Star Wars had a critical difference. It didn't need to work because it was all a giant bluff. Romney had little or nothing to gain by bluffing.
Back to the facts. The Romney team said the software was running 20-30 minutes behind. And pointed out that, as you suggest, there was no real way to test it in advance. But in that situation, you have to really over- design for scalability. Nick On Sunday, November 11, 2012, Dan Minette wrote: > Nick wrote > > >Sounded like a classic scalability problem. > > I would guess otherwise. This would be an interesting geekish debate to > have. My guess is that its akin to the problem with Star Wars software, > which was assumed to work first time untested. From what I read, their > software did not lend itself to real live testing before election day. So, > it glitched badly, as one would expect the first time in the field. > > My software work has often been with firmare that runs 20,000 feet below > the > surface, with no chance to fix anything once it goes downhole. Field > testing in real wells is essential, even for software that has run > perfectly > without intervention in the lab. It's easy to field test and fix software > that helps field operatives identify and talk with prospective voters > before > the election. If there's a major problem found in Cleveland in July, it > can > be fixed and the fix sent out nationwide in a few days. But, with the > Republicans, if I understand correctly, their software was for election day > only....counting voters off a list and then providing lists of pro-Romney > voters who haven't voted yet. If it glitches on election day, the best > programmers in the world couldn't get the patch out in time. > > That's my guess, anyways. Does anyone else want to play detective. :-) > > Dan M. > > > _______________________________________________ > http://box535.bluehost.com/mailman/listinfo/brin-l_mccmedia.com > >
_______________________________________________ http://box535.bluehost.com/mailman/listinfo/brin-l_mccmedia.com
