Given the discussion about the length and complexity of the Release
Notes, I thought people might be interested in some statistics.
I downloaded the individual branches for Fedora 7 through to Fedora 11
and ran the German po files through Publican successively (since the
German translation appears to be practically complete in each version).
The results show a general growth in the number of strings from each
release to the next; however, the number of strings requiring
translation in Fedora 11 is actually smaller than the number of strings
requiring translation in Fedora 10 (and the Release Notes for Fedora 11
contain 10% fewer strings overall than the Fedora 10 Release Notes)
Of course, this says nothing about the complexity of the material.
Total strings: 620
Strings unchanged from Fedora 7: 377
Fuzzy translations: 94
Untranslated strings: 149
Strings needing attention from translators: 243
Total strings: 812
Strings unchanged from Fedora 8: 329
Fuzzy translations: 194
Untranslated strings: 289
Strings needing attention from translators: 483
Total strings: 1080
Strings unchanged from Fedora 9: 141
Fuzzy translations: 409
Untranslated strings: 530
Strings needing attention from translators: 939
Total strings: 963
Strings unchanged from Fedora 10: 89
Fuzzy translations: 186
Untranslated strings: 688
Strings needing attention from translators: 874
If anyone would like to see the results of the string matching, just
send me an email. Also, if anyone would like to see the same analysis in
a language other than German, I'm happy to do this for you as well.
PS: If anyone knows of a tool or method to gauge the comparative
complexity of the texts, I'd be very interested to learn about it.