Re: linked open data and PDF

I haven’t dropped the thread, but I’d like to up-level a little:

- I think we might have different models for how documents typically get 
produced, and this difference in models results in different perspectives 
on what constitutes “easy” and “hard”.
  We should compare typical document workflows for the primary linked open 
data use cases (government transparency, scientific publication, regulated 
financial) to understand how data gets lost in the publication pipeline.

- I think there’s an assumption (sometimes, perhaps inconsistently) over 
whether the data wanted was actually already in the published document, or 
whether we were asking for more. For the primary use cases, the desire is 
for MORE data than what is currently published in the documents… not just 
the pretty chart but the raw data (suitably anonymized). The problem isn’t 
getting data OUT of HTML or PDF or whatever format, the problem is how to 
put (more) data IN. 

- On a different tangent: with the latest IE, all the major browsers 
(Chrome, Firefox, IE, Safari) come with their own built-in PDF display and 
interaction. The usability issues vary by browser and platform, but these 
are important browser features — Jacob Nielsen’s 2003 blog post about web 
usability of 11+ year old software is hardly relevant.
 insofar as you define “the Web” as including “what browsers do”, then 
PDF, just now, is more part of the Web than it has ever been before.


Larry
—
http://larry.masinter.net

Received on Friday, 20 February 2015 05:20:38 UTC