Home Page -> PerX -> The ugly, the bad and the good

Work in Progress

Below, there is a summary of my personal ideas and opinions on the standard protocols and mechanisms used by the PerX Project. This information is based on my own experience working with data providers, when trying to make them cross searchable by the PerX toolkit. It doesn't reflect the opinion of the PerX Project.

  • The ugly: Z39.50

    Z39.50 is the ugly guy here. I have put a quick presentation of Z39.50 hands-on here, in case you want to see the syntaxes and algorithms used by PerX to query Z39.50 targets. Following, I list the reasons that in my opinion make Z39.50 the ugly of the PerX movie:
    to be continued...

  • The bad: OAI-PMH

    A lot has been written about OAI-PMH. Too many papers, reports and discussions for a protocol that in my opinion is the nightmare of every IT staff maintaining a service that need to cross-search OAI repositories. Discussions on OAI, among ICBL researchers, have been occurred since some time ago. Basically there are a group that advocate the usage of OAI as a mechanism for exposing metadata and think that OAI is a good candidate to enable interoperability. Other group, I mean just me :-) think that OAI is simply bad. I have experimented that actually OAI is an inadequate mechanism to enable interoperability. "Coincidently" I have been discovering that I wasn't the only one "suffering" with the harvesting, normalization, enhancements and indexing of metadata obtained from OAI data providers.

    One important conclusion of PerX should be that the OAI-PMH, as it is currently specified and implemented, it is not meeting its initial purposes and that its validity, suitability and efficient for federated searching are questionable.

    Here is the list of reasons that we think make OAI the bad:
    to be continued...

  • The good: SRU

    In the last three months (Nov-Dec 06 and Jan 07) we have re-written large parts of the PerX toolkit in order to make it a SOA-compliant software and an SRU data provider. The motivations behind this significant revision were the embedding of PerX in a VLE and the use of PerX within a digital library context.

    List of reasons that support our advocacy for the SRU-SOA combination for practical and efficient interoperability:
    to be continued...

  • The OAIster Case

    I think OAIster could not be a suitable target for PerX. Please see below my points:

    1. Usage of OAIster is really low, for such a services launched six years ago: Average of search sessions (visits) in 2006: 547 visits/day. Compare with the current average visit per day of TechXtra: 540 visits/day
      Ref: http://www.cs.cornell.edu/people/simeon/workshops/JCDL2003/JCDL2003_OAI_Workshop_Kat_Hagedorn.pdf
    2. The search results returned by OAIster, in most of the cases, can be irrelevant for PerX audience. For example searching for the typical keyword "corrosion:" OAIster return 5,848 records (search avg: 26 secs.), being the top five hits:
      1.- A paper written in French (?) "La mesure de l'érosion dans le gypse/anhydrite des Alpes françaises du Nord. Méthodes et état des connaissances" Chardon, M., Association Revue de Géographie Alpine, 1996
      2.- A web site (so is that hidden web?) URL: http://www.science-and-research.com/science/chemistry/electrochemistry Electrochemistry Listings, Science and Research Inc. 2001
      3.- A web page (so is that hidden web?) URL: http://www.acoon.de/cgi-bin/showcat.exe?cat=Top/Science/Chemistry/Electrochemistry Electrochemistry Links, Acoon, 2003
      4.- A technical report in French (?) "Spectroscopies optiques non linéaires pour lanalyse des interfaces". Peremans, André AP et al, 2005
      5.- A thesis in German (?) "Silicium aus der Gasphase : Optimierung der Gasphasenabscheidung von Silicium aus Silan", Hofstätter, Michael, 2002

      In comparison, Google Scholar returns 943,000 hits (search avg: 2 secs.), being the top five hits:

      1.- Principles and Prevention of Corrosion. D Jones - Macmillan Publishing Company(USA) Cited by 450 papers
      3.- High Temperature Corrosion. P Kofstad Elsevier Applied Science Publishers, Crown Cited by 580 papers
      4.- Effect of 1-Hydroxyethylidene-1, 1-Diphosphonic Acid on the Corrosion of SS Cited by 60 papers
      5.- Corrosion Engineering. [Journal of The Electrochemical Society 126, 232C (1979)] MG Fontana, ND Greene, DD Mcdonald - Journal of the Electrochemical Society, 2006 - link.aip.org Cited by 508 papers

      In comparison, TechXtra returns 12,815 hits (search avg: 20 secs.), being the top five hits:

      1.- Electrochemical studies on carbon dioxide corrosion and its inhibition. ADT Thesis (1996)
      2.- Corrosion Detection in Airframes Using a New Flux-Focusing Eddy Current Probe citeSeer paper (1998)
      3.- The corrosion of iron and steel / by J. Newton Friend COPAC book (1911)
      4.- ECORR CAL Course Material for the Teaching of corrosion in Engineering HEA Engineering learning resource (2006)
      5.- Corrosion-Club.com Intute Key Website (2005)
  • Related Information

    An early, but most formal report on our initial findings using the above protocols can be found here.