3. Notable changes
Deposits
FundRef
Deposit
stand-alone CrossMark and FundRef data
Include
MathML in article titles
Include
abstracts (JATS)
Non-CrossRef DOIs as components
Text-mining support (metadata about licenses)
Queries
Can query on Orcids
Internal
Improved data replication
Hardware upgrades
System update 2013
12. In the works
Assigning
Multiple DOIsto books
Problem:
Book deposits from multiple sources had depended on multiple
resolution but the participants found the process cumbersome or
unworkable
Solution: Allow
multiple DOIs to be deposited for the same book content
from one or more members (different or the same prefix) and make
multiple resolution happen by magic behind the scenes.
Allow
references to non CrossRef DOIs
Problem:DOIs
assigned in DataCite to databases may be referenced
from journal articles
Solution:
Allow these DOIs deposited as <citations>. Then allow anyone
to perform Cited-By queries on these DOIs.
System update 2013
During the past year we’ve continued to make changes and improvement, including high profile initiatives like FundRef. Some are incremental improvements but many of these activities continued to address sustainability and performance issues since this work is never really DONE. I’m coming to believe that there is no light at the end of the tunnel.This year we had 2 or 3 significant events that resulted in temporary performance issues, each of which were overcome and in the end left us with a stronger system.While the rewrite has been completed for over a year, I’d have to say we are now in a phase of redesign. Much of it is happening organically with a few more substantial activities still on the drawing board.
FundRef was one of the larger initiatives we worked on this past year, and I should still working on.We’ve made changes to facilitate easier deposits of Crossmark and FundRef dataModified the deposit schema to support MathML, Abstracts, non-CR DOIs and text mining data
Overall system availability and performance is good or very good. However, as I mentioned on my opening slide we have had a few problematic episodes lasting days or reoccurring over a few weeks.Most recently due to growth in deposits we found that the basic query function is slowing down at times due to a buildup of alternate spellings for titles and a buildup of deposited title ‘noise’ caused by mistaken deposits of titles and the like.Addressing this will be a top priority immediately after the AMM.
These charts show what I mean, as you can see in the charts there a spikes where query performance slows to several seconds/query. This behavior became very pronounced on about October 22nd and has continued since , particularly between 4 and 9AM
The one statistic we all should work to improve is the participation rate in cited-by linking . Still under 50% of journal DOIs, and less than 33% of all DOIs are deposited with references. Thisis the count of cited-by links available in CrossRef as of last week
Deposit of DOIs with references really drop off going backwards through the 2000’s and flatten out at very low levels even for the 1990’sDeposit of references for backfiles !!
For the 12 years I’ve been doing these system updates for you, I’ve almost always stuck to facts and charts which only give a snapshot of the state of affairs. But this year I’d like to high light some of the main reasons for our success
While there are certainly a number of people that make CrossRef possible such as in accounting, marketing and those in the UK office, I’d like to take a moment to give credit the US based tech team. These folks mostly toil away in obscurity except when I force them to take the stage for a few minutes like at yesterday’s workshops.
Tim Pickard has been with us since August 2002.He’s our jack of all trades: unix system admin, network admin, SW tool master and laptop Mr fixit.
Jon Stark has been here since January 2004.He’s one of our software developers and is the guy responsible for all the regular (automated) reports you receive. Jon has also become our query algorithm master and is usually the guy who finds out why some query did not match.
Patricia Feeney joined in April 2007She’s our support manager and oversees all support ticket issues, she’s the author of our help documentation and has become the RI for our various XML schemas. Patricia has also served on a number of industry initiatives for NISO and SSP.
He joined in September 2009 He is our main system architect and has implemented nearly all of the system infrastructure code.Along with Tim and myself Andrew receives wake up calls at 2AM when some gremlin is causing mischief in the system
Vaishali Patel joined us in January 2011.She works for Patricia and is our support desk and triages or answers just about every ticket,
Mike Yalter is also a software developer and joined us in August 2012.Mike had taken over development of deposit processing which includes all fundref deposits and registry management.