Talk:PubMan Look-Up Services

From MPDLMediaWiki
Jump to navigation Jump to search

General[edit]

if cone (including the researcher portfolio) is intended to be available as software package that others may install, the additional search services integrated in the portfolio would have to be configurable, i guess. so some sort of widget/gadget mechanism/interface would be required.--Robert 12:38, 12 March 2009 (UTC)


Notes regarding WorldCat[edit]

  • the coverage of WorldCat focuses on book materials. Therefore only few appropriate hits may be returned for many [MPG] authors. Please note that Kurt Mehlhorn is a rather untypical example because he created a respectable amount of book publications. Anyway, the database DBPL lists an essentially higher number of publications. --Inga 23:00, 8 March 2009 (UTC)
  • only few Max Planck libraries report their holdings to WorldCat (see OCLC libraries) and therefore, the heading "search for an item in libraries near to you" could raise wrong expectations. --Inga 23:00, 8 March 2009 (UTC)
    • Perhaps the GUI team can check how to handle this (expectation = function). We could think about something like:
Search --Textbox-- in --Selectbox--
Where the selectbox may actually hold the logos of the lookup service, not only the name.--Friederike 08:40, 9 March 2009 (UTC)
Isn't it already something we can make use of Open Search? Search&Exporting service already offers open search integration --Natasa 12:31, 12 March 2009 (UTC)
An OpenSearch interface for Search&Export would provide an interface to PubMan data. I believe the use case considers the "other way round". --Inga 15:44, 13 March 2009 (UTC)
  • potential alternatives: Google Scholar might be a reasonable free alternative to WorldCat. In addition, vLib offers a similar search box service, but focuses on subscribed resources. --Inga 23:00, 8 March 2009 (UTC)
  • --Natasa 12:27, 12 March 2009 (UTC) Wonder why we should include the search box of Worldcat in researcher pages or PubMan pages? To my understanding in researcher portfolio pages we include links to Worldcat and other services that enable author-lookup. When user already uses the look-up link s/he is redirected to Worldcat and makes the search there. I think it makes actually sense to put a link to PubMan search (as normally, reseacher page is not PubMan, but is separate to PubMan). Also the idea of Google Scholar search is fine. VLib is actually also something we can consider for people coming from MPG (only? or all?) - as it will actually demonstrate easy linking of MPDL services.
  • --Ulla 18:35, 15 March 2009 (UTC): As always, we start simple;-): As no single source will match the needs of all MPIS and external partners, I propose to start with WorldCat and google scholar. Would not go for the proposed "Search item in lib near you". Additional sources will come up, has to be considered on GUI (see Rikes remark). In addition, the link to VLIb services (i.e. MPG services) is (unfortunately) old open issue. What do we need for it?
  • --Ulla 09:48, 16 March 2009 (UTC): Another input: As we might have variuous personIDs stored in Cone, (.e.g from KAKEN, ResearcherID, any other), we should think of kind of "dynamic" offer of Look-up in the respective source, depending on the IDs available for the person.

Notes regarding ISI Web of Knowledge[edit]

  • ISI ID - or no ISI ID?. If the article has originally been loaded from ISI, the metadata record may include an ISI identifier which would allow direct linking. Example on eDoc: http://edoc.mpg.de/403034 -> SFX menu offers direct link to corresponding record in WoS. --Inga 17:58, 9 March 2009 (UTC
--Ulla 18:45, 15 March 2009 (UTC): this was my understanding: we offer this for items where we have ISI ID.
Please note that the current specification of the workflow does not state this precondition. If you want to offer a search for the article title (with no regard to ISI ID), I still would suggest to restrict the option to publications which can be expected to be indexed by ISI WoS, i.e. English journal articles.
UPDATE: In April 2009 ISI released its Article Match Retrieval Service which "allows for a real-time lookup of bibliographic metadata such as Digital Object Identifier, author, source title, etc., against the Web of Science database using the institution's subscription entitlements" (see FAQ). This service may be used to offer the service even no ISI ID is known in advance. I updated the description accordingly --Inga 10:39, 15 July 2009 (UTC)
  • the information professionals in the MPG recommend to use the individual databases directly instead of the "All Databases" search. --Inga 18:15, 9 March 2009 (UTC)
  • --Ulla 18:45, 15 March 2009 (UTC) thanks for the hint...useful for GUI to understand more sources to come
  • Regarding "SFX service as being beyond PubMan's possibilities": Please note that PubMan was COInS enabled before, thus users with an appropriate plug-in can connect to their [institutional] link resolver. Anyway, the COinS implementation probably lacks a more accurate specification, PubMan_Display#COinS does not support various genres. --Inga 19:28, 10 March 2009 (UTC)
Please note that COinS are of advantage if a system does not know to which link resolver a specific visitor belongs to - this is very often the case for free available databases and repositories. Subscription based resources can map users to the institution granting the access. Integrating a lookup in an OpenURL resolver registry could bridge the gap. --Inga 15:18, 13 March 2009 (UTC)
  • Integrate ResearcherID Badge into the researcher portfolio if a researcherID id provided for this person. check out here.--Friederike 10:14, 15 July 2009 (UTC)
Wouldn't it be more generic to allow users to specific URLs pointing to user profiles created somewhere else? ISI's researcherID is just one example for various projects targeting "researcher identification" (see http://info.scopus.com/etc/authoridentifier/ and http://www.crossref.org/CrossTech/2007/02/crossref_author_id_meeting.html) --Inga 10:51, 15 July 2009 (UTC)
Of course it should be generic ;). But I think a person identifier will probably consist of a type and a value (like item identifier). And if the type of an identifier is a ISI ResearcherID we could display the ResearcherID badge. It anyway should be possible to enter whatever identifier someone likes.--Friederike 11:19, 15 July 2009 (UTC)
ya, but than you require to implement and maintain resource specific logics/syntaxes to relate IDs to URIs, badge requests and many more? --Inga 12:51, 15 July 2009 (UTC)
yes that's what I wanted to propose, similar to the item identifiers in the pubman presentation (dois are linked, edoc is linked etc.). sorry for confusion ;)--Friederike 12:57, 15 July 2009 (UTC)

Other ideas for researcher portfolio[edit]

  • any kind of visualization (tag cload, simile timeline) representing the existing information, see OCLC Identities for some ideas

Social Bookmarking[edit]

as far as i know citeulike is far ahead of connotea or bibsonomy in research/social bookmarking. this seems to corroborate it. so maybe before investing in integration with connotea, we might to make sure we can interoperate with citeulike.--Robert 17:59, 9 March 2009 (UTC) --Ulla 18:50, 15 March 2009 (UTC) thanks for the hint...quite convincing. Juliane could you update the resp spec?

  • Future Dev?: Allow a user to twitter a publication (Not sure if this page is the right place for this proposal...).--Friederike 14:52, 16 July 2009 (UTC)

User Interface[edit]

  • Even if we provide more than 2 bookmarking services I would go with displaying the logos, not providing a drop down list. A list of bookmarking logos are quite common use and i would not change it.--Friederike 08:19, 11 March 2009 (UTC)

--Ulla 18:50, 15 March 2009 (UTC)I agree

  • i'm somewhat sceptical about the usefulness of these buttons. as a delicious user myself (and once a connotea user) i never use these buttons, because all these bookmarking services provide bookmarklets or browser plugins to make posting easy - no matter what support the page to be bookmarked offers. i think it's paramount, though, to fix the pubman URLs, i.e. make sure that the URL in the browser's location bar of the item view is the canonical one.--Robert 12:03, 24 March 2009 (UTC)

--Rupert 14:31, 31 March 2009 (UTC) Basicly it's just a list of icons. That will come with a new group in "View Item Version". Thus it can be extended as needed.