Half the words in KTouch aren't real words or are incomplete

[Mackenzie Morgan]

Binary package hint: ktouch

I've just been playing with KTouch, and it seems there are a lot of words being mangled. Either the ends of words are being chopped up, or there's some German mixed into the English lectures. Examples:
salesladie
latencie
se
distinctivel
syllabicatin
amazo
fizze
pygm
summariz
metaphysicall
psychoanalyticall
heavyse
fraye
mestizo
yashmak
yukk
hygromete
glamoriziatio
synthesi

I'm using Hardy.

[Harald Sitter (apachelogger)]

Does this still happen in KDE 4?

[Mackenzie Morgan (maco.m)]

Yes, this still happens in KDE4 using the English auto-generated lesson.

[Harald Sitter (apachelogger)]

If you have time, it would very kind of you, if you would report this at http://bugs.kde.org and paste the report's URL here. Usually it takes quite long until one of the team has time to forward bugs ;-)

[Mackenzie Morgan (maco.m)]

Bug reported upstream

[Mackenzie Morgan (maco.m)]

Version: (using KDE 4.0.5)
Installed from: Ubuntu Packages
OS: Linux

Many of the words in the upper levels of the auto-generated English lessons are not real words, are misspelled, or are incomplete words. Often they look like German.
Examples:
salesladie
latencie
se
distinctivel
syllabicatin
amazo
fizze
pygm
summariz
metaphysicall
psychoanalyticall
heavyse
fraye
mestizo
yashmak
yukk
hygromete
glamoriziatio
synthesi

The expected behaviour is that actual words are used for spelling practice.

[Mackenzie Morgan (maco.m)]

Er, I meant to say "typing practice" but oops. The point is, practicing mistyping words just makes it habit, which is the opposite of the point.

[Mackenzie Morgan (maco.m)]

Also, this bug is in both KTouch for KDE4 and the older KDE 3.5.9 version.

[Anne-Marie Mahfouf (annma)]

Not sure what to do about it, the file should be all redone but I am not qualified to do so.

[Jiri Palecek (jpalecek)]

I see a similar issue in the Czech dictionary, where some of the words are German. Also, some of the words are Czech dirty words (eg. prdele and proser).

[Josh Leverette (coder543)]

mestizo is a word that describes a person of mixed descent (mix of European and American Native bloods), just thought i would throw that out there, but i agree with the rest of this

[Mackenzie Morgan (maco.m)]

Learn something new every day :)

[Trash-block (trash-block)]

In addition a lot of inappropriate words are included for an educational product: Words such as ASS, JACKASS, SCANDAL, CATCALL, ENEMA, CRACKHEAD, HORNY, SATANICAL. Some are misspelled, but it is obvious both by sound and the large quantity of inappropriate words.

[Anne-Marie Mahfouf (annma)]

If I remove the English file there will be no English training anymore which makes the program useless for English.
Isn't there a replacement or a patch for it instead? Same for Czech?

[Jarl Arntzen (jarl-arntzen)]

The same applies for the Norwegian file. My 9 year old son started complaining about the incomplete words in lesson 4 or 5, I think. Like others have mentioned, both because of the incomplete and the inappropriate words (both in terms of content and relevance), I somewhat regret having enthusiastically recommended Ktouch to his primary school.

Also, it makes the windows-based applications look so much better, since they make you type real words that are more relevant and fun for pupils at a primary school.

Now, since editing the training .xml-files is pretty easy, I'll simply go through and modify it by hand before submitting a patch. :)

[Anne-Marie Mahfouf (annma)]

Thanks Jarl! You can attach the new file in this bugreport or send it to me. I encourage the other reporters to do the same with other files which have problem, thanks. annma

[Jarl Arntzen (jarl-arntzen)]

Created attachment 53949
Added 300 words, corrected incomplete words, removed inappropriate words

- Added 300 most frequently searched words from the dictionary at sprakrad.no (Norwegian Language Advisory).
- Completed incomplete words.
- Replaced inappropriate words with better words. Ktouch can now be safely recommended to Norwegian primary schools.
- Spellchecked entire file.

[Jarl Arntzen (jarl-arntzen)]

BTW: The Danish training file seems really good with regards to use in schools or by students.

Not only does it cover punctuation and upper-case letters but it has some interesting facts-based lessons where you can learn about topics like the ice-age and practice speed-typing at the same time. I'll consider translating the whole thing to Norwegian.

[Anne-Marie Mahfouf (annma)]

SVN commit 1202668 by annma:

improved Norwegian training file, thanks a lot Jarl!!! I'll backport to 4.5 branch as well
CCBUG=166071

 M +217 -174 norwegian.ktouch.xml

WebSVN link: http://websvn.kde.org/?view=rev&revision=1202668

[Anne-Marie Mahfouf (annma)]

SVN commit 1202670 by annma:

backport of 1202668
CCBUG=166071

 M +217 -174 norwegian.ktouch.xml

WebSVN link: http://websvn.kde.org/?view=rev&revision=1202670

[Josh Leverette (coder543)]

well... anybody have any luck getting this fixed? i haven't checked ktouch in awhile

[Jarl Arntzen (jarl-arntzen)]

I've just _manually_ updated the Norwegian language file ( https://bugs.kde.org/show_bug.cgi?id=166071 ) but it was pretty quick and only took me about 3 hours to research and write up.
I think the result works quite well and I'd like to challenge others to do the same for their language.

BTW: Glancing casually at the Danish file, it seems that may be the gold standard in terms of educational value with not only very pedagogically correct lesson progression but also large sections of fact-based texts that actually teach you something about the world as you type. Maybe simply run it through Google translate and work manually from there in your language?

I havn't tried the scripts but since they are based on feeding in extracted lists from dictionaries such as aspell, it shouldn't be too difficult to tweak them to output comlete words only as long as you have enough short input words that cover the earlier lessons.