Design Doc for Linguistics-Oriented Data Sharing/Publishing Platform

Use cases

who would be interested in the platform?

1. the kwa lexicographer
   a traditional linguist that wants to do the right thing.
2. tracy: linguist that has done work, which should not be lost.
3. alicia: (big) data without time to annotate/analyse.
4. virach: competition on thai word segmentation.
   

1. the kwa lexicographer

linguist uses toolbox to write a lexicon on a kwa language dialect,
using custom markers (custom fields of lexical entries), thus toolbox doesn't work anymore.
what the linguist wants to do is prepare a printed book.
would be interested to get interlinear glossed examples for the lexicon

concerns:

• data conversion
• merging of data
• cross-platform publishing (including the non-web world)
  

extension (once on the platform):

• find new data (on kwa), transcribe it -> online collaboration for annotation
  

subfields:

• language documentation
• field methodologies
• phonology
  

2. traci: linguist that has done work, which noone should have to do again.
(radio shows, dictionary entries)
the squib
-> backup
-> interest in "publication clearance process" (the "lawyer button")


subfields:

• syntax

concerns:

• backup
• documentation
• reproducibility
  

extension:



3. alicia: there's (big) data without time to annotate/analyse.
(pacific nw phonetics) ("light annotation problem")
-> the case for video
-> the researcher that wants to reuse/reannotate data of others

subfields:

• sociolinguistics
• phonology
  

concerns:

• backup
• sharing with access control (graded access)
  

extension:


4. virach: competition on thai word segmentation
the tool builder
-> tools in search of data
-> results in search of judges

subfields:

• computational linguistics

concerns:

• find users for tools
• software as a service (potential for mashups)
• distribution mechanism
• attribution
  

extension:

with a platform that solves the concerns mentioned above, what else do we get?

foster the education interaction: connect students with community and
with data. (attribution will help here)

now it turns out that some concerns match up with extensions.
-> once the people are on the platform, the collaboration and network
effects set in.
- potential for citation, publication

Requirements

• data backup
  
• data conversion
• remote access
• publishing (legal/copyright/etc)
• graded access
• metadata: The platform should already provide for the basic metadata to allow for attribution (who created which data when). Tagging with language codes is necessary to enable network effects.

Potential

• platform to deliver software as a service