On Fri, 31 Mar 2000, Henry Nelson wrote:
> > that's horrible anyway. Having such fragments in the catalog rather
> > than full messages makes it impossible to make good translations
> > to languages where the word order should be different.
>
> Thank you. I simply do not have the programming skills to point out the
> cause of the problems I was having as a translator.
>
> I think you asked how we could get more people interested in translating.
> Well, fragmentation, the method that seems to have been adopted, is not the
> way to get rid of replications.
I think part of those sentence-fragment entries are jsut remnants from
the past, when most text strings where surrounded with gettext(...) in bulk
initially. Others (probably) came later (for example I have in mind, see
lynx_cfg_infopage() in LYReadCFG.h), but nobody brought the translators'
concerns to the attention of the author of the patch.
> I'm sorry, but _compromises_ will have to
> be made, i.e., strings will have to rewritten to be more inclusive, but of
> course not to the point that they become ambiguous.
Strings should contain whole sentences or messsages as far as possible,
not sentence parts. (I don't see how that would leed to ambiguity.)
> It's a tightwire to
> walk, I know. If no one has the time to do it, then stop mucking with the
> strings.
Did you complain at the point when the messages were changed?
> There's nothing more discouraging to a translator than to have his
> work thrown out the window by someone who isn't even aware of what he's done.
I'm sure there are things more discouraging - like the translation not
being used at all...
But it's part of normal procedure that messages may change. You should
complain about changes for the worse, not changes in general.
> Secondly, as I've mentioned before, the catalogue for Lynx is huge. Since
> there is no indication to the translator what the priority, high-incidence
> strings are, taking some of the bulk out might encourage some people to give
> it a try. I mentioned removing the .lynxrc stuff as a start; I don't think
> that is unreasonable.
Just make it clear that a translator doesn't have to translate everything,
if it needs to be made clearer. "Removing" stuff takes away choice that
is available now (whether to translate or not).
> If I were to press the issue, I'd ask for
> reconsideration of strings used only once, like the message you get when you
> press 'a' for the first time before you have established bookmarks.
I don't think that 'it appears only once' is a good reason for denying
the opportunity to translate it.
> Thirdly, there are quite a few strings which really have no need to be
> translated because they are as easily understood in English as in any other
> language. Again, this would require work on someone's part to list them up
> and get a "sorta" consensus. I'm talking about strings like the following:
> "Gopher Menu"
> "CSO Search Results" |_ probably combinable
> " Search Results" |
> "CSO index" |_ probably combinable
> "This is a searchable index of a CSO database.\n" |
These are mostly strings that appear in antiquated protocol modules;
just leave them untranslated if you want to, or provide the English
text as 'translation'.
Klaus
