Hmm. I'm a little late to the party here.
On Thu, Jan 28, 2016 at 11:23 AM, Dave Crocker <dhc@xxxxxxxxxxxx> wrote:
> The confusion on this is mixing 'status' with 'availability'. The
> fact that a document is no longer considered active does not mean
> that it should become inaccessible.
That's a great point; in fact, there are digital preservation folks
that champion dark archives, but my experience with them is that
they get full of cruft - when an archive is accessible then omissions,
repetitions, truncations etc. are more likely to be recognized and
corrected.
> Essentially, that reduces to requiring that future researchers act
> now to preserve ephemeral IETF materials that they might wind up
> needing.
>
> That, of course, isn't going to happen.
Typically, digital preservationists require three or more copies, with
checksums, that are independently maintained (different administrators
and software suites), each with institutional plans for how materials
would be migrated if the institution fails.
On Thu, Jan 28, 2016 at 2:31 PM, Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx> wrote:
> In short, just work out ways to preserve the bits. If you keep the
> bits you can be reasonably confident that the technology will be there
> to read them.
Sadly, this is not the case - there's a recent podcast from 'On The
Media' that Interviews Vint Cerf among others about digital
preservation. It's tricky.
You need to keep a bit of metadata, even for simple text files. Here
are some encodings you can run into:
$ iconv -l | wc -w
1173
However, that being said, IETF doesn't really have much in the way of
binary data (word perfect files, spreadsheet data, etc).
For the IETF data,, it's mostly about making sure the bits are periodically
migrated to up-to-date media.
On Thu, Jan 28, 2016 at 6:55 PM, Joe Touch <touch@xxxxxxx> wrote:
> The original authors had the right to prohibit distribution of that
> material, and some (myself included) have asserted that right
> vigorously. I.e., it's not merely a matter of what was "lost" but
> also "who has the right".
There are techniques for this among archivists, they can put embargoes
on the documents they archive.
An embargo can be open-ended, or it can have an explicit expiration date.
On Thu, Jan 28, 2016 at 7:10 PM, Brian E Carpenter <brian.e.carpenter@xxxxxxxxx> wrote:
> What I want to say is this: Archiving too much is just as bad as
> archiving too little, from the future historian's point of view.
>
> That's not to say that we shouldn't consider long-term archiving, but
> an important aspect is deciding what not to archive.
Well, the archivists have this saying: only 10% of what we archive is
ever of any use. But we have no idea which 10%.
You might have language scholars studying this stuff, as well as
the expected technologists and social scientists. Or next century's
version of post-modernists looking at g*d-knows-what.
On Thu, Jan 28, 2016 at 11:23 AM, Dave Crocker <dhc@xxxxxxxxxxxx> wrote:
> The confusion on this is mixing 'status' with 'availability'. The
> fact that a document is no longer considered active does not mean
> that it should become inaccessible.
That's a great point; in fact, there are digital preservation folks
that champion dark archives, but my experience with them is that
they get full of cruft - when an archive is accessible then omissions,
repetitions, truncations etc. are more likely to be recognized and
corrected.
> Essentially, that reduces to requiring that future researchers act
> now to preserve ephemeral IETF materials that they might wind up
> needing.
>
> That, of course, isn't going to happen.
Typically, digital preservationists require three or more copies, with
checksums, that are independently maintained (different administrators
and software suites), each with institutional plans for how materials
would be migrated if the institution fails.
On Thu, Jan 28, 2016 at 2:31 PM, Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx> wrote:
> In short, just work out ways to preserve the bits. If you keep the
> bits you can be reasonably confident that the technology will be there
> to read them.
Sadly, this is not the case - there's a recent podcast from 'On The
Media' that Interviews Vint Cerf among others about digital
preservation. It's tricky.
You need to keep a bit of metadata, even for simple text files. Here
are some encodings you can run into:
$ iconv -l | wc -w
1173
However, that being said, IETF doesn't really have much in the way of
binary data (word perfect files, spreadsheet data, etc).
For the IETF data,, it's mostly about making sure the bits are periodically
migrated to up-to-date media.
On Thu, Jan 28, 2016 at 6:55 PM, Joe Touch <touch@xxxxxxx> wrote:
> The original authors had the right to prohibit distribution of that
> material, and some (myself included) have asserted that right
> vigorously. I.e., it's not merely a matter of what was "lost" but
> also "who has the right".
There are techniques for this among archivists, they can put embargoes
on the documents they archive.
An embargo can be open-ended, or it can have an explicit expiration date.
On Thu, Jan 28, 2016 at 7:10 PM, Brian E Carpenter <brian.e.carpenter@xxxxxxxxx> wrote:
> What I want to say is this: Archiving too much is just as bad as
> archiving too little, from the future historian's point of view.
>
> That's not to say that we shouldn't consider long-term archiving, but
> an important aspect is deciding what not to archive.
Well, the archivists have this saying: only 10% of what we archive is
ever of any use. But we have no idea which 10%.
You might have language scholars studying this stuff, as well as
the expected technologists and social scientists. Or next century's
version of post-modernists looking at g*d-knows-what.
Thanks for your attention.
-Randy Fischer