4 ms·

The first thing that comes to mind is that any inaccuracies or vandalism are going to be stuck on their devices until they make an update.

abhaga 15y ago
Exactly. Whenever I read about one of these projects that are distributing Wikipedia content offline, my first thought is - How do they check all the content for vandalism? They most probably neither have the time nor the expertise. On the WP site, such things get detected and rolled back quickly and that is one of the reasons the whole thing works. Once you take that out, it is shaky territory.
- fusiongyro 15y ago
  They don't curate it at all, they just load snapshots. In practice, it's no shakier than before. Whenever you load a Wikipedia page, there is a certain probability it's vandalized. I'd be surprised if the probability were going down, since as it is more widely used it attracts more trolls. If the probability is staying relatively constant, it doesn't matter if you are looking at the current data or a snapshot from six months ago, the probability of a random page being in a vandalized state is about the same. And of course, if the vandalized page rate is going up, a snapshot would be better than the live site. You could argue that, if you discover a vandalized page, you can't do anything about it on the Wikireader. You could argue that once you find a vandalized page, it's going to stay vandalized until you load another snapshot. But those are consequences of the tradeoff for the size and network independence of the device.
  - abhaga 15y ago
    You are right when considering an individual. When you consider a group of people, the probability that all of them saw the vandalized version on Wikipedia site goes down as the number of people increase. With the offline version, it stays the same. If you consider the case of disadvantaged children whom some of these projects aim to reach, the whole peer group now has the same vandalized information and there may not be a secondary source of information.
    - fusiongyro 15y ago
      That's a good point. Ultimately it's going to come down to risk versus benefit.
swang 15y ago
Here is an idea for vandalism. Find the last non-bot, non-ip edit. Verify that this user has been on Wikipedia for a while, has a talk page with no vandalism warnings (again you'd have to check the history for the last edit made by someone other than the user), and/or has made more than 100 contributions. Won't get everything but should be effective.

abhaga 15y ago

Exactly. Whenever I read about one of these projects that are distributing Wikipedia content offline, my first thought is - How do they check all the content for vandalism? They most probably neither have the time nor the expertise. On the WP site, such things get detected and rolled back quickly and that is one of the reasons the whole thing works. Once you take that out, it is shaky territory.

fusiongyro 15y ago
They don't curate it at all, they just load snapshots. In practice, it's no shakier than before. Whenever you load a Wikipedia page, there is a certain probability it's vandalized. I'd be surprised if the probability were going down, since as it is more widely used it attracts more trolls. If the probability is staying relatively constant, it doesn't matter if you are looking at the current data or a snapshot from six months ago, the probability of a random page being in a vandalized state is about the same. And of course, if the vandalized page rate is going up, a snapshot would be better than the live site. You could argue that, if you discover a vandalized page, you can't do anything about it on the Wikireader. You could argue that once you find a vandalized page, it's going to stay vandalized until you load another snapshot. But those are consequences of the tradeoff for the size and network independence of the device.
- abhaga 15y ago
  You are right when considering an individual. When you consider a group of people, the probability that all of them saw the vandalized version on Wikipedia site goes down as the number of people increase. With the offline version, it stays the same. If you consider the case of disadvantaged children whom some of these projects aim to reach, the whole peer group now has the same vandalized information and there may not be a secondary source of information.
  - fusiongyro 15y ago
    That's a good point. Ultimately it's going to come down to risk versus benefit.

swang 15y ago

Here is an idea for vandalism. Find the last non-bot, non-ip edit. Verify that this user has been on Wikipedia for a while, has a talk page with no vandalism warnings (again you'd have to check the history for the last edit made by someone other than the user), and/or has made more than 100 contributions. Won't get everything but should be effective.