Re: lynx-dev gettext yes/no (was: structured description of lynx.cfg set

[Leonid Pauzner]

>      * From: Klaus Weide <address@hidden>
> On Mon, 3 May 1999, Webmaster Jim wrote:

>> Since these cases depend on first-letter, can we add numbers so that NLS
>> messages can be clearer:
>>
>> Allow? (Y/N/Always/Never) => Allow? (1=Y/2=N/3=Always/4=Never)

> The prompt was already too long to fit much other useful information
> on the same statusline, it is even worse in the translation for languages
> where words are on average longer than English - what's in the de.po
> for the cookie prompts is extremely ugly.  Making a change that takes
> up eight more characters, even for English, makes this only worse
> for everyone.

> I don't see a simple way to generalize the approach for yes/no in
> HTConfirm() to other keys, but there must be one...

In fact, we have a limited number of words we need to confirm for,
probably five to ten (perhaps Cancel, Download, Document, Link, in
addition to four above) so we need not "too generalized" solution.

The problem may come when several reserved words from the same querry begins
with the same letter for some (unknown) language as was pointed by Tom. But
this probably a translator's task to find synonyms or maybe a request for
handling several characters as an input for HTConfirm(). Anyway this is a
raw theory until we fall into a real problem.


>      Klaus




>      * From: Webmaster Jim <address@hidden>
>      _________________________________________________________________

> On Mon, May 03, 1999 at 07:36:57PM -0600, address@hidden wrote:
>> In a recent note, Webmaster Jim said:
>> > #define HEAD_D_L_OR_CANCEL \
>> >  gettext("Send HEAD request for D)ocument or L)ink, or C)ancel? (d,l,c): ")
>> >  gettext("Send HEAD request for 1)Document or 2)Link or 3)Cancel (1,2,3): "
> )
>> <sigh> We're on the threshold of reverting 25 years to the BASIC Startrek
>> game of the Altair S-100 computer:
>>     1) Fire Phasers
>> Why do I not perceive progress here?  But NLS accessibility is why
>> power switches are now labelled "0" and "1", isn't it?  :-)

> To keep the messages short (fitting in 80 columns ;-), I would leave the
> existing "english" answers working. I just couldn't think of a clean way
> to do this:

> Send HEAD request for (1,D)ocument or (2,L)ink or (3,C)ancel (1,2,3,d,l,c):

Adding numbers looks very ugly and really a bad idea IMHO.

> that would make sense when translated to other languages.

> ++++++++++++++++++++++++++++
> Marvin the Paranoid Android.