02/08/2008: "LSIDs"

They stand for life science resource identifiers, they are an OMG based standard, they offer nothing novel whatsoever while making the developers work a nightmare and apparently they are here to stay. I spent most of the week trying to educate people about the dangers of LSID identifiers. For those readers that do not know what LSIDs are, it suffices to say they are location independent identifiers. URIs they say poses a problem in that they idenfity location dependent resources whereas LSIDs along with the related infrastructure alleviates 'supposedly' this problem. It was really made up by computationally challenged bioinformaticians who spend no time trying to understand W3C specs and by rediscovering the wheel, they came up with the LSID spec. In all their grief they had a point. I remember doing some data mining for a bioinformatics projects and it was unbelievable how many different names a protein in a database would have. LSIDs propose a univeral persistent name for such data. This is not new however nor novel. URI variations such as XRIs and IRIs do the exact same thing and they are alot older specs. But why make our life easier when we can make it harder; hence another spec saying the exact same thing. I started thinking that maybe the W3C standards and more importantly their names may repell scientints and as such they feel the need to come up with their own specs. Who can blame them for trying to support it?

I spent most of the week trying to explain to people that you cannot carry out 'location independent' security, or provenance, or auditing for that matter because each domain has by definition its own gatekeeper. I must have writen about 2000 words in emails for this matter alone. I ended up explaining tokens, time spamps and the like unitl I got bored.

Hoping to a more intellectually stimulating week. Another 12 days to go until the Boston conference trip.