Open Science
short URL: http://tr.im/openscience
Update
The Open Science session proposal for the SXSW 2010 interactive conference has been accepted! Please go vote for it!
7 steps to Open Science today
recommendations to science publishers discussed 2009-190 at the "open science open data on today's web" session at SciBarCampPaloAlto (see after the recommendations for the outline of the talk), in rough order of priority (highest first)
- Just publish it! Whatever existing electronic format you have your papers, data in, whether HTML, PDF, .doc (MS WORD), CSV, just publish it on the Web. Don't worry about format conversion (for now). Just put the data out there so it's at least somewhat accessible to the web, search engines, applications etc.
- open robots.txt. As Obama did, no disallows for any content (only 2 lines in Whitehouse.gov robots.txt on inauguration. view current Whitehouse robots.txt), in contrast to previous Bush administration robots.txt (view Bush Whitehouse.gov robots.txt - over 2400 lines long!). An open robots.txt file is essential to letting search engines and archives (like the Internet Archive) archive scientific web pages for historical purposes.
- Public Domain License. Contribute every page to the public domain by linking it to the Creative Commons Public Domain dedication with rel="license". Even better: use Creative Commons CC0 for international public domain. Contributing to the public domain lets anyone create archives and in addition share them.
- Permalinks to versions. Make and maintain permalinks to the data. Version pages when major changes occur (more than just typo/spelling corrections), at least with suffix -YYYY-MM-DD snapshots of previous versions at same URL. Permalinks better enable online conversations about the content that is published, and provide good citations for other sites to use when mentioning, summarizing, or republishing scientific data. For more on this see: http://citability.org/
- Semantic HTML+microformats. Publish your reports and data with valid semantic HTML + microformats for max lifetime+access (as tweeted 2009-075). If your in-house web designers/developers don't know what that is, hire outside web designer/developer contractors that are experts in semantic HTML and microformats. Requirement: must have several live sites in their portfolio that have valid semantic HTML + microformats. Do not compromise on this.
- Empower independents. Empower developers, entrepreneurs, and the next http://govtrack.us as http://thomas.loc.gov did (as tweeted. ) by providing data plus design guides (thanks @simplescott)
- Re-use print IA for v1 usability. Parallel print equivalent information architecture (IA) for v1 usability. Many current users of scientific papers and data (papers that cite other papers) are used to citing (and thus trusting) print versions of scientific papers. On any v1 of a site for which print equivalents exist, simply re-use/mimic the existing organization of books, periodicals, sections, chapters, pages, paragraphs etc. when designing the URLs and other information architecture of the online version. Re-use, don't reinvent.
Tools
- Firefox Add-on: Dafizilla Table2Clipboard - use it to one-click copy any HTML table from a web page to CSV format and paste into Excel or whatever other old-school data-analysis tool you have.
History
Outline of talk given on 2009-109 (to be moved to its own wiki page)
Origins
- Much of this is based on my previously published suggestions for OpenGov, with iterations.
Citations
Even shorter URL: http://tr.im/opensci
Comments (0)
You don't have permission to comment on this page.