| =============================================================================== |
| De: F Wolff <[email protected]> |
| Per a: [email protected] |
| Assumpte: Re: article |
| Data: Wed, 16 Feb 2011 00:25:15 +0200 (15/02/11 22:25:15) |
| =============================================================================== |
| |
| ~ Ok, this is some notes from my read through the article. It mixes |
| ~ general and specific, large and small issues here, simply so that I know |
| ~ it is done and sent before I run out of time to write it nicely. Sorry |
| ~ if anything sounds blunt or harsh - definitely not intended. It is a |
| ~ nice article, I think. |
| ~ |
| |
| Abstract doesn't say much, but that was just an impression. |
| |
| I think there are more than 5 million Afrikaans speakers. I think we |
| need a better source here. |
| |
| ~The reference to Donaldson looks very strange - it is from 1993 but |
| ~comes after talking about the eleven official languages that only came |
| ~into law in 1996 (if I remember correctly). |
| ~ |
| ~ |
| ~ "parallel corpora for both languages" sound strange. "parallel corpus |
| ~ for the two languages"? |
| ~ |
| ~ http://opus.lingfil.uu.se/KDE4v2.php |
| ~ 404 - should be http://opus.lingfil.uu.se/KDE4.php I guess |
| ~ |
| ~ |
| ~ I guess this is the KDE translations that are pivoted over English? You |
| ~ could do the same for other stuff like Firefox, OpenOffice.org, GNOME, |
| ~ etc. with the big difference being that these Afrikaans translations |
| ~ were done by humans. Most of the KDE stuff is completely unusable for |
| ~ anything, in my opinion (last time I looked, anyway). We have a tool to |
| ~ do that in the Translate Toolkit - poswap. I think this will do that. |
| ~ For an easier ride, you could download the two language databases from |
| ~ open-tran.eu and link them up easily. Lots of data, but includes |
| ~ garbage, like KDE. This is also obviously pivoted, and in domains where |
| ~ I think we might differ more on a lexical level, and even stylistic ones |
| ~ (Dutch uses infinitives for UI buttons, whereas we usually use |
| ~ imperative). |
| ~ |
| ~ |
| ~ Figure 2 is really small to read. |
| ~ |
| ~ The section on number of nouns in Wiktionary is worded in a confusing |
| ~ way. I assume the 10610 is from the _English_ wiktionary? People not |
| ~ familiar with wiktionary might not understand why the Dutch nouns have a |
| ~ corresponding category on the Dutch wiktionary - where else would they |
| ~ be than the Dutch wiktionary (of course _I_ know :-) |
| ~ |
| ~ "After the morphological analysers for both languages were large enough" |
| ~ How do you measure the size of an analyser? Is "coverage" or something |
| ~ else meant? |
| ~ |
| Under 2.2.2 I really miss something about false friends. Van Huyssteen |
| and Pilon mentioned this, and I think it is an important to handle. It |
| is also easy to at least do something about it (like rip theirs). Their "review the list of false friends in .Van... to see if any translations can be improved" |
| license is CC-by, but I guess you can also get a list or two on the web |
| with different licenses, and we can maybe at least try to compile it by |
| hand. I think this is a shortcoming, specifically because it was |
| discussed in the only prior publication in this field. |
| |
| ~2.3.1: I _think_ Afrikaans has word-attached genitives (except that I |
| ~don't know what they are). They are only in archaic forms, I think, so |
| ~maybe you have a reference there? I'm not good with these things, but |
| ~maybe we should just make sure. |
| ~ |
| ~ "have + nie ..." - a bit strange mixing English and Afrikaans here. |
| ~ |
| In general, I find it a bit hard to work out if an example text (usually "in all examples, first line afrikaans, second line dutch,..." |
| nicely italicised) is Dutch or Afrikaans, so I guess an unfamiliar |
| reader might struggle as well. Maybe a convention of appending (af) or |
| (nl) afterwards might help in some cases, although it could clutter |
| things, so it is just an idea. It is probably only important in a few |
| cases. |
| |
| The 3 bullets at the end of 2.3.1 is a bit strange... where do they fit |
| in? |
| |
| ~2.3.2 - of course compounds can contain more than just nouns. Some forms |
| ~with adjectives also combine productively, usually in 3 part compounds. |
| ~ |
| ~ "infrastruktuurontwikkelingsplan" should be hyphenated as |
| ~ "in-fra-struk-tuur-ont-wik-kel-ings-plan" if it is going to be |
| ~ hyphenated. |
| ~ |
| ~ |
| I don't really get compound-L and compound-R, but maybe my background is |
| just lacking here. Do you mean that something like "ontwikkelings" (with |
| the -s-) can only occur on the "left" of a compound? |
| |
| ~ 2.4 needs some text to explain the numbers |
| ~ |
| 3 - the fifth evaluation was not done or reported on later (3.5), so |
| maybe take it out here? |
| |
| Should table 3 still get the "realistic" scores? |
| |
| ~ "Another issue are relative pronouns" - are -> is? |
| ~ |
| ~ |
| ~ I think the English is wrong here: |
| ~ |
| ~ Die sand het dan saam met die water weggespoel. |
| ~ Het zand heeft dan samen met het water weggespoeld. |
| ~ Het zand is dan samen met het water weggespoeld. |
| ~ ‘The sand is washed away along with the water.’ |
| ~ |
| ~ I think the Dutch here is passive past tense, so I think the English |
| ~ should be "The sand was washed away along with the water." A direct |
| ~ translation for the Afrikaans from me would rather be "Then the sand got |
| ~ washed away with the water." or "Then the sand was washed away with the |
| ~ water." but it could depend on context, I guess. |
| ~ |
| ~ Typo: chuncks |
| ~ |
| ~I think the correction to error 9 has more to it. Here the confusion of |
| ~the with-or-without-final '-e' in the adjectives of Afrikaans should be |
| ~taken into account, I think. If it is there, the adjective necessarily |
| ~describes the noun, but if it isn't there and the adjective can |
| ~optionally take it before a noun, it means that the adjective describes |
| ~the following adjective. I'm not sure how that would influence the |
| ~Dutch, but it seems as if the proposed solution might be a bit too |
| ~simple. Example: |
| ~ |
| ~Die onlangs verwagte reën het gekom. |
| ~The recent expected rain have came. (literal) |
| ~The recently expected rain arrived. (+-) |
| ~ |
| ~Here "onlangs" describes "verwagte", not "reën". Not sure if that is |
| ~relevant, though. |
| ~ |
| ~ |
| ~ The hyphenation is wrong for "meewerk". It should be "mee-werk" (even |
| ~ for Dutch, I assume) |
| ~ |
| ~No solution is offered for error 10. I guess it has to do with the way |
| ~the infinitive works in Dutch after "zonder"? |
| ~ |
| ~ |
| ~ "algemeen" not italicised at the end of 3.4.6 (there are some |
| ~ inconsistencies with Afrikaans/Dutch words that aren't always |
| ~ italicised) |
| ~ |
| ~ The rule about "klinkerbotsing" exists in Afrikaans as well. I think we |
| ~ just call it "vokaalopeenhoping", although I immediately knew what |
| ~ "klinkerbotsing" meant. |
| ~ |
| ~ |
| ~"’n Jaar later ruk die situasie in die land onder die indruk van |
| ~massabetogings hand uit." |
| ~I see this in [[af:Duitsland]], but it is a strange sentence, so if you |
| ~can use another example, it might avoid a double take by an Afrikaans |
| ~reader, as I had. The "onder die indruk van" sounds strange to me. |
| |
| 3.5 still outstanding. |
| |
| |
| I hope this helps. This work is really exciting! |
| |
| Friedel |
| =============================================================================== |