On 19/08/2011 21:09, Andrew Gray wrote:
On 19 August 2011 17:59, Charles Matthews charles.r.matthews@ntlworld.com wrote:
Hang on - much as I love the DNB, there is huge amount to know before saying spending money on it would be a good use. There are literally hundreds of Google Books keys for DNB volumes, some of which don't work in the UK but do work in the USA, etc.
To quickly follow up, there's also the complicating factor that not all DNB articles are public domain - at least, not outside the US, which isn't an issue for WM proper but is for us! Almost everything in the 1885 publication is *likely* to be PD - unless the contributors were very precocious junior academics - but the DNB included more recent content (supplementary volumes published every ten years or so) and those are a real minefield involving individual author dates. OED do have the metadata to figure this out on a life+70 basis, but it might not be trivial for them.
Indeed. The Wikisource author pages for DNB authors are not (quite) a complete set yet (some issues of identification in under 1% of cases); and not all death dates are therefore present. Some of the authors of the first edition did live on into the 1940s, meaning there are a few years before their articles fall into the public domain. But at least 99% of it is ex-copyright in the UK.
As for the supplements, the 1901 and 1912 supplements are PD in the USA and are therefore included in the Wikisource project. The 1912 authors are going to need more research, and are going to present more of a problem for UK copyright.
(It's also not clear if the versions published on the ODNB website are themselves the "clean" versions - they may have been tidied up, etc, which muddies the waters a little.)
They are apparently the versions from the 1912 edition, though I wouldn't stake my life on that. I believe what happened is that when OUP took over the DNB, during World War I, they must have decided to "freeze" the existing biographies to reprint. The subsequent OUP supplements are out of the picture as not PD.
Do we have a central list of DNB volumes for which WS does not have scans? I'm happy to spend an hour digging through Google Books / OL / etc...
So glad you asked ... it's not a question of not having scans as such. My basic list:
http://en.wikipedia.org/wiki/User:Charles_Matthews/DNB_scans
My finding list to zap Google Books links on enWP to the DNB:
http://en.wikisource.org/wiki/User:Charles_Matthews/DNB_referencing_data
The point being that such links are generally not readable here in the UK, so should be replaced by links to Wikisource versions as we proof-read them.
Working technical list on Wikisource:
http://en.wikisource.org/wiki/Wikisource:WikiProject_DNB/Progress
Outsiders tend to find WS a bit cryptic. A bot posted scans of all 63 volumes of the first edition; but there were several issues (flaws in the posting led to omitted and duplicated pages; some pages were in a very bad, illegible state; of choice of scans in some case the best scan wasn't used). Much has been done, using (a) the list of scans I made, which relates solely to archive.org, and then (b) other scans.
The real issue, however, is that mashups of scans are not easy to use because of the inflexibility of the djvu files used. Replacing one whole scan by another is a technical royal pain - I haven't done it myself, but I have heard the groans from those who do. Fixing one page at a time when it is bad currently is not a way that works (apparently).
The correct response to Rich's suggestion seems to be this: try to find a way of improving djvu handling on Wikisource/Commons, and you will have hit on something of major benefit all round to those working in this area.
Charles