Re: [nvda-translations] Romanian Translation

  • From: James Teh <jamie@xxxxxxxxxxxx>
  • To: Punga Dan <dan.punga@xxxxxxxxx>
  • Date: Tue, 22 Mar 2016 02:03:25 +1000

Hi Dan,

Copying nvda-translations so others can benefit.

Replies inline:

On 21/03/2016 8:55 PM, Punga Dan wrote:

Firstly, I don't know how the association works and where it can be found? I thought the system was the repository on Assembla and the mailing list on freelist.org <http://freelist.org>.
There are also automated processes that trigger when you commit to svn and on a schedule. I just need to teach this part of the system who you are. :)

And about the contact email and name, yes these are the ones. The last character in my family name is not an a but an ă (which is character a with a symbol on top that looks like a horizontal left parenthesis - so left paren rotated to the left), but for the purpose of our English-based framework, it's perfectly right.
We can actually put in the ă; the system will handle that just fine.

First issue is about the update process of the userGuide-newRevisions folder in the Romanian, ro repository.
I've recently done the proof reading of the userGuide. In fact I commited the full corrected version just now. What I found is that the ro repository doesn't get fed with the latest changes of the userGuide (so the userGuide-newRevisions folder).
Hmm. I thought I'd set things up so this should happen, but it seems it isn't working. I'll look into it. Sorry for the inconvenience.

Secondly, it's about the symbols.dic and more precisely the options of the file, how to interpret them and what's the best default version of it that should be included in the application.
...
Suppose I want to hear each occurrence of the characters left parenthesis and right parenthesis; doesn't matter if it's by character or standard/line reading. A simple math expression such as
(3 + 1) * 2 = 8
I would want it to be spoken as "left paren three plus one right paren star two equals eight"
Testing the expression above with the default settings for the two parenthesis symbols (which is level most and send to sinth always) doesn't speak the parenthisis symbols when reading a line with numpad8.
Now contrary to what I would expect, by reading the userGuide, or the guide for developers that has more info about the topic, raising the level from most to all doesn't change this behaviour.
The User Guide notes that the level field allows you to "adjust the lowest symbol level at which this symbol should be spoken". The important word here is "lowest". None is lower than some, some is lower than most, most is lower than all. So when you set it to all, you're saying that all is the lowest level at which the symbol will be spoken. Since there's no level lower than all, all is the only level at which it will be spoken. It'd be good to get your thoughts as to how you interpreted this sentence and why this wasn't clear.

And speaking of levels, what would be a concrete difference between some and most, given that none, all and character are self-explanatory?
Unfortunately, there's no good answer to this. Personally, I really dislike symbol levels for this reason; I'd prefer either none or all, but that's unpopular with users. I tend to think of them like this:
* some: This is the default. It's reasonable to assume that new users will be most comfortable with the way a human reads, so level some should try to achieve this. For example, when someone is reading aloud, they might read the plus sign, but they definitely won't read comma or full stop.
* most: Provides more detail than some, but without using words for symbols which are clearly communicated by changes in the voice (pauses, inflections, etc.). For example, we do include quotation marks, but we don't include sentence endings.

Jamie

--
James Teh
Executive Director, NV Access Limited
Ph +61 7 3149 3306
www.nvaccess.org
Facebook: http://www.facebook.com/NVAccess
Twitter: @NVAccess
SIP: jamie@xxxxxxxxxxxx

Other related posts: