This is a very complex question and probably has as many answers as
I was reading an article from Scientific American about medical data
security. It appears that medical data without personal identifiers has
been widely circulated for years. It is useful for long term
(longitudinal) studies to determine things like long term effects of
drugs, radiation exposure, etc. While the person is anonymous, the point
was made that through today's data mining techniques, it is relatively
easy to point back to the individual from the comparing identified
records with anonymous records. To be effective, the randomized data
over time needs to be tied to an individual. You can't compare Joe's
1957 and Jim's 2003 data and achieve useful results. You have to compare
Joe's 1957 and Joe's 2003 data to achieve useful results and with
tracking that data over time, even though you do not know who Joe is, by
comparing with other data, you can figure it out.
So, that was the article thesis, but data mining on both sides of this
question (anonymous and named) can create useful societal results.
Part of the challenge in archives, as I understand it, is learning what
is significant and what is chaff. It is difficult to predict what will
be useful in the future.
I will provide an example from my life of tape archiving. When I started
to get heavily into this, I subscribed to many of the same mailing lists
that you did. I found I didn't always know what I would need to know in
the future, so I made a point of archiving all the posts rather than
trying to figure out what was of interest and what was not. I then
decided that there were several lists that had high enough traffic that
I couldn't / shouldn't keep up, but I still kept all the posts. There
have been several instances where some search terms have given me a post
from one of those lists that helped answer a question I was having
perhaps half a decade after the original post.
It is easier and cheaper for me to keep it all (at least in regards to
email) than to spend the time sorting. I spend enough time just sorting
my general inbox every few months to keep it down to under 1000 messages.
On the other hand, I do not keep every project I've done.
From another perspective, my bank used to keep online data for me for
six to eighteen months, depending on account type. They really, really
wanted to stop mailing me paper statements, so one of the perks is that
they keep the data now for seven years--conveniently the time I need to
maintain records for the tax man. If I think of it, I will download my
year's credit card and main checking account information, not so much to
preserve it, I trust the bank to do that better than I can, but rather
now at tax time, I can review all the charges, then search for the
emailed invoice for many and make certain I have all the correct items
to deduct as business expenses. Much faster, and saves space in file
drawers and ultimately on storage shelves. Up until this year, we had
been saving a "book box" or 10-ream paper box worth of paper data for
accounting each year.
Digital images the same: I generally keep most of what I shoot as I
don't always shoot with a specific purpose other than, I LIKE THIS. So,
different versions of the same image may pertain better than others to a
later desire to create.
While still in the 1-2 TB class, we are starting to see our local
historical society's storage needs increase (10 years worth of data was
kept on a 320 GB HDD along with the computer's OS and program files). I
know have a 4 TB RAID-6 NAS unit there as we are adding video interviews
that are part of our historic collection -- and retrospectively
digitizing audio and video.
On 1/28/2016 6:46 AM, Tom Fine wrote:
> Which brings up the bigger issue -- do they need to keep all that data?
> We've just seen what happens when the government "security" forces build
> up a huge haystack -- they missed the needle in San Bernadino. I'm not
> convinced about capturing huge amounts of any data, just to keep it.
> This has been my argument about accumulating vs. collecting and
> archiving. Just because something was put to media doesn't mean it's
> worth preserving. In a world of limited resources, descisions need to be
> made by humans as to what is worthy of the efforts and money involved in
> digital preservation and storage. I would say the same is true of all
> data. In our world, those decisions are by nature aesthetic and
> sometimes political. It's how human culture works, it's constantly
> curating the past and making value judgements, and media archives are
> really just cultural archives. I think all of this will be even harder
> in a generation or so, because so much media is being created and thrown
> online every day. There's no effort involved in the "releasing"
> mechanism anymore -- you just thrown your production online and see if
> it sticks. Your "production" can be an artistic work such as music,
> fictional video, fictional writing, an artfully crafted documentary, or
> it can be pure opinion or noise or something in between. Any of it can
> be "released" through the same filterless, no-cost mechanisms. In the
> pre-Internet days, when mass media was manufactured physical media,
> humans had to decide what projects to fund through the releasing
> mechanisms, so some curating was taking place from the get-go.
> I know, I took this in a whole new direction ...I changed the subject
> line. I'm interested if this line of thought is being addressed in
> archiving circles and in schools where archivists learn their craft. My
> thesis is that creation of data is easier and cheaper than ever, but
> there are still major costs to archiving and storage, and thus more than
> ever we need skilled curators to cut through the noise and garbage and
> preserve what will matter in 100 years. For pre-digital stuff, there are
> even greater preservation costs in time and money, so I think more
> curating is necessary.
> -- Tom Fine
> ----- Original Message ----- From: "Corey Bailey" <[log in to unmask]>
> To: <[log in to unmask]>
> Sent: Wednesday, January 27, 2016 8:57 PM
> Subject: Re: [ARSCLIST] LTO vs HDD
>> Hi Tom,
>> The answer is relatively simple: Money
>> You and I think about storage in terms of a Terabyte or two. General
>> Motors and corporations of that size have to think in terms of
>> multiple Peta-bytes. LTO becomes the least expensive method. After the
>> data is on the tape, verification and migration is done robotically.
>> Those that are considering LTO need to know that the format (drives,
>> etc.) is only backward compatible for two generations and LTO-7 is on
>> the horizon.
>> Corey Bailey Audio Engineering
>> On 1/27/2016 4:36 PM, Tom Fine wrote:
>>> Could someone explain why a somewhat antiquated magnetic tape-based
>>> storage system is preferable to several copies across several hard
>>> drives? I just can't see any sense in using tape systems anymore for
>>> data security, but I'm not a computer-storage expert, just a guy who
>>> stores a lot of data.
>>> -- Tom Fine
>>> ----- Original Message ----- From: "Hood, Mark" <[log in to unmask]>
>>> To: <[log in to unmask]>
>>> Sent: Wednesday, January 27, 2016 6:41 PM
>>> Subject: Re: [ARSCLIST] LTO vs HDD
>>> Hi Richard,
>>> Thanks as always for sharing your experience and insights on all of
>>> Would you be comfortable sharing the make and model of the RAID-6 NAS
>>> units you are using, and any comments about how well they have performed
>>> to your expectations?
>>> Mark Hood
>>> Associate Professor of Music
>>> Department of Recording Arts
>>> IU Jacobs School of Music
>>> On 1/27/16, 3:36 PM, "Association for Recorded Sound Discussion List on
>>> behalf of Richard L. Hess" <[log in to unmask] on behalf of
>>> [log in to unmask]> wrote:
>>>> Hi, All,
>>>> I saw this thread and was going to ignore it, but decided not to once I
>>>> found out that RDX was HDD-in-an-otterbox merci, Henri, and thanks for
>>>> the image, Lou. Otters are wonderful--see "Ring of Bright Water" (The
>>>> book) and Point Lobos State Park.
>>>> LTO was around while I was still doing broadcast consulting and, at the
>>>> time (late 1990s, early 2000s).
>>>> I struggled long and hard about how to store things and realized if I
>>>> were going to become involved with LTO, I would need two drives (how
>>>> else can you be even remotely certain that your tapes are readable once
>>>> your single drive dies--I certainly saw that in the early days of PC
>>>> tape backup. At that point, the cost becomes excessive.
>>>> My philosophy now is: Any data I want to keep does not live solely on a
>>>> I have two in-house RAID-6 NAS units, one backing up the other; an
>>>> case of 2.5-inch HDDs off-site (2 TB 2.5-inch USB 3.0 drives are pretty
>>>> economical these days and are USB-powered).
>>>> One son has been migrated to the cloud where Dropbox backs up and
>>>> mirrors his two on-site laptops. Here, I harvest all new files (but not
>>>> updates to prevent pollution of existing files) and store them on my
>>>> RAID-6 NAS units to protect against a Dropbox failure or hacking. The
>>>> other son will do it soon, but the first one is potentially going far
>>>> away to school next fall for his Masters (Wichita and Edmonton are on
>>>> the list) so I wanted to get some closer-in history with the system.
>>>> RAID-6 allows the failure of any two disks without losing data and the
>>>> data does not have to be chopped up into 1 or 2 TB chunks as it does
>>>> with HDDs.
>>>> I do not keep CF/SD cards, I copy and verify the copy and then recycle
>>>> Richard L. Hess email: [log in to unmask]
>>>> Aurora, Ontario, Canada 647 479 2800
>>>> Quality tape transfers -- even from hard-to-play tapes.
Richard L. Hess email: [log in to unmask]
Aurora, Ontario, Canada 647 479 2800
Quality tape transfers -- even from hard-to-play tapes.