What is the effective lifetime of a database?

Yesterday’s blog has me waxing hypothetical, today it is philosophical. Does a database have a lifecycle? Is there a natural arc?

1. Someone starts to collate information on a scientific topic,

2. Realizes a list is too limited and needs a database.

3. Research topic is nascent so the database captures content and serves as focus or hub to community.

4. Gets research funding

5. Then it gets too big for initial developer enlists help support Or moves on.

6. Database growth slows down as field maxes out or interest wanes.

7. Database logins decline.

8. Funding ceases and database not updated.

9. Bought or Dies.

10. could not think of a tenth point – can you? Is this arc realistic?

OK so that’s one way to look at it. Are there examples to illustrate this?

We are definitely seeing interest in drug repurposing. We have seen FDA collate lists of compounds repurposed already and we have seen NCGC put together a first draft of a database. I think there is a gap here for some entity to take just the drugs and combine with literature content and put it out as a database that you or I can build on which is free and structure searchable. So we are in that ascendent phase.

Next we need the community to rally around the database and add content. Fine, so we can explore all potential targets with all compounds, this could be via physical testing or in silico and enable various types of analysis. The scientist could add their own information publically to the database (this would be a condition of use – not going to get anywhere without a bit of collaboration). So we see content increases. I predict peoples interest in this approach will eventually shift as it becomes an accepted method, then it will wane after they either find it gives us cures to all diseases or fails as the panacea people expect. In this case I give it’s lifecycle as less than a decade worst case. But even knowing that, we should give it a try just because we know something has a finite lifecycle. Nothing lasts forever – and I am being optimistic.

No comment yet

1 ping

  1. Egon Willighagen says:

    10. original owners make the data Open Data
    11. Dr. Who picks it up, and reboots the process to step 3

    The Dr. Who model was introduced to me by Peter Murray-Rust:


Leave a Reply

Your email address will not be published.

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=""> <s> <strike> <strong>