Re: Adding spellchecker stats to existing d-i stats

[Davide Viti <zinosat@tiscali.it>]

Il Fri, 28 Apr 2006 07:20:14 +0200, Clytie Siddall ha scritto:

> 
> On 28/04/2006, at 9:09 AM, Davide Viti wrote:
> 
>> I'd be interested in adding infos obtained with the spellchecker to
>> the statistics pages, so that everytime a new error is spotted  
>> translators
>> know about it.
> 
> It's a very useful suggestion, but you said in your previous mail  
> that because Alioth runs Sarge, there is no aspell-vi installed, I'm  
> assuming because Sarge did not support Vietnamese.
> 
> Can we install on that machine aspell-vi, and the aspell dictionaries  
> available for other newly-supported and soon-to-be-supported  
> languages, so we can review them more effectively as part of the D-I  
> stats?

I'm not sure Alioth admins would install backported packages to a
production machine.
I think the most important things translators should be automatically 
notified about are not typos, but things like wrong/missing variables,
broken multiselect entries and such: things which badly break the
installer.
As for spotting typos, you can easily download the "vi_all.txt" file which
is updated everyday around 11 UTC and spellcheck it locally on your 
mahine using your favourite spellchecker (not necessarily aspell).
If you want to obtain the list of unknown words as it would appear in the
spellchecker page, you can run the following:

aspell list --lang=vi --encoding=utf-8 < vi_all.txt | sort | uniq -c >
unkn_vi.txt

hopr this helps,
Davide