4 thoughts on “considering OAI-PMH

  1. Pingback: Tweets that mention considering OAI-PMH -- Topsy.com

  2. Hi,

    I was wondering if you could name an aggregation protocol that is not ‘reviled’ by whoever has to use it. Aggregation is just like that!

    I am wanting to suggest the use of OAI-PMH in up coming projects. I feel there must be an alternatives but have yet to find them. The main issue I have is that the main feed technologies don’t appear to allow you to go back in time. If you find out that your harvester has been unreliable for the last couple of weeks you can’t go back and say “give me everything that has changed since…” and page through to the present day. (I may be missing something in the ATOM spec but I don’t see it).

    People seem to complain about OAI-PMH but as you point out that is usually to do with metadata standards not the protocol. Nobody actually proposes an alternative.

    What am I missing…

  3. Hi Roger,

    I agree that many issues with aggregation are ultimately about underlying data quality. There are some issues with OAI-PMH in itself as it’s currently specified which are discussed in the above captured tweets – however, for all the problems which it may have, my colleague (Phil Barker ) was recently pointing out that one of the difficulties with OAI-PMH is that is it’s implemented by comparatively so few pieces of software that if you want to write a harvester or endpoint interface and address some of the underlying data quality issues you probably have to write everything yourself. Phil would contend that most feed technologies are so ubiquitous (Rss/ Atom) that there is a much greater chance the there are freely available software libraries addressing data quality issues that you can draw on to create your software.

    in terms of going back in time … this is perhaps the greatest weakness of RSS/ATOM as it’s often implemented. The issue isn’t in the spec (as i understand it) but that it is often implemented as a (10 most recent items) newsfeed – the spec can be implemented to provide a complete listing of resources (for example listing everything or everything matching a particular criteria). If you have control over the resource providers you could use rss in this way quite effectively. however, if they’re already set up and supporting OAI-PMH, or using commercial software (which is likely to restrict how you can output the rss) – it may be more straightforward to keep using OAI-PMH (depsite the above limitations)

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>