[sword-devel] ztext file format...

[Wade Balzer]

I have looked, but I cannot seem to find any information regarding the
structure of the ztext file format.

 

I understand that there are 3 files.   For example, looking at the KJV New
Testament, there are:

 

nt.bzs

nt.bzv

nt.bzz

 

I can only look at so much C++ code before I get lost in what the code is
doing.

 

My goal is to create an open source C# reader from scratch if possible.   I
just need to understand the file format, and how it is compressed, and how
is it indexed. 

 

Perhaps, it is all explained in some header file that I missed, but I hadn't
found it.

 

I understand that the file is compressed in a block size of either verse,
chapter, or book.   How is that determined?

 

Also, when you do a search for, lets say, the word, "God", does the search
uncompress each block and search the uncompressed block?  Or is there an
index that is created for each word?    The reason I ask is because the
search is reasonably fast, and I am curious how it can be fast with the
constant decompression of each block.    I know that in my own tests, the
time to uncompress a raw text bible could take 3-5 seconds depending on the
speed of the computer, not including the time to search for the text which
would take another 2 or 3 seconds.   The SWORD lib could search for "God" in
a compressed Bible and return results nearly instantly.

 

Any help would be appreciated.

 

Wade

 

 

 

 

 

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20100518/be68ab26/attachment-0001.html>