Hi,
please see my comments below.
Yours,
Per Tunedal
On Mon, Nov 19, 2012, at 23:06, Francis Tyers wrote:
> El dl 19 de 11 de 2012 a les 22:56 +0100, en/na Per Tunedal va escriure:
--snip--
> >
> > On Mon, Nov 19, 2012, at 22:39, Francis Tyers wrote:
> > > El dl 19 de 11 de 2012 a les 22:32 +0100, en/na Per Tunedal va escriure:
--snip--
> > > > On Sun, Nov 18, 2012, at 14:22, Francis Tyers wrote:
> > > > > El dg 18 de 11 de 2012 a les 12:01 +0100, en/na Per Tunedal va
> > > > > escriure:
> > > > > > Hi,
> > > > > > I've started some work on abbreviations for the pair Swedish -
> > > > > > Danish
> > > > > > (sv- da). I imagined that this would be a well-defined, small task
> > > > > > of
> > > > > > limited scope, but it isn't.
> > > > > >
> > > > > > Some questions:
> > > > > >
> > > > > > 1. Internationally accepted standard abbreviations
> > > > > >
> > > > > > I'm talking about abbreviations for countries (e.g. SE for Sweden),
> > > > > > for
> > > > > > languages (e.g. sv for Swedish), for currencies (e.g. SEK for
> > > > > > svenska
> > > > > > kronor - Swedish crowns), for measurement units (e.g. kg or m) and
> > > > > > possible some other domains. These abbreviations are common to all
> > > > > > languages and it seems unnecessary to add them to all language
> > > > > > pairs.
> > > > > > Couldn't they be include by default in Apertium? Now they are
> > > > > > marked as
> > > > > > unknown words and thus get unnecessary attention when post editing.
> > > > >
> > > > > There complications to "including them by default" for example "kg"
> > > > > and
> > > > > "m" are not the same in all languages. For some languages, codes may
> > > > > have unexpected ambiguity with other words, thus decreasing the
> > > > > performance of the tagger. So no, I don't think it's a good idea to
> > > > > include them by default.
> > > > >
> > > > > A good idea, would perhaps be to have a page on the Wiki with a list
> > > > > that people can copy, paste and check. Or a script in the
> > > > > trunk/apertium-tools which autogenerates dictionary entries for common
> > > > > abbreviations which can then be manually checked before being added.
> > > > >
> > > >
> > > > Is this a suitable GCI task?
> > >
> > > Sure, would you like to mentor it ?
> > >
> > Yes, it might be interesting! When do I have to be available?
>
> If you'll be checking your emails once every couple of days from the
> 26th November to the 14th January, you should be fine.
OK. I don't know anything about scripting in Linux, though. But a
Wiki-page would be fine for a start, wouldn't it? Can anyone write in
the Wiki?
>
> > > > > > 2. An abbreviation that hasn't any equivalent in the other language
> > > > > > in
> > > > > > the pair. It has to be translated by an expression.
> > > > > > How to treat them?
> > > > >
> > > > > No idea, give some examples.
> > > >
> > > > Well, you might often abbreviate a word or an expression in one
> > > > language, but never dream of abbreviating in it the other (or be
> > > > reluctant to doing so). Example:
> > > > Danish "ejd." = Swedish "egendom". And further there are phenomenons in
> > > > one country that simple doesn't exist in the other country. e.g. Danish
> > > > "D." = Swedish ("-") (We don't have any "Dannebrogsordenen" in Sweden!)
> > > > or even Danish "AB" = Swedish "-" ( It doesn't exist, I have to use
> > > > the Danish full word "andelsboligforening" or some Swedish related
> > > > phenomena like "bostadsrättsförening" (can be abbreviated to bfr.) or
> > > > "bostadsförening" (very rare, as only a few old ones exist, due to the
> > > > fact that it's illegal to create new ones.)
> > >
> > > This sounds really infrequent.
> > Well, the Danish "AB" is very frequent, and must be added as the Swedish
> > abbreviation "AB" is something completely different (and very frequent).
>
> I didn't really understand your example, can you phrase it better ? What
> is the translation(x) you want, when translating 'AB' to Swedish. What
> is the translation of _that_ word(x) when translating to Danish ?
Danish "AB" = "andelsboligforening" in Danish, and would have to be
translated to "andelsboligforening" in Swedish!
Swedish "AB" = "aktiebolag" in Swedish, and should be translated to
"A/S" (= aktieselskab) in Danish. (It signifies "Ltd.")
>
> > > In that case I might do something like:
> > >
> > > da:
> > >
> > > D.<abbr> (LR)
> > > ejd.<abbr>
> > >
> > > da-sv:
> > >
> > > D.<abbr> = Dannebrogsordenen<abbr> (LR)
> > > ejd.<abbr> = egendom<abbr> (LR)
> > >
> > > sv:
> > >
> > > Dannebrogsordenen<abbr> (RL)
> > > egendom<abbr> (RL)
> > >
> > > This would allow you to translate from Danish->Swedish, but wouldn't
> > > mess up the analysis in Swedish. If you don't understand this, then I
> > > would just leave it until it starts causing serious translation problems
> > > (maybe in 10 years).
> > >
Hmm. Looking once more at your suggestions above I get confused. Swedish
is the left language in the pair, haven't you mixed up the directions?
The Danish word "ejd." should be translated to the Swedish noun
"egendom", but the Swedish noun "egendom" should be translated to the
Danish noun "ejendom" (signifies "property").
Like this:
<e r="RL"><p><l>egendom<s n="abbr"/></l> <r>ejd.<s
n="abbr"/></r></p></e>
<e> <p><l>egendom<s n="n"/></l> <r>ejendom<s
n="n"/></r></p><par n="_ut_ut"/></e>
> >
> > I thought of something like that, but dreaded some trouble if I called a
> > noun for an abbreviation. What about the tagger training for instance?
> > The tagger wont be confused?
> >
> > BTW those RL:s wont trouble your disambiguator, I hope? Any description
> > of it to read?
> > Would it be useful to try it out?
>
> RL in this case means 'generate only'. E.g. you only include the path
>
> egendom:egendom<abbr>
>
> when compiling the generator. In the analyser it won't be included, so
> you'll just get whatever the existing analysis is for egendom.
>
> Fran
>
------------------------------------------------------------------------------
Monitor your physical, virtual and cloud infrastructure from a single
web console. Get in-depth insight into apps, servers, databases, vmware,
SAP, cloud infrastructure, etc. Download 30-day Free Trial.
Pricing starts from $795 for 25 servers or applications!
http://p.sf.net/sfu/zoho_dev2dev_nov
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
