November, 8th 2009
We need a place to keep our notes related to KDE Educational software and Anne Marie would like to write me about them, therefore I thought it might be useful to have a page with subpages that is part of the project itself where I add thoughts, wishes and problems.
First of all: now we have to work with what is there and possible and we do need users that give us feedback.
The first thing we need to consider is: you can have all the neatest software in the world and maybe even have it localised by one persone, if nobody can really use it, it is as if it was not there. Our target are less resourced languages and with them many things are simply different. Well, the fact that they are less resourced does not mean they are not relevant: the represent approx. 80 % of the market. Many rules we use for the big relatives cannot be applied.
Like I said in Localisation the other way round, the first thing we need is contents, that is educational contents. And we must create them in such a way that they can be re-used in multiple ways. For now all approaches are going through English as the "mayor layer", and for many things we can go that way, BUT we should keep in mind what Martin Benjamin and Kevin Scannell tell us in the talk they had with Bèrto the other day. A moment will come, and I suppose it is also near, when we need special solutions for the most particular languages of the world that also reside in the poorest regions. One of our first approaches is Ambaradan, because it shall work without needing a server, allowing for contents distribution in a very different way to what we are used today.
I talked with Bèrto about my problems on "quantity of files" and frequent releases, needing also bug reporting in terms of data problems from users.
Uploading single files is not an option - let's do some calculations: right now we have terms for Parley in 27 languages, but only publish the data of 22, because the other 5 still have too many bugs and need first some more entries and proof reading.
This means that one language produces 21 files combining it with the other. Now I am not going to create all 21 possibilities, but for now I will concentrate to have English, German, Spanish, French, Italian, Russian, Portuguese and Finnish. this means 21 x 8 right now only for Parley.
Then let's add KHangMan and Kanagram: We have 38 categories by now (tendence growing 50% of the data is still not tagged) and let's assume 15 useful categories per language.
It is too much to be uploaded manually in particular during this phase where there are still bugs around. It could also well be that some categories become more finegraned like: animals of Africa, animals of Northern Europe etc. ... it's impossible to digest.
And now we are talking 20 languages - and what when we start to talk 200?
In this phase we will create a repository with Mercurial where we store all these files and then create projects where people can write down their bugs. It is very relevant for me to get in touch with those people who report bugs: they are potential contributors.
I would like to reach the stage where I can say: "Thanks for reporting this bug. Since there could be something else, would you mind to have a look at ... and eventually correct what you find wrong". In the work we do relationships are really relevant.
Only really stable versions should go to KDE-files. That is versions where you know for sure that you will not touch them again soon. I uploaded the other Parley files there and we have to upload some new versions and some additional data, to keep people interested in what is going on and maybe starting to contribute themselves, but is is going to be a reduced amount, because nobody of us really has the time to do things. For now KHangMan and Kanagram files cannot be uploaded there, so sending them to the maintainers telling them to upload is just double work that is not needed. Get new hot stuff is great: but for now it works only for the mayor languages I suppose. Having things in various places, connecting them and giving links to people is an advantage for all projects.
In this world right now the normal conversation about the aid of developement is considering the subjects of medicine and infrastucture, sometimes about the aliment and then maybe about the education. In this case I would like to make a statement about the relevancy of the native tongues.
When we launch new versions about the programs those are mainly in the major spoken languages. This is causing a problem with less-resourced languages. Many times a minor speaker starts to underestimate their own native tongue, when it is not visual in everyday life. But if a language is seen by own eyes in common situations, it is ”proofed” to be good for education etc. This is a reason, why it is necessary to make games in the less resourced languages.
I find it funny, that a person who is fully native and total speaker of Diola Fogny told me a while ago, that I am the first person he knows that is writing his native tongue. And we are talking here about an educated man working in Europe with university degree.
We, my husband Ismaila Sané and I, made main work for collect the first material for KDE Khangman in Dyo (Diola Fogny). We have been testing the game by our own children and now we are giving other Diola people the chance to try the game. They are all thrilled! This is the first time ever a Diola person can write and play with computer in his/hers own native tongue. I trust that this is a way to make many things come true.
With the game a child can study the alphabets, learn the right spelling, get more vocabulary if a child of an emmigrant, get computing skills while playing, get a new way of learning by solving the problems mathematically etc.
It is important to a child to learn by its own native tongue. That's why we must go on with evaluating the games and programs not only as codes and tricks – also languagewise.
If we get now a newer version of program X in the major language of one country. It is good – for a while. BUT if there is not version made also for smaller languages, the minor speakers are starting to use the major language pack. This will head to the situation, that the future generations are feeling the minor language unnecessary and will change their own native tongue by the time.
We have to lower the digital divide and vanish the inequality in the means of the working and educating systems. We can make is easy – just by creating the games in the minority native tongues!
Well, the wider your offer, the more potential users, but also here we need to consider that the biggest part of the planet has special requirements.
Children like graphics, stories, and games with "levels". My two are those who tell me what they would like to see.
Let me take Tux as an example:
Like Tux collects fish and gets points it could also collect correct words and destroy the wrong ones. It can pick the right equation results out of the sea (fish having the results on their side or back). Finished all 10 equations you go to the next level.
You can have various bags and collect words having to put them in the right bag.
There are many, many possibilities.
One thing to consider is the mobile sector - I just found this link:
which leads to:
The funny thing is: I actually thought about Nokia to talk to for mobile phone applications for the educational sector. Now I don't have the time to read up, but: this is really something to consider. In many countries people don't have money for a computer, but they buy mobile phones. It is one of the reasons why all our i-iter installations like http://dyo.i-iter.org also have a .mobi version:
More about the i-iter network can be found here:
So whenever you write applications: keep the mobile phone sector in mind - on the majority of the market it is the future. This was also confirmed during the last conference where we presented our projects: http://agis09.org .
What IMHO is also relevant is that the langauge for the application's UI can be changed without needing to have the locale supported by the desktop.
Now if you read the article about Localization the other way round one thing is clear: it is not relevant that the whole Desktop is localized, it is not immediately relevant that the software is immediately localized, but it is relevant that the contents are there, that they can easily be translated, created and distributed. All consuming less time possible: there are never enough people to do all that is needed, we will never make it to get all, but we can reach as much as possible by using tools, sometimes very simple tools.
Why KDE with Parley, KHangMan, Kanagram? Well they all can use data that was created outside the software itself and they can share the very same data repository. From the repository you extract the data by query and from there you convert it into the needed .kvtml format. For now we can do this for Parley using a script by Burkhard Kück to kvtml1 format having also "lessons". I tried to create the file for kvtml2 and reached also the stage that a file without categories can be created for kvtml2, but not being a programmer I was not able to change the behaviour in such a way that the category part works with kvtml2.
The step from Parley to KHangMan and Kanagram in kvtml 2 format is not "far", so Iwill try to look at this myself as well and if it works: good, if not: I will need help.
Gwenael Casaccio sent me a first script in Smalltalk, which is the language we would prefer for the reason that it allows for fast applications in the mobile phone sector. I understood a good part of the code he sent me, but was not able to actually change the way it works. The logics is very different and with someone teaching me I would probably learn quite easily how to do things myself.
But seriously: for now I need to concentrate on project management and data management. We do have translators, we do have langauge enthusiasts and I hope we will have some people caring about code that helps to get data in the right fashion over time.
At this stage I actually don't know if there is further KDE-Software where one can add data outside the file that needs to be localized for the UI, but if there is: please let me know. Knowing will not mean that Iwill start to immediately work for it, it simply means: I'll start to integrate as soon as I see the chance to do things trying always to minimize time expenditure.
Consider that what we are doing for Parley for example for Udmurt (udm), Joola Fonyi (dyo), Veps (vep) also leads to a dictionary for these languages and with some tweaks can lead to a spell checker. And if we are really lucky and can get Part Of Speech tagging, then also Apertium is not so far away.
All three of them can be used for many additional exercises in terms of geography, science, history. For elementary school designs like in KHangMan and Kanagram are great. I will see if we can maybe use some pictures my kids made and transform them into background images. Why not involve kids in that?
Parley in terms of design is a bit too adult, so a "nicer" interface would be great, or maybe really different software for kids using parley code, but for example limited to one kind of exercise: one for multiple choice exercises, one where you need to insert a written answer.
One thing I am missing in parley is the "how to do the exercise" description. But on the other hand: this brings additional text to be translated and could be difficult for some kinds of exercises. So I am for now not sure about this feature, that is: how to implement it. Imagine the situation where you want a child to train country names in a foreign language by doing question/answer. The child is very likely able to aswer questions like "Which is the captial of Germany?" by indicating Berlin, getting used to the foreign terminology, but it would not be able to read and understand the instructions in English, but only in the native language.
When it comes to which users we address the answer goes from very generic to very specific:
1) Whoever is interested in learning a language or contents in a specific (mainly less resourced) language in any which way is a potential user.
2) Depending from the applications and terminology groups different user groups are addressed:
2.1 ) Primary school age: this is our first main focus. Children that see their mother language written for the first time. Of course we are also producing for the big languages, but the main focus of Vox Humanitatis is languages of less resourced cultures.
2.2 ) University level/teachers: the main problem we have in less resourced cultures is that on one hand we have to get teaching on the way, and so where you start is primary school, but on the other hand you have the situation that teachers don't even know the language they should teach. Therefore they also need to learn their own language, but on a different level, because they need to be able to explain things to children.
2.3 ) Middle and high school level is the intermedium part. Most of what is necessary for them will be there resulting from points 2.1 and 2.2.
Which software for who:
1) Primary school level: the easiest places to start is KHangMan and Kanagram. For Parley the contents creation gets too complex to reach high level results for primary school for now: but in a second stage we will go that way.
2) University level/teachers: Parley - it is in some way the "all-rounder" for adults.
3) Middle- and highschool: a mix of 1 and 2
Sabine Emmy Eller and Outi Sané