Web Search Alert OLA Superconference 2002

1
Web Search AlertOLA Superconference 2002

• Gwen Harris
• goharris_at_websearchguide.ca

2
Watson, what do you see?

• www.websearchguide.ca
• Sherlock Holmes and Dr Watson went on a camping
  trip. They set up their tent and tucked in for
  the night. Around 3 am Holmes woke and said to
  Watson, look up at the stars. What do you
  deduce. For rest of story go to LaughLab
• http//www.laughlab.co.uk/press/pressinfo.html
  

3
Good and Bad

• 2001 Annus Horribilis
• Trend Document Clustering
• Google reigns but others wait in wings
• Tools to watch

4
2001 Annus Horribilis

• About.com cuts 300 guides
• Excite.com under new management
• Northern Light dims
• Pay to be found
• Pop-up ads like crows on a rooftop

5
1) About.com

• About.com cut 300 guides
• Many Canadian - canadactualite, canadahistory,
  canadamusic, canadanews, canadapolitics and
  canadateachers
• About 3.0? More e-commerce.
• Find guides at www.formeraboutguides.com

6
2) Excite

• Excite Eviscerated
• Excite_at_Home paid 7.8B. Sold to Infospace for 10
  M
• US portal carries on. Others gone.
• Search engine gone.

7
3) Northern Light

• Northern Light Dims
• Closed public search engine set back for
  natural language query
• Bought by divine content management content
  integration
• Yahoo is selling documents from the Special
  Collection. (Advanced Research Documents)

8
4) Pay to be found

• Paid Placement - Overture is everywhere Excite,
  Yahoo, Altavista, Hotbot, AOL, MSN
• Paid Inclusion Adopted by Inktomi, Altavista,
  AlltheWeb, Looksmart, Yahoo
• Lines blur - editorial content v advertising. MSN
  combines editors picks, most popular and ads in
  Featured Sites
• Will yellow pages model take over?

9
5) Pop-Up, Pop-Under

• 30 of top U.S. websites use push advertising
  such as pop-ups.
• AnalogX POW free pop-ups only
• WebWasher free for home and educational use
  ads, scripts, pop-ups.
• CookieCop from ZDNet free cookies, pop-ups
• AdSubtract Pro 29.95 US ads, pop-ups, etc.

10
5) Pop-Up, Pop-Under

• Gary Price recommends CTRL W.
• Opera browser 6.0 accept or refuse popups or
  open in background

11
Document Clustering

• Analyze search results to find words and patterns
  in common. Group into topics.

• Vivisimo
• Wisenut
• Teoma
• AllTheWeb
• Infonetware - Real Term
• QueryServer

12
Data Visualization

• View documents as stars in a galaxy, connected
  nodes, or as hills in a topography.

• Map.net from Antarcti.ca (Screen shot)
• WebMap (Screen shot)
• WebBrain
• ThinkMap Plumb Design Visual Thesaurus

13
Google

• Web pages 1.5 b indexed, 2 b reach. 25
  non-English
• Filetypes pdf, doc, xls, ppt 35 million
• lives of the saints picks up stop words
• Wildcard three mice - for a word
• Headline news 100 leading English newspapers
  indexed hourly
• Translate language tools

14
All The Web

• 625 million pages, 115 million images
• Freshest index 9 to 12 days
• Topics takes first 200 results. Eg roch
  voisine
• 3000 news sources
• Query rewrite natural language
• Extensive customization

15
Freshness

• AlltheWeb 9 to 12 days
• Google 28 days
• Inktomi 45 days but Best of Web every 9 days
  (20 of database)
• Altavista 45 to 60 days

16
Altavista

• Problems in 2001 not updating index
• Dropped date of web page.
• Improved phrase identification
• AND, OR, AND NOT, NEAR in basic search
• Dropped the . Is an AND engine (Jan 30)
• Advanced search more proximity controls
• Promises localized results
• Future????

17
Altavista - Proximity

• Advanced Search
• (anthropology within 5 polynesia) and not amazon
  
• (anthropology 5 polynesia) and not amazon
  also means within 5
• (anthropology lt polynesia) anthropology
  before polynesia by 10 words or less

18
Canada

• Altavista Canada improved index
• Google Canada largest and best
• MSN Canada Inktomi has added regional
  identification
• Canadian-Links 55,000 sites, 1000 categories.
  Aim is to index every Canadian site.

19
Canada

• Canada
• AV.ca ??
• Google 2.5 m
• MSN.ca 14,538
• MSN.com 7,336

• Roch Voisine
• AV.ca gt 150
• Google 1,480
• MSN.ca 3,219
• MSN.com 3,215
• Yahoo.ca 6
• Looksmart Can 4
• Canadian Links 12

20
Document Summarization

• Surfwax SiteSnaps and InfoCubbySelect a search
  set, run a meta-search, summarize selected pages,
  save to an InfoCubby, and share.
• Copernic Summarizer Web browser utility
  process pdf, doc, txt. Picks out high quality
  keywords (47.99 US until Jan 31)

21
Summary

• Dont need except for stop words
• Less use by search engines of subject directories
  BUT more use of document clustering for grouping
  by topic
• Freshness getting better
• New tools for the researcher
• No free lunch advertisements and for-fee search