m EDRA P Attanasio G Marangoni P Mazzucchi

m. EDRA P. Attanasio, G. Marangoni, P. Mazzucchi, N. Mezzetti, G. Scipione A semeiotic view of DOI applications General approach and concrete experiences Washington, 22 June 2007 www. medra. org Handle System Workshop

Summary Ø What do we mean as “semeiotic approach” Ø Relations between DOI and other identifiers: § The actionable ISBN (aka book-land DOI) Ø How a semeiotic approach influences the Multiple resolution development www. medra. org § Menu based vs interim page § The MR value chain

What do we mean as “semeiotic approach” Ø Users are familiar with systems, which are characterised – apart technicalities – by signs Ø Such signs are used both by machines and by people Ø For instance, we can tell people: § “The DOI is an opaque string: please, do not infer anything from that” § But we cannot avoid that people read strings giving the meaning they like (and instruct machine to read strings in unexpected ways) www. medra. org Ø Having a semeiotic approach simply means to try to consider signs as an opportunity instead of a problem

We may be not familiar with the semeiotic vocabulary www. medra. org Ø Sign = Signifier (e. g. the ISBN number per se) + signified (what is intended to mean) Ø Signification = the process of decoding signs passing from the signifier to the signified, thus creating meaning Ø Code = system of signs Ø Denotation = the first order of signification: the primary intention of signification Ø Connotation = second order of signification, introduced by the actuation of meanings deeply rooted in one social environment § e. g. : one ISBN is a product identifier for a certain book (denotation) § some people frequently signify the ISBN saying: “since there is an ISBN, that product is a real book and not grey literature” (connotation)

But we are familiar with many semeiotic phenomena Ø www. medra. org = 130. 186. 85. 120 § Using the first was a fantastic semeiotic idea o That also generated interesting connotations, i. e. second signified coming from the same signifier (e. g. when we speak about “dot-coms”) Ø However, the URL syntax become less semeiotic-effective when you use URLs such as: www. medra. org § www. publishing-watch. org/documents/PMW-o-2004081006%20 European%20 Book%20 Publishing%20 Revised%20 Final%2010%20 August%2020 04. pdf § That’s for: Publishing Market Watch – Sectoral Report 2: Book Publishing, by Turku University and Rightscom, Turku (Fi), August 2004, Ø The use of the DOI may have also a semeiotic value: § http: //dx. doi. org/10. 1234/whateversuffix or simply doi: 10. 1234/whateversuffix § Think how grateful editors will be when preparing a citation for a

First application of the concept: Actionable ISBN or Bookland DOI Ø All people in the book trade are familiar with this set of signs § The ISBN number on the top § The barcode (only for machine) § The GS 1 number on the bottom Ø They are “signifiers” that carry some “signified”, which become "meaning" within contexts of use www. medra. org § The publishing prefix of an ISBN does not mean that book is currently under the control of that publisher, just may suggest it. . . § Any number is opaque, in some extent, but is interpreted by humans that accept some risk of misunderstanding Ø There is a need to maintain the ISBN as “primary identifier” of text monographs (books), exploiting all “denotations” and “connotations” associated to the ISBN signs

Actionable ISBN project: the syntax Ø Syntax combining DOI and ISBN (a brilliant idea by Julian Sowa of Nielsen Book. Data) Ø E. g. : doi: 10. 978. 8807/701689 Number reserved (together with 979) by CNRI to the ISBN community 10. 978. 8807 www. medra. org DOI prefix Number identifying that string as a DOI in the Handle System / 701689 suffix ISBN publisher prefix ISBN book identifier (including check digit, without hyphenation)

Benefits ØAll users (publishers, agencies, bibliographic DBs, web-based systems, libraries, etc. ) only need one number-string (the ISBN’s) to manage a corresponding DOI ØConsequences www. medra. org § Development of new applications for the book supply-chain is made easier § Machine can be easy instructed to obtain DOIs from ISBNs in automatic way § Final users with a book in hand may easy type a valid address for that book

Value of ISBN actionability An example: look at Google Book. Search www. medra. org Users can reach services to buy the book

How such links are made? Ø Links include the ISBN as a parameter in queries to dynamic webpages. The following is the list of the URLs: § http: //www. amazon. com/exec/obidos/ASIN/1740593405 § http: //search. barnesandnoble. com/booksearch/isbninquiry. asp? r=1& ean=1740593405 § http: //www. booksense. com/product/info. jsp? isbn=1740593405 § http: //www. hoepli. it/libro. asp? ib=1740593405 § http: //www. internetbookshop. it/serdsp. asp? isbn=1740593405 § http: //www. libreriauniversitaria. it/BUS/1740593405 www. medra. org Ø In fact ISBN, per se, is not actionable Ø This implies that § Who creates the link has to know the syntax of the queries § The publishers of the cited objects have no rule on that (may just, sometime, opt-out)

Value of using directly actionable ISBNs Ø Using the Actionable ISBNs and multiple resolution, publishers may include any (and all) the links they like, without need to ask any third party to do anything Ø Two key features: § All publishers regardless their size may add links according their needs § Links may be used within any web page Ø We are piloting this using within the VTO (Volltextsuche online) project in Germany, but it is usable: www. medra. org § by any publishers registering “bookland DOIs” § by any webpage, simply using a DOI syntax Ø Multiple resolution is essential for this application

Multiple resolution and semeiotic Ø There are different types of MR (see the Chuck Koscher’s presentation later) § We are moving within the domain of “multiple choice” MR, not considering at this stage “contextual” MR www. medra. org 1. Menu based MR

Multiple resolution and semeiotic Ø There are different types of MR (see the Chuck Koscher’s presentation later) § We are moving within the domain of “multiple choice” MR, not considering at this stage “contextual” MR www. medra. org 2. Interim page MR

How to select one? Ø The first is much better when you consider “look and feel” criteria and usability § Higher usability: one click less to reach the targeted resource § More elegant: users never leave the webpage unless they actually wish to move to another resource www. medra. org Ø Not surprising, larger efforts, in the DOI community, was devoted to the first So: we decided to select the second

Reasons for such a contradictory decision Ø There is a value chain in the use of Multiple Resolution C Final user We page editor e. g. one publisher assigning DOIs to books B DOI-RA Registrant A e. g. one newspaper website citing that book Note that the user is a reader of B and not of A www. medra. org Ø The registration agency should consider § All the needs along the value chain § The semiotic codes used by the different players

Consequences Ø A cannot say to B: “please, cite my book using this new code” § This is necessary with menu based MR, where the webpage must be created with additional, not usual, scripts Ø A must say to B: “please, cite my book using this value within the code you are familiar with” § All the webpage editors need to cite is a URL (a DOI incorporated in a proxy service) www. medra. org Ø There is not only a matter of simplifying the work for generating pages with very beautiful menu based multiple resolution Ø It is also a matter that you have to convince people to learn another code § And that code has already some connotation: e. g. “it is referred to journal articles: why to use for books or others” § While an actionable ISBN brings more appropriate connotations, as well

A second problem: what should the MR contain? Ø MR interim page (and menu as well) are based on: § Taxonomy of target resources § Labels to be shown to users Ø Also in this case: much better to start from what users are more familiar with www. medra. org § So we started from ONIX relevant code list § The final objective is to include such taxonomy within the ONIX for DOI registration schema § But having ONIX-like codes immediately have some semeiotic value Ø Once again: we should be aware that we bring in the pilot denotations and connotations associated to ONIX Ø Moreover: separating the list of target resource types and labels facilitate multilingualism § (a problem we immediately have between Italy and Germany)

What we show in our interim page Very sort explanation to final user One button to view complete metadata www. medra. org The actual multiple resolution Few metadata (planned: author, title, publisher and © year)

Finally: one unexpected semeiotic lesson in our experience Ø When we decided to implement interim pages, we would like to have more elegant one Ø We decided to produce pop up windows because (denotation) § They are more similar to menu § You can have the previous page visible on the back Ø However, pop up windows have now negative connotation within the web environment Ø So that, many browsers block them in default configuration www. medra. org Ø We have just decided to turn back to normal pages!

Thank you for your attention Questions and information Ø piero. attanasio@medra. org Technical aspects Ø gabriella. scipione@cineca. it Ø nicola. mezzetti@medra. org www. medra. org Genuine semeiotic experts Ø giulia. marangoni@medra. org Ø paola. mazzucchi@medra. org