What Brian Cant Never Taught You About Metadata

Transcript

1. What Brian Cant Never Taught You About Metadata a presentation by drew mclellan
2. What Brian Cant Never Taught You About Metadata a presentation by drew mclellan Everything You Know About Metadata is Wrong
3. What Brian Cant Never Taught You About Metadata a presentation by drew mclellan How I Learned to Stop Worrying and Love The Data
4. geek in the park metadata, html, robots, 1970s/80s childrens television programming, tofu, truth, honesty, and some made up rules stated as absolutes.
5. enough about you, let’s talk about me for a minute. allinthehead.com edgeofmyseat.com webstandards.org microformats.org drew mclellan
6. enough about me, let’s talk about Cant for a while. brian cant
7. Brian Cant taught us lots of things.
8. Brian Cant taught us lots of things. 20% 80% Brian Cant Other sources everything I know
9. 1% 99% Other important stuff Metadata of that 80%
10. THERE’S MORE THAN BRIAN WAS LETTING ON.
11. Brian taught us to share. sharing is important
12. The web is all about sharing just ask Humpty and Jemima.
13. We use the web to share like Brian taught us. knowledge information data the web is primarily a tool for sharing
14. What are we sharing? all types of data common obscure
15. Common data names addresses dates & times things for sale reviews
16. Obscure data your auntie’s hat collection every place Paul McCartney has sneezed since 1962 how many days a web server has been up
17. this is why we publish it All data is potentially useful to someone else.
18. Brian taught us to tell the truth
19. Data is only useful if it’s correctly described
20. Data is only useful if it’s correctly described but we’ll come onto that in a bit.
21. So, metadata then
22. So, metadata then Metadata is data about other data enables you to unlock data
23. Bus timetable
24. Audio
25. Photographs
26. Metadata is everywhere Often hidden away, but doesn’t deserve to be.
27. The more exposed metadata is, the more useful it is and the more useful the original data becomes.
28. Rule #1 Beware dark data.
29. Rule #1 Beware dark data. Hidden data gets forgotten and goes out of date.
30. it’s not complicated metadata is simpler than it sounds
31. sunny
32. sunny yesterday’s weather:
33. sl6 8aj
34. sl6 8aj postcode:
35. 1980-02-21
36. 1980-02-21 date of birth:
37. Information is data put into context data is grand on its own without context it cannot inform
38. Metadata puts data into context turns it into information information is even better than data
39. information is 3 times better than data 0 20 40 60 80 Information Data Betterness
40. Metadata isn’t new to the web no more than stalking is new to Facebook
41. XML is a good example <building> <colour>orange</colour> <type>house</type> <doors>1</doors> <windows>0</windows> </building>
42. XML is a good example define your own schema describe the data you have
43. Semantics and metadata aren’t identical concepts different ideas on the web there’s a lot of overlap
44. HTML has a basic set of tags Some enable us to communicate meaning Some put data into context Often both these things
45. HTML has a basic set of tags Some enable us to communicate meaning <p> <h1> <h2> <h3> <h4> <h5> <h6> (useful but not great metadata)
46. HTML has a basic set of tags Some put data into context Often both these things <title> <address>
47. HTML enables us to add metadata the HTML class attribute <span class=“name”>Drew</span> this is a very useful technique
48. HTML enables us to add metadata this makes HTML extremely flexible a good thing indeed
49. There’s a really obvious example of metadata use in HTML surely you’ve already thought of it
50. HTML META a.k.a. meta tags been around since HTML 2
51. HTML META keywords description author copyright date Dublin Core the spec lists no legal values
52. HTML META <meta name="keywords" content="vacation, Greece, sunshine" /> <meta name="description" content="My holiday in Greece" /> <meta name="author" content="Drew McLellan" /> <meta name="copyright" content="Drew McLellan 2008" /> <meta name="date" content="2008-06-12T12:03:56+0100" /> <meta name="DC.identifier" content="http:// www.ietf.org/rfc/rfc1866.txt" />
53. The use of META elements hasn’t been plain sailing
54. The use of META elements hasn’t been plain sailing many web designers don’t know how to use them properly leading to inconsistent use dark data
55. Many misunderstand the purpose I’m looking at you, web marketeers and you, so-called SEO experts
56. Many misunderstand the purpose META tags aren’t for search engines META tags are used by search engines
57. Many misunderstand the purpose META tags aren’t for search engines META tags are used by search engines META tags are for describing the data
58. Many misunderstand the purpose to provide a means to discover that the data set exists and how it might be obtained or accessed; and to document the content, quality, and features of a data set, indicating its fitness for use
59. Rule #2 The more you lie, the less you can be trusted and the less valuable your info becomes.
60. Rule #2 This is something Brian Cant taught us.
61. Rule #3 The fewer distinct consumers, the less valuable the metadata becomes over time.
62. Rule #3 Only search engines really used META keywords, descriptions Authors began writing targeted for search engines “how do I get well ranked?” vs. “how do I describe this data?”
63. Rule #3 Search engines can no longer trust keywords, descriptions Abuse has spoiled it for everyone Brian Cant never said anything about that.
64. What have we learned so far?
65. What have we learned so far? Sharing is good - the web is for sharing Metadata isn’t new IRL or on the web HTML gives us ways to express metadata It only works if we tell the truth
66. We need thems robots on our side Part 2:
67. or against us robots are either with us
68. so we’d better co-operate we don’t want them against us
69. and effort robots can save us time
70. yay.
71. tofu robot says: data is everywhere
72. There are lots of idioms for data Opening times Event details Addresses
73. Idioms are good they’re not always formal you don’t need to be formal to be understood
74. Informal is good but consistency is important let’s look at why...
75. Humans are quick to adapt we can easily re-evaluate and adjust we can climb stairs without a trip to the workshop
76. Robots prefer patterns they rely on known patterns patterns can be formal or informal must be consistent and repeatable
77. Humans like patterns too we like routine we like repeating patterns robots like patterns because they are repeatable we like patterns because we don’t want to think thinking is hard, uncomfortable and inconvenient.
78. thinking 45% 29% 4% 21% Hard Uncomfortable Inconvenient Prone to error
79. So as it turns out what’s good for thems robots is good for us too
80. it’s not complicated metadata is good - so we want to use it our metadata challenge need to embrace reusable patterns avoid dark data avoid specific data for any consumer make it easy to be truthful embrace existing idioms reuse existing technology
81. remember this? <span class=“name”>Drew</span>
82. remember this? <span class=“name”>Drew</span> avoid dark data avoid specific data for any consumer make it easy to be truthful embrace existing idioms reuse existing technology
83. remember this? <span class=“name”>Drew</span> avoid dark data avoid specific data for any consumer make it easy to be truthful embrace existing idioms reuse existing technology need to embrace reusable patterns
84. microformats just a bunch of patterns
85. names and addresses hCard - based on VCARD given-name family-name email url tel title org street-address locality
86. names and addresses hCard - based on VCARD <p class=“vcard”> The announcement followed calls by <span class=“org”>Apple</span> <span class=“role”>Chief Executive</span> <span class=“fn”>Steve Jobs</span> earlier this year... </p>
87. names and addresses hCard - based on VCARD <p class=“vcard”> The announcement followed calls by <span class=“org”>Apple</span> <span class=“role”>Chief Executive</span> <span class=“fn”>Steve Jobs</span> earlier this year... </p>
88. events and dates hCalendar - based on iCAL dtstart dtend location url description summary
89. reviews hReview item reviewer rating description summary photo
90. relationships XFN contact acquaintance met co-worker friend colleague neighbor child parent sweetheart crush me
91. many more licenses tags date-based feeds directories products payments geolocation more
92. remember this? <span class=“name”>Drew</span> avoid dark data avoid specific data for any consumer make it easy to be truthful embrace existing idioms reuse existing technology need to embrace reusable patterns
93. remember this? <span class=“name”>Drew</span> avoid dark data avoid specific data for any consumer make it easy to be truthful embrace existing idioms reuse existing technology need to embrace reusable patterns
94. Brian Cant never knew this but I bet he’d be thrilled.
95. microformats are good a humane method for using metadata on the web easy for us to implement readable by our robotic friends
97. hCard
99. hCalendar
102. For robot masters http://microformats.org/wiki/parsers http://tools.microformatic.com/
103. For humans http://microformats.org/
104. For humans http://microformatique.com/book/
105. For humans http://oreilly.com/
106. What Brian Cant Never Taught You About Metadata. So that’s Thank you. http://allinthehead.com/presentations
107. http://flickr.com/photos/gperez/4393118/ http://flickr.com/photos/warmnfuzzy/466382466/ http://flickr.com/photos/stevegarfield/194648339/ http://flickr.com/photos/donsolo/2385041554/ Creative Commons photos used: