[tei-council] parsing odds

[Sebastian Rahtz]

Christian Wittern wrote:

>In trying to do something useful with the P5 sources, I tried to
>inject them into eXist.  Using the scripts on sf did not produce
>anything at all (eXist complains about data not being allowed in the
>prolog of teinames.xml etc.)
>
can you expand on this? what scripts? if there are errors,
lets fix them....

> so I used client.sh to upload the
>sources.
>  
>
look at the Makefile target "split", it makes a set of useable XML files
from each chapter, just for this purpose. thats how the eXist database
used by Roma is peopled.

>To my surprise, I found that 29 of the .odd files would not parse. The
>reason is that they use entities which have not been declared within
>the file (since there is no internal subset in the file, appearently
>they are only ment to be referenced from the driver file). 
>
correct. thats the way it is designed to work. for good or bad.

> One
>solution that has been proposed for this general problem is xinclude
>(a recommendation now since Dec.2004). This allows to have internal
>subsets, but files can nevertheless be referenced from a driver file.  
>  
>
one downside is that it limits you to tools which understand
XInclude. support is by no means universal, I think.

I can't quote chapter and verse, but Lou and I have certainly
argued around this, and experimented, on several occasions;
I know I made the XInclude  argument several times, but
we didn't adopt it.

I think its a red herring, to be honest.

-- 
Sebastian Rahtz      
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

OSS Watch: JISC Open Source Advisory Service
http://www.oss-watch.ac.uk