The subject of archiving is one near and dear to my heart. Call me a data junkie. I hate to throw anything away, even tho the useful nuggets may be outnumbered by the junky ones. Data mining is so much more rewarding when there’s more data to paw through.
And some data is irreplaceable. Who are we to decide what should be kept and what should not? Things change, people change, styles change, values change. What is junk to day might be gold tomorrow. Save it all, I say, and let the future sort it out!
But in a practical sense, we are left with a cost/benefit ratio to decide what to keep. Obviously, there has to be a level at which it is not affordable to keep some stuff. So what is that level?
Adam Osborne once wrote, upon noticing the steady trend towards the cost decline of data storage, that eventually the cost of storing a byte would be so little as to be effectively zero. At that point, he said no one would ever delete files, as the effort required to delete them would exceed the cost to retain them.
While I’m not sure I will go that far, Adam had a point. I have a friend who takes beautiful digital pictures, prints them out, then deletes them from her storage drive. Yet the cost of saving them would be about 25 cents for several hundred (on a CD or DVD) and she is the beneficiary of a doctor’s income, so 25 cents is not hard to come by. Yet she makes not attempt to save them.
This is really foolish. Imagine how valuable these pictures might be to her children’s children’s children a hundred years from now? Imagine what it might be like to have similar data from 200 years before now? Beyond priceless.
Therefore, I urge the SDMB admin team to make every effort to preserve as much data as possible. If logistics make it temporarily difficult to keep data online current, at least store it in the best manner possible for future restoration.
And if the problem is SQL and/or vBulletin’s limitations, the time is long overdue for those programs to be revised to handle a million times more data. The history of computer progress is a continual one of expanding limits; of RAM addressability, hard drive partition size, cluster sizes, and database record limits. Time once again to boost them up.
Make a trend line – show the cost of increasing storage vs needs. At present trends, will next year’s needs be met with next year’s budget? I think it will.