I was thinking of cross-referencing names and such things against the verse. Then having a table which would store all the variants for a word (not sure if we'd want that to be by verse, or variants across the whole bible?).
We could easily write something that parses biblical texts looking in particular for verses, and then automatically (maybe with a SOUNDEX or similar functions) tries to match different variations of the words. Then a manual check would have to be done to ensure the automated process is correct...
So for example, say we decide to primarily index against Ramathaim against 1 Sam 1:1
a table called reference_to_variant would perhaps look like
ScriptureRefernce,PrimaryVariant (or an id)
1Sam1.1 - Ramathaim
other_ref - other_place
...
another table would be indexed by primary_variant (or an id)
and have the following rows:
Ramathaim - Ramathaim Ramah
Ramathaim - Ramathaim-zophim
Ramathaim - Rama-thaim
The latter would be populated by a parsing of many biblical texts and would take the first table as its source. So it would go through 1Sam1.1, then other_ref, and so on, looking for the word being the closest match (or matches), and flag uncertainties.
We could also write something to populate the initial table based on our choice of the primary version, say by looking for all words that do not start a sentence but have capital letters, or load up a dictionary, and look for all words that do not match any word contained in the dictionary...
Have I missed the plot? or does that make sense?
Chris
I've got a question about how we're intending to link from the Bible
text to the databases given that different translations may render the
same name differently, and the impact that has on database creation.Here's an example: 1 Sam 1:1 starts (in NRSV)
"There was a certain man of Ramathaim, a Zuphite from the hill country
of Ephraim, ..."We want to be able to link from Ramathaim to the article on Ramah (it's
another name for the same place).Unfortunately, other translations render this name differently.
* KJV, RSV, ESV (and several others) have Ramathaim-zophim.
* NIV follows NRSV (both have Ramathaim-zophim in footnote).
* CEV, GNB, NLT have Ramah.
* NAB (New American Bible) has Rama-thaim
* NET Bible has Ramathaim Zophim (no hyphen)
* Young's Literal Translation has Ramathaim-Zophim (with capital)This is a slightly pathological example since there's both a LXX/MT
disagreement here, and a linguistic problem as to whether "Zophim" is
part of the name or not. NRSV is following LXX, KJV etc MT. However it
illustrates the kind of problem we face. Perhaps we'd also want to make
the same link from the Hebrew and Greek LXX texts.My understanding is that we're intending to make these links by a text
search, which in this case would mean that we have to (manually)
cross-reference the various renderings to the Ramah article. If we're
doing that, then I think we need to decide which translations we're
primarily supporting. Checking the 15 or so I have access to is not
going to be practical! Also, we need to specify whether hyphens and
upper/lower case are significant (I would assume not).Alternative solutions include:
* tagging (semi-automatically but requiring manual input) the Bible
texts in each translation.* linking from the verse rather than the word. This means we only need
a global 1 Sam 1:1 -> Ramah link, but we lose the ability to click on
the word. This will still need manual input. (There's also cases where
one translation will have a given name in a particular verse and another
won't.)Thoughts?
--
Posted By Tyndale STEP Project to Tyndale STEP - Programming on 11/12/2009 06:40:00 AM
No comments:
Post a Comment