- From: Bill Roberts <bill@swirrl.com>
- Date: Thu, 25 Mar 2010 13:48:57 +0000
- To: Danny Ayers <danny.ayers@gmail.com>, semantic-web@w3.org
I worked for many years on issues around long-term preservation of digital material. Danny's point about longevity of identifiers is a good one and it deserves attention, but I can safely say that information expressed as RDF will be far easier to preserve than the vast majority of stuff currently around. It has a well-documented standard syntax and explicit semantics. Even if URIs stop resolving and are replaced by new URIs, at least they can still serve their purpose as unique identifiers and in most cases it would be possible to match old identifiers to new ones where necessary. Contrast this with the current home of a lot of data, inside Excel and Word files where the ability to understand the data is closely linked to a particular application. The features of the semantic web that make it good for present day interoperability between applications also make it good for interoperability between today and unknown future systems. Cheers Bill Roberts On 25 Mar 2010, at 11:38, Danny Ayers wrote: > On 25 March 2010 11:48, Jakub Kotowski <jakubkotowski@gmx.net> wrote: >> Danny Ayers schrieb: >> > The data on the planet now is probably as fragile. >>> >>> What *is* the digital equivalent of parchment? >> >> Digital data definitely is fragile but usually it is seen so because >> recording and playback methods quickly become outdated. There already >> are people looking at this problem: >> >> http://en.wikipedia.org/wiki/Digital_Dark_Age >> >> Perhaps the problem of naming should be brought to their attention. > > Thanks Jakub, interesting material. > > Long-term naming really does seem a bit of a pain - immediate reaction > would be to go for names that are forcefully protocol independent > (i.e. URNs), but that would be ignoring all the goodness of the Web, > and in any case URIs as names are technically protocol-independent. > > Perhaps things will improve when we see more focus on personal > (/agent/business entity etc) -oriented naming directly, with WebIDs > and so on, rather than having to live with the current model where > naming is devolved to 3rd party service providers. > > While Steve Pemberton's approach seems a bit idealistic, utopian even, > I do believe he's hitting the nail on the head in many respects - > > http://homepages.cwi.nl/~steven/vandf/2008.03-website.html > > Cheers, > Danny. > > -- > http://danny.ayers.name >
Received on Thursday, 25 March 2010 13:49:35 UTC