Old technologies can be fascinating and informative for best practices for new technologies. However, they can also be broken systems that burden users and developers who are trying to use them and work around them. PURL servers are one of the broken ones in need of replacement.
Like MARC, PURL servers are a dated concept that lead to failures. Yet, now PURLs are PURLz and are gaining new adopters, even though the US Government Printing Office had a PURL server outage that lasted over a week (Aug. 28 and still in process of correction Sept. 4). The design of PURL servers makes mirroring difficult. Thus, using PURLs to have persistent uniform resource locators can result in link failures even when resources are available, as is the case with the GPO documents right now.
With limited budgets and limited people resources, replacing a technology that isn’t clearly and provably broken can be difficult, even for the technologists who see the broken technology as looming disasters waiting to happen. While GPO’s painful downtime (and downtime is always painful) is awful, that pain could be harnessed to form a wonderfully clear argument that articulates the problems with PURL servers and advocates for their replacement. Hopefully GPO will be able to use this to replace their PURL server, as they’ve considered. Then, hopefully they can help others in migrating to technology that is more resilient and then won’t subject us all to unnecessary and painful downtime.
2009-09-07