SLAC Inputting - Spires legacy workflow

Introduction

Here I have inlcuded 4 flowcharts to describe the current slac inputting process. These should be considered when desigining the SystemDesignBibEdit to take are of make of the ComparisonSlacFermilabDesyCernEnrichmentScripts I will include some proposed screenshots of how we have considered a web version of this process could look. Currently it is a command line process, but we made some pencil and paper mockups of web based versions before starting the inspire project.

Included are charts for arXiv harvest (basically just OAI) the main inputting flow, and the sub-processes for authors and citations. Not (yet) inlcuded are processes for adding from holding pens or adding long(>15-20) author lists (giva).

As I mention on the diagrams, the idea in these processes is to make the inputters job as efficient as possible, since that is the limited resource. The routines should be as smart as possible to guess the correct answers, which should be confirmed or selected by the inputter. Inputters should almost never type anything, as this takes too long and leads to typos. Everything should be optimized for the common cases.

In regard to eprints, the common case is:

• New paper (not a duplicate)
• 5 or so authors - no institutions harvested, but listed on pdf.
  • Multiple groups of authors, each group of 1-2 from same inst.
  • Authors from same inst as thier last paper
• Most (20-30) citations extracted correctly, but 1-2 with an error in extraction (usually missed completely, or error in paper itself)

Inputting: Main Process

Inputting: Authors

Inputting: Cites

Inputting: Giva

Inputting: Holding Pen

• The merging of the two records (one from the holding pen, one from HEP) is a special case of the more general situation that should have a good web interface: the merging of two records. The interface here might be configurable to keep certain things depending on the source, but in general one wants to display them side by side and choose what to keep and what not to keep... with suitable defaults.
• The checking for duplicates is also something that occurs in other places.
• This process then, is really just the combination of two more general procedures (dupe checking and merging) with the slight differences that here we keep the candidates in a separate pen (could just as easily hide them) and that if no dupe is found we must add the candidate, and may need to do some checking there...

Spires:Email(RT)

In order to completely specify the system shown in Harvesting:General here is a rough diagram of the process for handling the Email input from users. This is an inherently manual process, however, we have automated some aspects of it, most notably the citation corrections that are most frequent.

Harvesting: arXiv

Harvesting: General

Spires:System

This is an older diagram of the web infrastructure used by spires. It is unlikely to be useful for inspire, but is included for completeness.