[Dict-common-dev] Re: Including aspell pwli files in the dict-common policy?

[Domenico Andreoli]

On Thu, Jun 27, 2002 at 12:34:19PM +0200, Agust?n Mart?n Domingo wrote:
> 
> Many of such things should come from the aspell people. For instance 
> aspell dicts naming scheme. I think that the widespread use is 
> aspell-language_code. This implies that all variants for the same 
> language should be provided in the same package. Since it is also 
no, why should be this way?

i don't think we should enforce a strict naming scheme, people is still
able to read description of the packagesa. i propose the slighter cheme
aspell-<language_code>[-whatever]

so that a person looking for, say, portuguese he can search for
aspell-pt. he can then find one or more packages which description is
to be read.

we cannot encode everything in the package name nor we want to include
all the variants of a language in the same package. if i'm interested in
it_IT maybe i'm not interested in it_CH. enforcing strict ownership of
the variants to the same package is for sure inefficient. it is better
to not be too picky here.

> desirable that the same ispell dict source be the aspell dict source the 
i'm missing the point here

> naming scheme should be extended to variants having different ispell 
> sources, e.g., portuguese european and portuguese brazilian, coverint 
> not only the language code but also a variant code. Unless you are 
> thinking about a different naming scheme, what I doubt.
> 
indeed :)

> I think we should currently add to the policy requirements all aspell 
> dicts must have. At a later stage we might thing of some sort of helpers 
i think that a naming scheme and the restriction on the pwli files would
be sufficient.

for instance, aspell-fr is completely broken. it doesn't provide a file
named fr-aspell.pwli . i don't really know how it can work.

instead aspell-de, provides de-aspell.pwli but also some other files
like de_DE-40-aspell.pwli that do not seem very "pwli-compliant". this
is maybe due to lack of the correct antries in regions-to-spelling.map,
the upstream tryed (he succeeded for some strange coincidence) the
fancy name.

at the moment the lack of policy cannot make situation worse.


-----[ Domenico Andreoli, aka cavok
 --[ http://filibusta.crema.unimi.it/~cavok/gpgkey.asc
   ---[ 3A0F 2F80 F79C 678A 8936  4FEE 0677 9033 A20E BC50