Vinyl, Digital Content, Data Loss, The Long View
By pmonday on Oct 02, 2006
My personal content problem started several weeks ago when I figured out that I could not purchase CDs or MP3s of some of the groups I listened to when I was a teen in the 80s: Game Theory, Trip Shakespeare, Dali's Car, some OMD albums, etc... As I was wandering around my basement on Saturday, I uncovered my old vinyl albums and there the music was, as well as a solution! Why re-purchase music or even search around the net when I already own it? The problem now was that I didn't have a media reader (otherwise known as a "record player").
On Sunday, I went to Best Buy and purchased a Sony PS-LX250H "Turntable" (record player). My kids were fascinated by it. I also bought a "record cleaner". I hooked the turntable up to the amplifier, and wired the RCA jacks to the Line-in of my soundcard. I then upgraded my Real audio player so I could record MP3s from the Line-in. Off to the races. I converted my Arcadia and Game Theory albums this morning. Cool beans, 2 down and about 15 more to go. Many of my albums were replaced as I purchased CDs, its those nuggets that I wanted to convert. So, maybe in about another month, I will have my content duplicated and be able to conveniently listen to my vinyl content.
Now, interestingly enough, when I say "content duplicated", I actually mean, content copied to an adequate quality. I'm going from an analog signal to 192 Kbps...lots of loss. There is a noticeable difference between CD Quality vs. 192 Kbps and some say there is a noticeable difference as you go down from vinyl analog to CD Quality. I don't notice the Vinyl to CD Quality (personally), but as I listened last night I was struck by some of the warm sounds off the albums. These seem to get lost as you move to digitized 192 Kbps (btw, I can't stand 128 anymore, you lose a lot of the information).
All of this is going somewhere relating to technology and the storage industry, honestly. Abstractly, here is the problem I solved: I have lost content over the years due to lack of availability of a media reader that conformed to the specifications of my stored content. Here is the primary problem with my solution: No matter what technology I use from today, I cannot effectively "copy" my content, I can only re-represent it in today's technologies (losing whatever benefits of the previous technology there may have been).
This problem is actually a problem in data centers. So much so that the practice of "data migration" is a common service. Think about it, this gentleman was thinking about a book "Dark Ages II: When the Digital Data Die" figuring how alarmist it is. He then went on to explain that he often zips his data up to his Zip Backup drive. I realize that Iomega still sells Zip Drives, but seriously...think about this. Zip will die as so many tape formats have died. Got any data on an 8" floppy? Where are you going to go to recover that data? What about on a tape casette that you stored information on from your Commodore or TI-64? These are peculiar in that you aren't dealing with just a hardware solution, you have a format issue in that for the data written, you don't just need the hardware that can read/write the data, you need a particular OS that can read/write the data format.
In special cases, a company should not just store the tape volumes of their backups and archives in a salt mine, but they will also want to store the hardware and software required to read those tape volumes. Ensuring the means to read backup information is available is as important as the content itself.
Now, this all flashed me back to 1999 at JavaOne. The very intriguing keynote was on The Long View, or the Long Now, or something like that. The talk was about building a clock with technology that would transcend the changes in time. How do you assemble a clock so that when it needs maintenance in 500 years, the generation can maintain that clock? Can you build a clock that never wears out? Can you build a clock with Solaris 11 and Sun will maintain backward compatibility from the new Solaris 10,003 version so you just have to load the new version? Will there even be electricity in 500 years, what about 10,000? I don't recall that much of the talk, just that it was one of the first talks that really made me think deeply, what a complex problem.
Going to "The Long Now" web site, you find a quick picture of layers of time that include:
With each of those categories having a longer time to cycle. Now, if you want to build the longest lasting clock, you should rely on stable elements of nature that can be counted on. Radioactive decay could play a big part in that, orbital time, solar energy. Each of these change, but not as fast as the technology that goes into building your Swatch.
To get some perspective, if you are really, truly interested in data storage, backup and archiving. You HAVE to take the long view of things. Ask the hard questions:
- What is the worst case in terms of time for having to retrieve the data I just stored?
- What is the worst case for technology shifts that would make the physical media that the data is stored on useless because the technology to read the data is no longer available?
From there, you are forced to architect a backup solution that may not only include physical media with data, but also physical devices and networking capabilities to read the physical media.
Did you realize that some media (such as CD-ROM and WORM technologies) have only a 5-10 year lifespan whereas many papers have a 500 year lifespan, and at 500 years we are in a crisis and losing valuable information that is a part of our history. The problem is identified in this brochure I just found as I was Googling. What a crazy world when our new technologies often have a shorter lifespan than the oldest technologies (a carving in a rock).