SpeciesBank Dreams and Realities

1
SpeciesBank Dreams and Realities

• Rainer Froese
• IfM-GEOMAR
• rfroese_at_ifm-geomar.de
• March 2005

2
The SpeciesBank Dream

• ... a computer interface to the Internet
• able to find, combine and present data
• in a way that would be meaningful and useful
  to the person who issued a query about a species.

3
Reality Check Background

• Two-third of all major software projects fail
  (IHT 25.1.05)

4
SpeciesBanks, what are they not?

• Regional or global checklists
• Purely distributed systems
• Google mix of good and garbage
• Three year projects
• Amateur products
• Specialist products
• Committee products
• Community products

5
Reality Check Users

• Taxonomists?
• Decision makers?
• Stakeholders?
• Nobody?
• Depends on usefulness
• Mostly interested public students
• Few specialists

6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
Reality Check User Needs

• Politically very important
• Boring at best (motherhood statements)
• Typically misleading
• Most users dont know what they need
• Scientific approach analyze actual usage of what
  is available

10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
About 10,000 visitors per month
16
What Determines Usage

• Quality and accuracy?
• Recognition of scientists behind database?
• MoUs?
• Beautiful interface, fancy tools?
• Content common names, photos, summaries
• Simplicity of interface (e.g. Google)
• Number of clicks needed response time

17
(No Transcript)
18
What Determines Usefulness?

• Actual use
• What is not used is useless
• How about yourself (the custodian)?

19
Number of strategies used by phylogenetic Classes
plotted over number of recent species in the
Class, with linear regression line forced
through the origin slope 0.37 r2 0.9754 .
20
Orders per Class plotted over Species per Class.
Sarcopterygii, Elasmobranchi and Actinopterygii
fall nearly on a hypothetical straight line
through the origin slope 0.37.
21
Orders per Class plotted over species per Class
for four Kingdoms and 415,000 species the
dotted line indicates the maximum number of
Orders per species in a Class slope 0.37.
22
What is the Best Quality Assurance ?

• Scientific degree of encoders?
• Double-encoding?
• Hierarchy of checking?
• Usage by custodians!
• Usage by others!

23
Speed of Data Flow

• What determines speed of data flow?
• Bandwidth?
• Trust!

24
How to Prioritize Data Entry

• What approach is best when prioritizing data
  entry?
• User need analysis?
• Importance and quality of data?
• Opportunism! Enter what is ready for entry.

25
Enemies

• Who are your most dangerous enemies?
• Critiques?
• Jealous colleagues?
• Unconvinced donors?
• Institutions!

26
Data Encoders

• Who are the best data encoders?
• Students?
• Long-time staff?
• Experts?
• Women!

27
Members of the FishBase Team in 1998
28
Back to Dreaming

• Building the AllFish Species Portal
• Form Consortium of respective SpeciesBank
  Custodians and Institutions
• Agree on Concept, Standards and Protocols
• Use FishBase Interface and Servers
• Have small AllFish Encoder and Programmer team
• Find modest funding from different donors
• Have AllFish up-and-running within one year

29
Dont Dream ItBe It