Exactly. For one thing, the National Corpus of Polish files were posted
on the Web separately as the source TEI, the relevant schema and the
relevant ODD, but no attempt was made to make them usable *in situ*.
The idea is: download the desired parts of the set, and you can thus
assemble a TEI-compliant (sub)system, on your disk. Only half of this
lunch is free. The microwaving is on you...
On 11/21/2013 07:38 PM, Stuart Yeates wrote:
> Not all of these defects disqualify files from being TEI, and some of them have relative straight forward technical fixes...
> If you commit your script to the github repo, I'll see what I can do.
>> On 22/11/2013, at 3:46, Sebastian Rahtz <[log in to unmask]> wrote:
>> I amused myself looking at these 258 examples and seeing whether they were valid TEI.
>> It is rather sad to say that only about 30% seem to pass this simple test. Some reasons include
>> * TEI P4 not P5
>> * incomplete fragment
>> * no namespace
>> * reference to non-existent DTD
>> as well as the straightforward “wrong against the TEI” ones.
>> What conclusions one can draw is another matter :-}