[Dxspider-support] a Chinese translation?

[Dirk Koopman]

On Fri, 2006-11-03 at 23:31 +0800, Michael Chen wrote:
> Hi, Dirk and others,
> 
> How to add a new translation to the spider? 
> 
> What should I do besides adding to new entry to /spider/perl/Messages?
> How to make set/language cn working?
> 

There is a more fundamental problem that needs to be solved. Basically
character sets in the cluster world are not well defined. We are
currently using a mixture of various IBM codepages/European ASCII and
ISO 8859-1 (or whatever it really is, can't remember the exact number). 

What we need to do is standardise on the only one that is really viable
- and that is UTF8.

In principle that would not be too hard, provided that I upgrade the
perl requirements to 5.8.1 and do some work with "encode" to change the
messages to UTF8. 

That would move the node to UTF8, you could modify the Messages file,
show/language and set/language appropriately and move on from there. I
suspect then that a load of other little strings would need to be added
to the Messages (or some other) file. 

However, we then have the almost insuperable problem (that already is
showing itself) of the mixture of codesets. We are already getting a
mixture of UTF8 and one (or sometimes more) of the European codesets in
spots - to the extent that we get duplicate spots with slightly
different comments.

So, anyway, these are the basic things that need to be done. It isn't
particularly hard and one could probably do 95% of it by adding a 'use
utf8' in cluster.pl, doing the three files I suggested and then fiddling
with the deduping algorithms until the dupes went away.

Hope that helps

Dirk G1TLH