[sword-devel] How to get vowels in the Arabic bible module?

[Chris Little]

jonathon wrote:
> Chris wrote:
> 
>> We can assume the Word documents are a good source, so proofing is
> 
> I've seen way too many documents that are not good sources, despite
> purporting to be so,to assume that they are good.

Fundamentally, we at CrossWire do not purport to be a text authority and 
we don't just want to make modules. We want to write Bible software (I'm 
stealing that line from Bob at Logos), and part of doing that happens to 
entail building a nice set of resources.

So we do depend on the quality of upstream content providers (same as 
everyone).

>> unnecessary beyond verifying that the Sword module is a good representation of the Word document originals.
> 
> There is a reason why Logos has somebody proof read their
> resources,even when they get clean copy directly from the publisher.

I don't know how consistently they do that. It's great if they do, but 
we should bear in mind that this thread is about an Arabic Bible. We 
don't have the resources to have someone proof the Arabic Bible. I'm 
pretty sure that Logos didn't pay anyone to proof their Arabic Bible 
either. I trust that if they had, they would have corrected their 
spelling of van Dyck's name and maybe given the title as it appears in 
American editions and the LOC.

Beyond Arabic, we have tons of minority language resources that we would 
have no hope of ever finding proofreaders for. Following the requirement 
of finding fluent speakers to do the conversion & proofing would mean we 
would have to ignore these languages and resources forever.

>> The "fluent in the target language" suggestion is just ridiculous.
> 
> Format shifting always introduces errors.  The only issue is how
> severe those errors are.

Always? Certainly not.

Format shifting will almost always lose some information from the 
source, but if it is done skillfully, there should be no loss of 
important information. For example, with the Arabic Bible, of line 
breaks, paragraph breaks, and page breaks, only paragraph breaks are 
retained in the OSIS edition. They're the only ones I assume to have any 
meaning, the others just being the result of rendering on a page. And 
having compared to a page in an early print edition, I know the Word 
docs don't represent the original anyway.

That kind of information loss can be inherent to format shifting, but 
there's no necessity that important information be lost or that errors 
be introduced. When errors do crop up down the line, it's usually 
indicative of some error or inconsistency upstream.

--Chris