Re: ISSUE-61: R-archiving appears to be out of scope. we must ask christophe, who put it in

Hi all,

Also +1 for not specifying the format, actually ;-)

We can let the archive decide what to accept and what not. This is what we
do at DANS, we accept anything serialized with a w3c standard. If you come
with your own format we will refuse to archive the file.
The rationale is that it will be easier for us to be sure files are still
readable on the long term. There will (always) be tools for parsing these
accepted formats.

Christophe

--
Sent from a device that knows everything about me but is still useless for
typing mails. Sorry for the brievety and typos...
Op 31 okt. 2014 12:36 schreef "Augusto Herrmann" <augusto.herrmann@gmail.com
>:

> Hi all,
>
> while I agree that we shouldn't recommend on a particular format for data
> dumps, there are benefits in recomending an archive format that allows for
> storing data files (in any format) and also metadata along with the data.
> Preserving metadata alongside the data is very important for digital
> preservation. I know of two competing formats that could be useful for that
> purpose:
>
> BagIt
> * IETF draft specification: http://tools.ietf.org/html/draft-kunze-bagit
> * Wikipedia: https://en.wikipedia.org/wiki/BagIt
> * Blog article about it: http://inkdroid.org/journal/2008/06/06/bagit/
>
> Data package
> * Draft specification: http://dataprotocols.org/data-packages/
> * Blog article about it:
> http://blog.okfn.org/2013/04/24/frictionless-data-making-it-radically-easier-to-get-stuff-done-with-data/
>
> Even though they are similar in technical implementation, BagIt is more
> focused on digital preservation and more widely used among digital library
> practitioners (including, for example, the Library of Congess). On the
> other hand, Data package aims at making open data easier to manipulate (the
> so-called "frictioness data") and is more used by open data practitioners
> (including, for example, the Open Knowledge Foundation).
>
> Best regards,
> Augusto Herrmann
>
> On Fri, Oct 31, 2014 at 7:30 AM, Ghislain Atemezing <
> auguste.atemezing@eurecom.fr> wrote:
>
>> Hi all,
>> [Following Max comments ] (*)
>>
>>
>> By the way, I don’t think we should in any case recommend that data be
>> dumped using a particular technology like JSON-LD. If we say anything, we
>> should recommend dumps in a format that is appropriate, widely used and can
>> be expected to be supported over a long timespan.
>>
>>
>> +1. I am also in favor of not recommending a particular serializations
>> for a dumped dataset. Well, although if there is a good resolution for a
>> “suitable format” in publishing data, we can adapt it for the dumps as well.
>>
>> Best,
>> Ghislain
>>
>> (*) Sorry again for not being able to follow all the rich discussions
>> that you had yesterday :(
>>
>
>