You are viewing a plain text version of this content. The canonical link for it is here.
Posted to l10n@openoffice.apache.org by Göran Andersson <go...@init.se> on 2013/05/18 20:26:12 UTC

Updated Swedish Dictionary

An update to the Swedish hunspell dictionary is available at

  http://www.dsso.se/ooo_swedish_dict_2.16.oxt


You should consider replacing the ancient version 1.43 with the new one.
Are there any brave Swedes who wants to test version 2.16?

Re: Updated Swedish Dictionary

Posted by Andrea Pescetti <pe...@apache.org>.
On 22/05/2013 Göran Andersson wrote:
>> 1) Send me the updated dictionary files (if they are the same as in the
>> link you sent earlier, just confirm it and I will take them from there)
> You could use version 2.16 from the link. I will not release any updates
> before the OO 4.0 release unless someone has  suggestions for improvement.

OK, thanks. I've now uploaded a new release of that extension:
http://extensions.openoffice.org/en/node/5959/release

>> 4) I'll update the sources to include the latest version in OpenOffice 4.x.

This is done too:
http://svn.apache.org/viewvc?view=revision&revision=r1486367
Thanks for testing the new dictionaries in 4.x.

> Hunspell v1.3.x, as linked into OpenOffice 3.4.1 and 4.x, is fine!

OK. I've raised the extension requirements to OpenOffice 3.4.1 or later, 
so people won't complain due to limitations of old OpenOffice versions.

> Since you asked, I have two minor wishes though:
> 1) The option "Check words with digits" should be enabled for Swedish. Is
> it possible to enable it by default in the Swedish build? Or at least to
> include a hint about this issue in the documentation for Swedish users?

The place for this would be somewhere under
http://www.openoffice.org/sv/
There's probably a lot to be changed there, including broken layout. 
Feel free to suggest changes to those pages.

> 2) Colon should be included as a word character when checking Swedish text.
> (I understand that this probably won't be implemented in the near future...)

Right... Well, it helps to capture these problems in Bugzilla if you 
can: https://issues.apache.org/ooo/ and flag them as "enhancement". So 
please open an issue about this and, while honestly it's unlikely to see 
it addressed before 4.0, it can get addressed when we have volunteers 
able to work on it.

Regards,
   Andrea.

---------------------------------------------------------------------
To unsubscribe, e-mail: l10n-unsubscribe@openoffice.apache.org
For additional commands, e-mail: l10n-help@openoffice.apache.org


Re: Updated Swedish Dictionary

Posted by Göran Andersson <go...@init.se>.
> 1) Send me the updated dictionary files (if they are the same as in the
> link you sent earlier, just confirm it and I will take them from there)
>

You could use version 2.16 from the link. I will not release any updates
before the OO 4.0 release unless someone has  suggestions for improvement.


> 3) I'd appreciate that a native speaker (like you!) has a look with the
> latest 4.x snapshot, to check that the extension works (I can do it too,
> but I can't judge on the dictionary quality)
>

I've tested the dictionaries (sv_SE and sv_FI) in the 4.x snapshot. Both
work as expected.
Nevertheless, I wouldn't mind if someone could give some feedback on the
linguistic qualities.


> 4) I'll update the sources to include the latest version in OpenOffice 4.x.
>

Thanks!

Nothing forbids from updating Hunspell in OpenOffice 4.x if desirable. We
> just need to be in contact and I'll need some specific test cases and
> feedback.
>

Hunspell v1.3.x, as linked into OpenOffice 3.4.1 and 4.x, is fine! (Well,
it could use some features to enable better suggestions, e.g. ordering the
suggestions according to word frequency.) However, hunspell in Openoffice
3.4 and older has some bugs that sometimes make the dictionary look
retarded.

Since you asked, I have two minor wishes though:

1) The option "Check words with digits" should be enabled for Swedish. Is
it possible to enable it by default in the Swedish build? Or at least to
include a hint about this issue in the documentation for Swedish users?
Test case: the words
   22-nanometertekniken
   33-årigen
should be underlined in red, but not
   22-nanometerstekniken
   33-åringen

2) Colon should be included as a word character when checking Swedish text.
This is just as important as including the apostrophe as a word character
when checking English text.
Test case: the two words
   Volvo:ns
   vd:erna
should be underlined in red (spanning the complete words including the
colon), but not the words
    vd:arna
    dvd:ns
(I understand that this probably won't be implemented in the near future...)

Re: Updated Swedish Dictionary

Posted by Andrea Pescetti <pe...@apache.org>.
On 21/05/2013 Göran Andersson wrote:
>> [Andrea] http://extensions.openoffice.org/project/aoo-dict-sv ?
>> We will take the latest version of that extensions when we package
>> OpenOffice 4.x
> Does this mean that OpenOffice 4.x will include the old and incomplete
> version of my Swedish hunspell dictionary unless someone agrees to be the
> maintainer of the official extension? That would be very unfortunate.

No, it just means that we will take the latest version from there. If 
nobody is maintaining it but you are able to provide updated files to 
include in it, then it's fine and I can continue to maintain it (as 
packager only, always ready to reassign it to a native speaker).

> For several reasons, I will not become the official maintainer any
> OpenOffice extension. Perhaps you could find someone who understands
> Swedish and is willing to test the new version of the dictionary? Then, if
> the tests shows that the upgrade is worthwhile, maybe you, Andrea, could
> update the extension?

Let's proceed this way:
1) Send me the updated dictionary files (if they are the same as in the 
link you sent earlier, just confirm it and I will take them from there)
2) I'll put a new version of 
http://extensions.openoffice.org/project/aoo-dict-sv online
3) I'd appreciate that a native speaker (like you!) has a look with the 
latest 4.x snapshot, to check that the extension works (I can do it too, 
but I can't judge on the dictionary quality)
4) I'll update the sources to include the latest version in OpenOffice 4.x.

> The hunspell version used by
> OpenOffice 3.4.0 and older has a few severe bugs (e.g. resulting in
> malformed words as suggestions) and missing features which are truly
> essential when spellchecking Swedish texts.

This is interesting to know. Nothing forbids from updating Hunspell in 
OpenOffice 4.x if desirable. We just need to be in contact and I'll need 
some specific test cases and feedback.

Regards,
   Andrea.

---------------------------------------------------------------------
To unsubscribe, e-mail: l10n-unsubscribe@openoffice.apache.org
For additional commands, e-mail: l10n-help@openoffice.apache.org


Re: Updated Swedish Dictionary

Posted by Göran Andersson <go...@init.se>.
>
>  You should consider replacing the ancient version 1.43 with the new one.
>>
>
> Would you consider to become the new maintainer of
> http://extensions.openoffice.**org/project/aoo-dict-sv<http://extensions.openoffice.org/project/aoo-dict-sv>?
> We will take the latest version of that extensions when we package
> OpenOffice 4.x
>
>
Does this mean that OpenOffice 4.x will include the old and incomplete
version of my Swedish hunspell dictionary unless someone agrees to be the
maintainer of the official extension? That would be very unfortunate.

For several reasons, I will not become the official maintainer any
OpenOffice extension. Perhaps you could find someone who understands
Swedish and is willing to test the new version of the dictionary? Then, if
the tests shows that the upgrade is worthwhile, maybe you, Andrea, could
update the extension?

Note: My dictionary is used by many other applications. E.g. LibreOffice
uses v2.14 and Google Chrome uses v2.12, so it's not completely untested.
Still, there are some differences in the tokenizers and hunspell versions
used by these applications, so a specific test of the dictionary in Apache
OpenOffice v3.4.1 or later is required. The hunspell version used by
OpenOffice 3.4.0 and older has a few severe bugs (e.g. resulting in
malformed words as suggestions) and missing features which are truly
essential when spellchecking Swedish texts.


>
>  Are there any brave Swedes who wants to test version 2.16?
>>
>
> This is still important if you wish to have feedback about the dictionary
> quality.
>
> Regards,
>   Andrea.
>

Re: Updated Swedish Dictionary

Posted by Andrea Pescetti <pe...@apache.org>.
Göran Andersson wrote:
> An update to the Swedish hunspell dictionary is available at
>    http://www.dsso.se/ooo_swedish_dict_2.16.oxt

Thanks! I confirm installation works fine (means: it can be installed 
and it provides good-looking suggestions, I cannot judge the dictionary 
quality) for me in the latest OpenOffice 4.x snapshot.

> You should consider replacing the ancient version 1.43 with the new one.

Would you consider to become the new maintainer of
http://extensions.openoffice.org/project/aoo-dict-sv ?
We will take the latest version of that extensions when we package 
OpenOffice 4.x. If you are available, feel free to contact me in private 
(pescetti AT apache.org) and we can have the extension reassigned to you.

The only change you would need to make is changing the extension ID in 
description.xml to "org.openoffice.sv.hunspell.dictionaries", for 
consistency.

> Are there any brave Swedes who wants to test version 2.16?

This is still important if you wish to have feedback about the 
dictionary quality.

Regards,
   Andrea.

---------------------------------------------------------------------
To unsubscribe, e-mail: l10n-unsubscribe@openoffice.apache.org
For additional commands, e-mail: l10n-help@openoffice.apache.org