- From: Leigh Dodds <leigh.dodds@talis.com>
- Date: Wed, 24 Aug 2011 10:44:01 +0100
- To: Gannon Dick <gannon_dick@yahoo.com>
- Cc: public-lod@w3.org, John Erickson <olyerickson@gmail.com>, patrick@durusau.net
Hi, On 23 August 2011 15:17, Gannon Dick <gannon_dick@yahoo.com> wrote: > Either "Linked Data ecosystem" or "linked data Ecosystem" is a dangerously flawed paradigm, IMHO. You don't "improve" MeSH by > flattening it, for example, it is what it is. Since CAS numbers are not a directed graph, an algorithmic transform to a URI (which *is* a > directed graph) is risks the creation of a "new" irreconcilable taxonomy. For example, Nitrogen is ok to breathe and liquid Nitrogen is a > not very practical way to chill wine. A URI isn't a directed graph. You can use them to build one by making statements though. Setting aside any copyright issues, the CAS identifiers are useful Natural Keys [1]. As they're well deployed, using them to create URIs [2] is sensible as it simplifies the process of linking between datasets [3]. To answer Patrick's question, to help bridging between systems that only use the original literal version, rather than the URIs, then we should ensure that the literal keys are included in the data [4]. These are well deployed patterns and, from my experience, make it really simple and easy to bridge and link between different datasets and systems. Cheers, L. [1]. http://patterns.dataincubator.org/book/natural-keys.html [2]. http://patterns.dataincubator.org/book/patterned-uris.html [3]. http://patterns.dataincubator.org/book/shared-keys.html [4]. http://patterns.dataincubator.org/book/literal-keys.html -- Leigh Dodds Programme Manager, Talis Platform Mobile: 07850 928381 http://kasabi.com http://talis.com Talis Systems Ltd 43 Temple Row Birmingham B2 5LS
Received on Wednesday, 24 August 2011 09:44:28 UTC