Talk:Offline Projects/Content Development

Namibian Topics

Hi all, not sure if anyone is watching this, but I'll try it anyway:

We want to distribute an offline Wikipedia version to Namibian rural schools. The full story is here: en:User:Pgallert/You_can_help!#You can help creating a special offline Wikipedia for Namibia. We think of a combination of Wikipedia:0.8 plus articles on Namibia.

We discussed which articles should be included and think it would be best to have a complete snap shot of Namibian content, in other words: We would like to have all articles included that are tagged for the WikiProject Namibia.

First question: Is there a possibility to create such a list automatically?

Regarding article versions: There is very little vandalism on Namibian articles, and the few articles that encounter vandalism are all on my watchlist, I think. We feel it would be most effective and least work to just take all current versions of them. After all, we also want to pave the way for learners to become editors, so the occasional mistake in an article could encourage them to start contributing. Not to mention the enormous task to check 2,000 articles for best versions, I am essentially the only experienced editor of our team.

Second question: Are there any objections against this approach? (To take all current versions, and not to check each of them separately)

Third question: Is there an automated way to produce this list? I understand the software needs permalinks, not just links.

Thanks for reading, Best regards, Pgallert (talk) 13:30, 7 September 2012 (UTC)

Hi, Pgallert!

First question: well, yes and no. No easy way on the wiki itself, but a very easy way if you're willing to run a command-line script (or have a friend who is [and you do -- me! :)]). You can use the pywikipediabot tool-set to generate an on-wiki list of articles in a category, recursively (i.e. including all its sub-categories), with a command such as:

python category.py listify -recurse 

Then typing e.g. Namibia (for w:Category:Namibia), then typing a name for a new page under user own userspace, e.g. User:Pgallert/NamibiaList, and the script will create a list like this one. Note that it can take a while, as in half an hour or more; it's not a very efficient script, and some categories have huge trees under them.

Second Question: if it works for you, it works for us! :) Just note that it could theoretically mean some page would be vandalized just before the offline content is generated, and then your collection is stuck with it. It's no problem to fix, except if you've already deployed this offline version somewhere remote, it can be expensive (certainly time-wise) to replace.

Third question: yes, it shouldn't be difficult. Feeding a list to the Permalinks tool would be one way of doing it.

Hope this helps! Asaf Bartov (WMF Grants) talk 22:03, 10 September 2012 (UTC)

Hi Asaf, thanks for the response. Now, a recursive category is not quite the same as a tag for the WikiProject. But our next excursion is on Oct 5, so there is not much time left. If you wouldn't mind, could you produce the list of articles somewhere in my user space? I'm currently essentially without Internet access (this page took 20 minutes to load, and I'll still have to see how to save it), those are the pleasures of living far away from everything.

Before feeding it into the permalink tool, I guess I'll first have to browse the list for odd entries. Thanks so much in advance, Brgds, Pgallert (talk) 07:09, 14 September 2012 (UTC)

Because of time constraints I will just use your list for now. Our firewall does not let python commands through. Unfortunately it is still some manual work, because Permalinks needs comma-separated lists, and many article names have commas. It further breaks with a variety of special characters, and it does not want to accept url encoding, either. Anyway, I'm working on it. --Pgallert (talk) 12:50, 28 September 2012 (UTC)

Partnerships

I sincerely hope that the Wikimedia Foundation (and other movement entities) will spend more resources in getting the content right. I know that there have been some past investment/resources dedicated to improving the software bit, and I hope that someone, somewhere will step up on improving the content. After all, this product will only achieve the desired impact and goal (of increased use/reach or whatever) if it is tailored for the target niche. Getting the right content is a lot of work, but in the end, it's worth it. I would rather you spend 10K on getting the right content rather than spending 1K on deploying/distributing a product that students/teachers won't use. Abbasjnr (talk) 10:58, 23 November 2012 (UTC)