KTouch lecture contains a swear word

[evenjn]

Binary package hint: ktouch

I am using ubuntu 10.4.
ktouch version: 4:4.4.2-0ubuntu2

I was following a lecture to learn Dvorak (Training > Default Lectures .. > Dvorak (auto-generated))
Among the various English and pseudo-English words KTouch asked me to type, I encountered "cunt".
I imagine it could be a source of trouble if Ubuntu is ever used at school to teach children how to type.

[evenjn (evenjn)]

It was level 8. I think that it is possible to reproduce this bug by following the lecture at level 8. The words start to repeat after a while.

[Ryan Macnish (nisshh)]

I don't think this bug is relevant, the word does not appear to be used in a swearing context. Therefore i'm marking this as invalid.

[Michael Hall (mhall119)]

If this word appears on a computer used in a school in the USA, the software would almost certainly be banned from use immediately and permanently. The word has no legitimate use, it is a vulgarism, and should not be included in any educational packages.

[Tom Hoffman (tom-hoffman)]

While there is a slippery slope here -- where do you stop? -- in the US this is considered about the most offensive word there is (moreso, I gather, than in the UK).

Its appearance could throw a whole class of 10 year olds into complete disarray, and if it showed up, say, in initial testing would probably disqualify the product entirely.

[Jonathan Carter (jonathan)]

Please keep tis bug report open. I agree that swear words is an old-fashioned concept that's becomming less relevant to young people today. However, Tom is right, there's still currently plenty of good reasons why that word shouldn't be included in the dictionary. I can certainly imagine it spinning a typical classroom out of control.

This isn't a case of censorship or being 'old people', it's a case of what's appropriate. We also don't want to have warnings slapped on Edubuntu for parental advisory, etc.

So, even though the swear word idea is kind of outdated, please don't close it, it's certainly valid!

[Tom Hoffman (tom-hoffman)]

Also, I don't know how this application actually works, if this word is explicitly included in the dictionary of practice words, just take it out. If it is occasionally randomly generated and would mean starting a new blacklisted list of all possibly objectionable words in each locale, I can understand why that would be an annoying path to start down (but it is still probably necessary).

[Harald Sitter (apachelogger)]

As I understand it the word is randomly created, hence only blacklisting is an option and as far as blacklisting goes...

a) requires a framework to do this

b) requires translators to actually know all and ever swear word that appears in their language *and* *every* associated/derived language that uses the same translations *and* *every* language that might be spoken by a considerable amount of people in the language's area (or at least languages of which common swear words might be appearing in that area)
 ...e.g. the german translation would have to contain all swear words appearing in standard german *and* all german dialects (including swiss german) *and* english *and* french *and* danish *and* italian *and* ...

simply put, you would need one large list that contains *every* swear word on this planet, after all it might also be that a chinese swear word appears in a school in the US and a chinese student reads it as a swear word -> pretty much same outcome as if it were english

c) finally a simple isequal comparision would not suffice but a regex is needed, since a swear word might also appear as part of another word (i.e. in german to form new words you combine existing ones without seperating whitespace)

so we are talking about thousands of words (and phrases for that matter) that have to be blacklisted by means of (*expression*), and of those thousands of expressions some might be 3 or maybe just 2 letters long, rendering the pool of possible letter combinations pointlessly narrow

[Alkis Georgopoulos (alkisg)]

Grepping the kdeedu source for that word showed 3 occurrences of it on kdeedu-4.4.2/kiten/data/edict. I don't know if that file is actually used by ktouch, but maybe the word should be banned from there too.

[evenjn (evenjn)]

I don't want to seem overly optimistic, but I can reproduce the bug fairly easily on my computer. That means that whenever i run the said lecture at the mentioned level 8, after a few (~30) words, I come across the said word. This is reproducible across sessions - i.e. I reboot and try again and it's still there. If the content of my lecture has been automatically generated, my experience suggests that it is not created on-the-run. If someone else can reproduce this bug, the chances are that it was automatically generated by the developers of KTouch, once and for all, and therefore there would be hope that the solution is within reach of human capabilities.

[evenjn (evenjn)]

Well, it could also be that the lecture is generated on-the-run out of the words in the dictionary Alkis mentioned. In that case, I guess we are out of luck.

[Harald Sitter (apachelogger)]

Right then.

on #9: kiten is a japanese reference/study/dictionary tool, I would very much appreciate not proposing censorship of dictionaries, especially when they have nothing to do with ktouch, thanks.

on #10: I looked into this and while the words are autogenerated (from aspell dicts) the training files are static distributed, which makes this very fixible, if not forever, but at least until someone feels the need to regenerate the training files.

Fixed upstream, will land in maverick with KDE 4.5.1
http://websvn.kde.org/?view=revision&revision=1167745

[evenjn (evenjn)]

I just want to mention an open source "bad" content filtering project (http://www.poesia-filter.org).