Some of our clients have learners that are accessing content mainly through ship-based or extreme northern hemisphere satellite connections. The higher latency of these connections causes the checknet alert to be thrown when the user's network connection hasn't been dropped. I think it's a good idea to provide a setting (with very strongly worded precautions) allowing admin users to disable the network check. I See Dan's resistance to the change in MDL-48419. We happen to have a couple of high-profile clients for whom occasional loss of data resulting from dropped network connection is preferable to their users having to repeatedly close the alert. Finally, I have to acknowledge some degree of situational irony in submitting this improvement, since I pushed very hard for the implementation of the checknet funct for a very long time (and it has dramatically reduced our SCORM support workload). We would probably disable this setting by default for most clients using the config file, but providing an admin setting allows us to avoid having 15 or 20 sites with their own custom repositories. This issue is fairly common in some rural Alaska villages where online learning is one of the only avenues for professional development/career advancement, and I would hate for this very important functionality to become another obstacle for those learners.

Alternatives I've considered:

• An admin setting allowing admin users to bump up the timeout for the checknet request. (Seems kind of scorched-earth to me, but we could provide a dropdown with preset options so admins couldn't set it really high...)
• Stop throwing alerts after the third (?) or so alert. We've done or due diligence and warned the learner numerous times, if they want to continue, that's up to them.