On Wed, 18 Jul 2001, James wrote:
Any comments on i18n POD?
My feeling is there should usually be
at least English POD. Sounds like a lot
of work to update alternate languages as well.
There have been recent queries on p5p regarding translations of fixed pod
sets (e.g. those that came with perl 5.6.1) to other languages. I think
it is widely supported as a "good thing" but it may be unlikely that the
translated pods would be included in the perl tar ball in future versions
- so as to keep tar ball size down and to avoid maintenance problems with
languages that the perl tar ball maintainers are not qualified to keep up
That said I can offer this bit of advice: for widest distribution try
to restrict your pod to to the 7 bit ASCII char set. I found out that
with the 8 bit ISO-8859-1 chars in perlebcdic.pod that there are some
*roff implementations that do not grok the Latin-1 Char set well (nroff on
locale C Solaris 2.7 being one, the GNU version of nroff on OS/390 being
another). Note that the pod spec allows for html inspired E<code> names
for the printable Latin-1 characters but little else.
Having said that I think that the question actually pertains not to wide
distribution of pod, rather to narrow distribution of pod. E.g. using
scandanavian pod with an appropriate char set. Given my experience with
the *roff implementations I guess I would recommend testing things out
with whatever pod2* tools you intend to use. A limited example of this
would be writing pod for translation only to html: you might make more
liberal use of the L<> construct than you would if the pod needed to go
both to pod2man and pod2html. Likewise if I knew that the Linux
implementation of pod2finnish could easily grok ISO-<mumble> then I might
not care at all if the nroff on Solaris 2.7 does not handle the
ISO-<mumble> char set all that well.
I think that maintaining a list of such cross-system incompatabilities
would be as daunting a job as, say, specifying which characters in
Mac-Roman do not map well to the latest rendition of Windows codepage
1252; which is to say a combinitorial explosion that would be difficult to
verify for accuracy.
Bjoern Hoehrmann wrote:
Something I _really_ miss, is a section about i18n in POD, e.g. how
should one encode my name in some .pod document and how is this method
supported across Terminals/Tools/etc.
Great work anyway :-)