Two random thoughts
May 10th, 2007 | Published in Browsers, Web, Work | 6 Comments
On Download History — It might be cool if my browser history also indicated from which sites I’ve downloaded files so I could find those sources again. How separate should Download History be from History?
On Smart Agents — Are services like StumbleUpon a rudimentary sort of “SmartAgent” where, instead of developing machine intelligence, we harness the “wisdom of crowds” to find new undiscovered content that fits into patterns of interest we develop over time?
May 11th, 2007 at 12:47 am (#)
Very insightful. Even if download history is kept as a separate list (which has merits), it would be useful to have an “open host page” link in the context menu (or even directly visible in the list).
May 11th, 2007 at 1:47 am (#)
Very insightful. Even if download history is kept as a separate list (which has merits), it would be useful to have an “open host page” link in the context menu (or even directly visible in the list).
May 11th, 2007 at 7:06 am (#)
As a followup to my plug for downloads to remember their source, http://www.dria.org/wordpress/archives/2007/05/09/550/#comment-27254, the idea was brought to me back in 04/05 by a PhD student at CMU. He used tracking of link text, titles, etc up to the download point to enrich the file with data to support file system search:
Connections: using context to enhance file search
http://www.lcs.ece.cmu.edu/~soules/papers/sosp05.pdf
Craig A. N. Soules, Gregory R. Ganger.
Symposium on Operating System Principles, October 2005.
Another way to look at this is how should download history affect web history? A download is a strong vote of confidence or interest in history. See http://doi.acm.org/10.1145/1059981.1059982 for other strong predictors. This is the kind of interpretation of history required to make the wisdom of the crowds, well, wise.
May 11th, 2007 at 8:06 am (#)
As a followup to my plug for downloads to remember their source, http://www.dria.org/wordpress/archives/2007/05/..., the idea was brought to me back in 04/05 by a PhD student at CMU. He used tracking of link text, titles, etc up to the download point to enrich the file with data to support file system search:
Connections: using context to enhance file search
http://www.lcs.ece.cmu.edu/~soules/papers/sosp0...
Craig A. N. Soules, Gregory R. Ganger.
Symposium on Operating System Principles, October 2005.
Another way to look at this is how should download history affect web history? A download is a strong vote of confidence or interest in history. See http://doi.acm.org/10.1145/1059981.1059982 for other strong predictors. This is the kind of interpretation of history required to make the wisdom of the crowds, well, wise.
May 11th, 2007 at 10:24 am (#)
A colleague of mine wondered once why we couldn’t get “download source” put into file metadata, and I admit every time I have a file and have forgotten where it came from, I think the same. I actually archive this info manually for academic papers, so I can see where he was coming from.
Download location history would be a great start! I’d say it should stay with the download history and just be an extra bit of info like the download date is.
May 11th, 2007 at 11:24 am (#)
A colleague of mine wondered once why we couldn’t get “download source” put into file metadata, and I admit every time I have a file and have forgotten where it came from, I think the same. I actually archive this info manually for academic papers, so I can see where he was coming from.
Download location history would be a great start! I’d say it should stay with the download history and just be an extra bit of info like the download date is.