Re: [lmi] [PATCH] fix InputSequenceEditor error messages

[Vadim Zeitlin]

On Sun, 07 Aug 2011 17:09:05 +0000 Greg Chicares <address@hidden> wrote:

GC> On 2011-07-12 14:31Z, Vaclav Slavik wrote:
GC> > 
GC> > please apply this simple patch on top of the others.
GC> > It fixes diagnostics messages to use human numbering instead
GC> > of zero-based.
GC> 
GC> Actually, lmi uses zero-based ordinals elsewhere, eg, on the "Reports" tab:
GC>   Column 0
GC>   Column 1
GC>   ...
GC> (and implicitly, I would say, within the input-sequence grammar), so my
GC> preference would be to maintain consistent usage (and therefore not apply
GC> this patch).

 Hello,

 I agree that consistency is good but I've never seen 0-based numbering in
any grid-like control in the user interface (even programmer-oriented
ones). Are you sure you don't want to change the column numbering to start
from 1 as well?

GC> But if wx has a function to transform an int to an ordinal
GC> string that cannot be confused with a cardinal number, eg:
GC>   "Invalid keyword \"%s\" on %s row.", value.c_str(), OrdinalOf(row + 1));
GC>   row=0 --> "1st"
GC>   row=1 --> "2nd"
GC>   ...
GC> then I wouldn't object to using that. (The high-scoring answer here:
GC>   
http://stackoverflow.com/questions/69262/is-there-an-easy-way-in-net-to-get-st-nd-rd-and-th-endings-for-numbers
GC> appears to be correct for English.)

 Yes, and English is good enough for LMI but not for wxWidgets in general
and it's all but impossible to implement it for all languages in practice
(perhaps we could provide it for the languages we have translations for by
reusing the information about plural forms in the message catalogs but I'm
not sure if it's really possible and it wouldn't be enough anyhow). So I
believe the best would be to add this function to LMI itself.

 However even without it I still think that using 1-based numbering is the
right thing to do. Using "error in row 0" is just too machinocentric.

 Regards,
VZ