On Sep 28, 2009, at 12:37 PM, Lynn W. Taylor wrote: > I see your point. It uncovers one that is IMO, far more important.
Not really ... because the other issue you see is not really an issue and does not lead to repetitive downloading of the exe that will never be used. > So, I wrote on that point, and now I'm at fault because I did not > limit my comments to only what you said? No, those were the rules you say I should adhere to as I understood you ... so, live by the rules you would have others live by ... > Not downloading excess files is a good general principle. There is > a big discussion about code complexity, how it would impact > intermittent crunchers, and the relative "cost" of tracking what is > absolutely needed vs. downloading against future need that needs to > take place before one decides if it should be fixed. > > I do know from four decades of writing code that something can be a > good idea, and add so much code that it isn't worth it. > > But at the end of the day, I don't care about one extra download. > It seems to me that it is a relatively small problem. > > ... and at the end of the day, if I'm not doing the work, and I'm > not paying the bills, then all I can do is make suggestions. > > Having the antivirus swoop in and delete files, triggering an > endless loop of downloads is a *real* problem -- or so it seems to me. It is, and it is out of the scope of BOINC. And, in theory, the code that is detecting the missing files and driving the extra downloads should be removed... thus making the system simpler. And it is not really the cost of the download... I agree it is trivial... it is the fact that each of these downloads was triggering a McAfee event that is the issue. For a novice in BOINC that is security conscious this might be the trigger to have them leave BOINC. We already have a retention problem and one of the concerns that has been mentioned over and over is about security and the possibilities of downloading malicious code... _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
