Cori wrote:How would you automate around Wikipedia pages having different URL formats, different page sections ... or simply not existing at all? Not to mention books in languages other than English, which ought to go into that language's Wikipedia if at all possible?
In practice, suggestions for changes to the software are noted for consideration when we have funding available. However, our highest priority is keeping things running as smoothly as they do now (which in itself is not free.) I wouldn't see this as a high priority, not least because there's the risk that by automating it, we'd again fall foul of Wikipedia policy. We don't have the money to do things which are risk-free, let alone exploratory coding.
Yes, it would be great if we could "automagically" update wikipedia and maybe someday we could. What I have in mind is much less ambitions and would only require minor changes to our programs.
This is what I have in mind:
We modify our data structures to include two flags on every completed project. The first flag would indicate whether the link(s) in wikipedia had been completed, a Yes/No wikipedia updated flag. The second flag would indicate if there was a wikipedia page or not.
Once our data contained this info, it would be a small matter to provide access to this info in some way so that whoever wanted to contribute to our wikipedia effort could check and see if the link already existed, and if not, either link to the existing wikipedia page, or, create the page to link to. This second flag would exist as date stamp indicating that on such-and-such a date there was no wikipedia page.
With this information easily available, potential project helpers could quickly discover where there time could be used effectively.
Without such a flag there would be no way to keep track of what was linked on wikipedia and what wasn't, and without such a process in place I think it would be difficult to encourage any LibriVox folks to contribute.
Regarding the programming cost, I can't estimate too closely because I know nothing about the architecture of our system, nor how easy the code is to maintain. Assuming an industry standard regarding these two items, I would estimate it would take less than a week, that is 40 hours, to make the needed modifications.
If we do a cost/benefit analysis, I wonder where it would come out. Our purpose here, I think, is to get people to actually listen to our recordings, rather than just creating them and putting them on the web.
With this in mind I can think of no single thing we could do toward LibriVox's goal that would help us the more than putting us on wikipedia.
And regarding possible issues with wikipedia, we do need to research this. I suspect there will be much less resistance to our desires than there was so long ago.