with numerous GSoC projects. You could port that.

* Initially my idea was that only, porting TuxType to XO. But, during IRC discussions I was informed that it can not be ported because its built using SDL which is still not very well supported on OLPC-XO. I'm still open to the idea. Just need to know whether its possible or not. Thanks for the feedback. [[User:Prakhar | Prakhar]] 21:30,13th April 2008 (UTC)

----

* SDL works fine on the XO. I just finished playing with/updating TuxPaint (which uses SDL) and it works fine. Problem with TuxType is that until very recently it assumed 640x480 and the code did not adapt to the current screensize. The current trunk in SVN has come a long way to make this happen. So I am seriously looking at helping getting TuxType to work on the XO. [[User:Tvoverbeek | Ton van Overbeek]] 2008-09-16

== Multiuser Activity ==

Some suggestions on ways that 'LetsType' could be ectended into a multiuse activity.
* Two or more students
* Competitive (eg. like a 'handball game')
** The current 'winner/king' sets the typing activity for the other participants (the serve)
** Other participants then respond as quickly and accurately as they can (the return)
** Players are ranked and the best 'return' then becomes 'king'.
* Instructive - Like competitive, but the Leader is fixed.
** Students only see their own results.
* Competitive with teams (students can arrange thenselves into groups/teams from 2 to number-of-players/2 in size).
* Non-competitive
** Some form of collaborative effort.
[[PaulSchulz]]

== Design Ideas ==

Some good suggestions here. I am a learning designer with two XOs and two XO users (one aged six, one aged four) and have some thoughts for you. If you haven't seen it yet, there is a decent spongebob squarepants typing tutor. The one I have is on CD-ROM. I'll have to dig up the disc to give you more info as I'm sure it's released under some kind of license.

Another suggestion is look at the Sesame Street website. They seem to me to have a solid understanding of the early learning a child has with a computer. They have some really simple exercises like pressing any key to make Elmo giggle, or identifying the letter that makes a certain sound. While your project's goal is more limited, teach basic typing skills, there might be some ideas from Sesame to glean about amount of graphics, interactivity, etc.

I suggest you consider for the design of the first module a simple structure like what is described on the previous page as a bubble with a letter in it. Start with single lower-case letters. I do not know how other language keyboards are defined, but you'll want to start with letters only, not special characters (but I recognize in some languages the letter set is greater than 26 due to the existence of accented characters, etc.)

You can set parameters on when to "pass" a student out of this first level. Typically the best way to do this is performance-based. Use the video game metaphor you're already working with. Pass a student out when they have successfully hit (within a reasonable time frame - 3 seconds is probably right, or maybe 5 seconds. I'd have to observe kids working on keyboards to give you a better range, which I can do) 35 letters, allowing for up to 10 mistakes (so after they've seen a total of 45 letters, 10 of which they missed but 35 of which they hit successfully.) Alternatively, you could choose to pass them after they hit 20 letters in a row correctly. This is a more stringent requirement, it would be harder to hit 20 correct than 35 of 45 correct.

Second level of this module should be two letter combinations. Similar requirement for exit. You can think of these "mistakes" as "lives" ala video game metaphor. You should come up with a way to visually represent to the child how many "lives" they have used up and/or have left. Frown faces would be one option (someone should give input on whether this would be demotivating to kids.) I am trying to think of other options that would be language-agnostic, as my first inclination was to use Xs. Maybe if Tux is used as a mascot, it could be Tux with a surprised expression on his face?

Third level of this module should be simple three-letter combinations that actually form English words. This part of the program would need to be customizable for other languages. These strings should be WORDS, not arbitrary letter combinations. Kids need to learn to type words, not asdf. Now, one caveat - I'm not familiar with the ins and outs of the typing educational world. Perhaps there are strong evidence-based pedagogical reasons to teach kids the asdf jkl; string typing. That's how I learned to type a million (Oh, OK, only about 30) years ago and I'm not permanently scarred. However, I think words are more meaningful for kids. If you want to get fancy, you can have each word associated with a small image, which is loaded with the word and the kid has x many seconds to type it.

Variation on this level to make it more complex would be to reduce the number of seconds the kids have to type the word.

You could continue with this type of model for four, five, six, seven letter words. A child who can accurately type 7 letter words within a restricted time limit is probably an excellent typist. :)

You will need to address the question of whether backup/delete keys will work and how they function within these timed/accuracy modules. It's a personal pet peeve of mine that most online typing tests (not tutors, but speed tests) do not permit backspace to fix typos. I end up creating MORE typos because I have such an ingrained use of the backspace key in my typing (I counted - I used the backspace key more than 5 times in that one sentence alone, not counting the twice I hit it in this parenthetical comment.) But that's a personal bias from an adult who learned to type a long time ago. Your mileage may vary.

I'm happy to talk more about design ideas if these sound like they're going in the direction you need. I'm eager to see more practical young-child oriented software for the XO so I'd like to support you and any other developers as much as I can. Please contact me via my talk page (which I think will work...) Good luck with this project! [[User:KayTi|KayTi]] 00:24, 14 May 2008 (EDT)

Latest revision as of 01:47, 17 September 2008

Hello Everyone! Please give your feedback and comments here. It will help me in analyzing myself and my mistakes. Thanks for participating.

This sounds promising! You mentioned balloons, I think something like that is a good way to do it, you can have them either floating around the screen and they get popped when you hit the key, or you can have them float into a word where you have to pop them left to right (for western typefaces, at least). It would also be nice to be able to see your progress over time, maybe something like charts of words per minute for free floating characters (characters per minute?) and a different series for words, it can be motivating to see how far you've come already!
Also, it seems to me like it would be a good idea to implement support for multiple languages from the start. I'm new to the OLPC myself but I imagine there is a python native and OLPC standard way to do this that isn't too hard, plus most of your stuff (beyond basic characters) would come from dictionaries or word lists anyway. As you said this is a requested app, so if you have multi-language support by the time you bug test (end of the summer your timeline says), then others could just take it and add localizations without worrying about having to test for regressions themselves, getting it out there that much quicker
Bobbypowers 17:59, 28 March 2008 (EDT)


Other keyboard layouts

  • Dvorak keyboard, please
  • Consider adding one-handed keyboard layouts for the disabled

There are 30 modern writing systems, mostly alphabetical, but some requiring IMEs, including Chinese hanzi, Japanese kana and kanji, Korean hangeul and hanja, and Ethiopic. I can assist with layouts and elementary exercises, and in some cases with fairly detailed information on the order of lessons.--Mokurai 05:17, 2 April 2008 (EDT)

not porting is an advantage????

Somebody might not understand Open Source development. You do way better if you can share code.

Tux Type is a program that could be ported. It's quite alive, with numerous GSoC projects. You could port that.

  • Initially my idea was that only, porting TuxType to XO. But, during IRC discussions I was informed that it can not be ported because its built using SDL which is still not very well supported on OLPC-XO. I'm still open to the idea. Just need to know whether its possible or not. Thanks for the feedback. Prakhar 21:30,13th April 2008 (UTC)

  • SDL works fine on the XO. I just finished playing with/updating TuxPaint (which uses SDL) and it works fine. Problem with TuxType is that until very recently it assumed 640x480 and the code did not adapt to the current screensize. The current trunk in SVN has come a long way to make this happen. So I am seriously looking at helping getting TuxType to work on the XO. Ton van Overbeek 2008-09-16

Multiuser Activity

Some suggestions on ways that 'LetsType' could be ectended into a multiuse activity.

  • Two or more students
  • Competitive (eg. like a 'handball game')
    • The current 'winner/king' sets the typing activity for the other participants (the serve)
    • Other participants then respond as quickly and accurately as they can (the return)
    • Players are ranked and the best 'return' then becomes 'king'.
  • Instructive - Like competitive, but the Leader is fixed.
    • Students only see their own results.
  • Competitive with teams (students can arrange thenselves into groups/teams from 2 to number-of-players/2 in size).
  • Non-competitive
    • Some form of collaborative effort.

PaulSchulz

Design Ideas

Some good suggestions here. I am a learning designer with two XOs and two XO users (one aged six, one aged four) and have some thoughts for you. If you haven't seen it yet, there is a decent spongebob squarepants typing tutor. The one I have is on CD-ROM. I'll have to dig up the disc to give you more info as I'm sure it's released under some kind of license.

Another suggestion is look at the Sesame Street website. They seem to me to have a solid understanding of the early learning a child has with a computer. They have some really simple exercises like pressing any key to make Elmo giggle, or identifying the letter that makes a certain sound. While your project's goal is more limited, teach basic typing skills, there might be some ideas from Sesame to glean about amount of graphics, interactivity, etc.

I suggest you consider for the design of the first module a simple structure like what is described on the previous page as a bubble with a letter in it. Start with single lower-case letters. I do not know how other language keyboards are defined, but you'll want to start with letters only, not special characters (but I recognize in some languages the letter set is greater than 26 due to the existence of accented characters, etc.)

You can set parameters on when to "pass" a student out of this first level. Typically the best way to do this is performance-based. Use the video game metaphor you're already working with. Pass a student out when they have successfully hit (within a reasonable time frame - 3 seconds is probably right, or maybe 5 seconds. I'd have to observe kids working on keyboards to give you a better range, which I can do) 35 letters, allowing for up to 10 mistakes (so after they've seen a total of 45 letters, 10 of which they missed but 35 of which they hit successfully.) Alternatively, you could choose to pass them after they hit 20 letters in a row correctly. This is a more stringent requirement, it would be harder to hit 20 correct than 35 of 45 correct.

Second level of this module should be two letter combinations. Similar requirement for exit. You can think of these "mistakes" as "lives" ala video game metaphor. You should come up with a way to visually represent to the child how many "lives" they have used up and/or have left. Frown faces would be one option (someone should give input on whether this would be demotivating to kids.) I am trying to think of other options that would be language-agnostic, as my first inclination was to use Xs. Maybe if Tux is used as a mascot, it could be Tux with a surprised expression on his face?

Third level of this module should be simple three-letter combinations that actually form English words. This part of the program would need to be customizable for other languages. These strings should be WORDS, not arbitrary letter combinations. Kids need to learn to type words, not asdf. Now, one caveat - I'm not familiar with the ins and outs of the typing educational world. Perhaps there are strong evidence-based pedagogical reasons to teach kids the asdf jkl; string typing. That's how I learned to type a million (Oh, OK, only about 30) years ago and I'm not permanently scarred. However, I think words are more meaningful for kids. If you want to get fancy, you can have each word associated with a small image, which is loaded with the word and the kid has x many seconds to type it.

Variation on this level to make it more complex would be to reduce the number of seconds the kids have to type the word.

You could continue with this type of model for four, five, six, seven letter words. A child who can accurately type 7 letter words within a restricted time limit is probably an excellent typist. :)

You will need to address the question of whether backup/delete keys will work and how they function within these timed/accuracy modules. It's a personal pet peeve of mine that most online typing tests (not tutors, but speed tests) do not permit backspace to fix typos. I end up creating MORE typos because I have such an ingrained use of the backspace key in my typing (I counted - I used the backspace key more than 5 times in that one sentence alone, not counting the twice I hit it in this parenthetical comment.) But that's a personal bias from an adult who learned to type a long time ago. Your mileage may vary.

I'm happy to talk more about design ideas if these sound like they're going in the direction you need. I'm eager to see more practical young-child oriented software for the XO so I'd like to support you and any other developers as much as I can. Please contact me via my talk page (which I think will work...) Good luck with this project! KayTi 00:24, 14 May 2008 (EDT)