The ar Xiv eprint archive CS 520 guest

The ar. Xiv eprint archive CS 520 guest lecture Simeon Warner (simeon@cs. cornell. edu) Simeon Warner - 21 April 2003

What is ar. Xiv? • Eprint - lots of meanings, here: pre-print, post-print, no-print, refereed or unrefereed, openly accessible via internet • 230, 000 papers (35 k/year, ~150/day) • Mostly physics, some math and cs • Mostly Te. X/La. Te. X source • Automated PS/PDF production • 17 mirror sites world-wide, nightly update • Secondary submission site in Lyon, France Simeon Warner - 21 April 2003

Full-text downloads/week Simeon Warner - 21 April 2003

History • 1992 - hep-th created (email reflector, messages were saved). ~200 users. • 1992 - ftp interface. hep-ph… added • 1993 - web interface • 1994 - data on remote sites moved to main site, they become mirrors • 1995 - automatic PS from Te. X • 1996 - PDF generation, mirror network grows • 2003 - ~70, 000 users Simeon Warner - 21 April 2003

Not often cited as HCI exemplar Simeon Warner - 21 April 2003

Submission form Simeon Warner - 21 April 2003

Submission methods Simeon Warner - 21 April 2003

Submissions to different areas Simeon Warner - 21 April 2003

Why is cs archive failing? • Why are physics/math working? § Strong pre-print culture in physics § Active promotion from inside communities • CS less Te. X based • CS has stronger homepage culture? • CS has Citeseer (aka Research. Index) • CS researchers hate ar. Xiv interface? • CS has more conference publication? ÞLESSON: Can’t apply `one size fits all’ Simeon Warner - 21 April 2003

Different habits in CS (1) Simeon Warner - 21 April 2003

Different habits in CS (2) Simeon Warner - 21 April 2003

Current problems at ar. Xiv • Too much admin time § 150 new/day, 30 replacements/day • Inappropriate submissions § Encouraged by ar. Xiv’s popularity § Should be “peer reviewable” quality § Must be appropriate to subject areas • Copyright PS/PDF § Don’t (can’t? ) automatically detect • Submission size § (otherwise) good tools produce bloated figures Simeon Warner - 21 April 2003

Discipline-Based Repositories, Institutional Repositories, Open Access… • Driven by cost, ideology, self-promotion… • ar. Xiv is exemplar archive § Largest, best-known, discipline based § Need open archiving • Current interest in institutional repositories § Slice the other way to disciple based § Coexist? • Metadata sharing - OAI § Essential infrastructure for more disperse model than ar. Xiv Simeon Warner - 21 April 2003

Publishers and Peer-Review • Peer-reviewed publication necessary for academic rewarding (promotion, tenure, prestige) § System slow to change • Physics publishers accept/live-with ar. Xiv § Can even submit to journals via ar. Xiv § Not true in chemistry/biology • Peer-review costs ~$500/paper § Where would that come from without subscriptions? Simeon Warner - 21 April 2003

Resources • ar. Xiv: http: //ar. Xiv. org/ • Front (interface to math ar. Xiv): http: //front. math. ucdavis. edu/ • Algebraic Geometry and Topology (ar. Xiv overlay journal) http: //www. maths. warwick. ac. uk/agt/ • OAI: http: //www. openarchives. org/ • BOAI: http: //www. soros. org/openaccess Simeon Warner - 21 April 2003