[sword-devel] AmTract Encoding

[DM Smith]

On Oct 25, 2007, at 5:19 PM, Eeli Kaikkonen wrote:

> On Fri, 26 Oct 2007, Eeli Kaikkonen wrote:
>> I protest against this very strongly. All I get with Sword library
>> encoding system is problems after problems. Why use "latin1" which is
>> not latin1? Why not use real latin1? Why use latin1 at all? It is  
>> easy
>
> At least all references to "latin1" should be removed and the name of
> the real encoding should be used in all documentation. That would be
> honest. Latin1 is a common synonym for iso8859-1, at least in unix
> world.

This would be my preference as well.

I got worked up about this a couple of years ago. JSword was breaking  
because it assumed that Encoding=Latin1 meant iso8859-1 and not  
cp1252. Being a purist, it really bothered me. I know Microsoft calls  
it Latin1 and many follow, but slapping a wine label on a catsup  
bottle does not turn it into wine.

But it is too late to change Sword now. It is an issue of backward  
compatibility. There is no way to teach older apps to understand an  
Encoding=cp1252. It would only be of value going forward (i.e.  
MinimumVersion=1.5.11).

As to re-encoding the modules into UTF-8. That is possible, but  
doesn't help. Some people keep the older modules around and the  
current software still needs to work with those having Encoding=Latin1.

As I see it all future modules should be encoded as UTF-8. I think  
ASCII is a proper subset of UTF-8. So it should work for those as well.

But Chris gave the solution: On Linux, every module that is not UTF-8  
should use the Sword filter to convert it to UTF-8.

Serving Christ Together,
	DM