http://liinwww.ira.uka.de/bibliography/index.html - Good resource for papers in computational mathematics?

http://compgeom.cs.uiuc.edu/~jeffe/compgeom/biblios.html - By Jeff Erikson, good resource of things to search in computational geometry

http://ubiety.uwaterloo.ca/~tveldhui/tools/ - Good start page for searching things, looks heavily inspired by some of fravia's suggested search pages.

http://infomine.ucr.edu/cgi-bin/search?physci - search for information in engineering?

http://www.ams.org/mathscinet/

http://ubiety.uwaterloo.ca/~tveldhui/tools/litsearch.pdf - Great slides on how to do a literature search in computer science. I'll summarize some of the sailent points in the above talk.

• flow of information through computer science: http://www.lib.uwaterloo.ca/usered/grad/researchskills/flow_of_info.html
• ACM CSUR is an entire journal devoted to surveys http://portal.acm.org/browse_dl.cfm?linked=1&part=journal&idx=J204&coll=ACM&dl=ACM
• Great graphic on how an idea can be searched in the research (attached.)
• Google Scholar http://scholar.google.com/
  • Harvests publications from lots of other sources (CiteSeer? , ACM, IEEE, etc.)
  • Excellent to find an individual paper by title/authors: if it is online somewhere, changes are good Google Scholar will find it.
  • Pretty good at keyword searching. Indexes the text of the papers, which is very useful.
  • Bad: Lots of crosstalk with other disciplines (physics, medicine, biology) (e.g., try searching for “aspects” or “unification”)
• CiteSeer? http://citeseer.ist.psu.edu/ http://citeseer.csail.mit.edu/
  • Crawls the web and finds papers. Therefore heavily biased towards recently published papers (i.e., 1997-present)
  • Has two search modes: a keyword-type search (not very good), and a citation search, which is amazingly useful
  • Citation search will weight search results by how heavily cited they are.
  • Good for finding pivot papers (heavily cited papers) about a particular topic, and for browsing the citation graph forward/backwards.
  • Often excruciatingly slow and/or too busy to answer queries.
  • Good for answering questions like, “What is so-and-so best known for?” and “Has so-and-so published anything with impact in the last ten years?” and “How cited is so-and-so?”
  • To search for a person "Paul W. McCartney? ," have to do something like "(Paul McCartney? ) or (P McCartney? ) or (Paul W McCartney? ) or (P W McCartney? )" to account for the ways people might cite papers.
  • Example: (Alan Turing) or (A Turing) or (A M Turing) or (Alan M Turing) Look at the bar graph at the bottom — note the war years spent at Bletchley. (Turing was convicted for homosexuality in 1952, and committed suicide in 1954.)
  • CiteSeer? ’s list of most cited papers are worth browsing. If you’re looking for worthwhile papers to read, this is a good start. I used to organize a reading group (“High-Impact Papers in Computer Science”) where each week we’d pick a paper from the CiteSeer? top cited papers list, read and discuss it.
• Collection of Computer Science Bibliographies http://liinwww.ira.uka.de/bibliography/index.html
  • You have the paper, but you need the bibtex.
  • Search for ”The First Few Words of the Title” and you’ll probably find it.
  • Beware of the CiteSeer? bibtex entries: they are almost never in a suitable form for citing in papers.
  • Indexes many papers whose fulltext is not available online, so has larger coverage than CiteSeer? or Google Scholar.
• ACM Guide http://portal.acm.org/guide.cfm
  • Combines bibliography data from many sources
  • Full-text access to ACM publications (but, annoyingly, will not even link to others like IEEE, Springer, Elsevier, etc.)
  • Good for browsing the citation graph insofar as it lies in ACM publications
  • Fairly awful search capabilities compared to google or citeseer
  • Gives BibTeX? entries, but screws up accents (get HTML escapes like é — ugh)
• UMI Dissertations http://wwwlib.umi.com/dissertations/search
  • Good for finding dissertations on topics close to yours.
  • Restrict subject area to “computer science”
  • Dissertations often contain well-written introduction and survey sections. If you can find one in your area, it may be a perfect starting point.
• DBLP http://www.informatik.uni-trier.de/~ley/db/
  • 610,000 publications
  • Bibliographic (no fulltext)
  • More complete coverage than CiteSeer? (e.g., citeseer only has about 30% of the articles in DBLP for 2000; less than 10% for decades before that)
  • Very complete for database theory!
• Good search terms for surveys tutorial, summer school, handbook, dissertation, thesis, survey, introduction
• Develop a wiki page as you search
  • Effective search terms
  • key researchers
  • key papers.
• Find pivotal/canonical papers: these are heavily connected papers that either
  • Everyone cites, because they established the field or made an important contribution or made something understandable;
  • Cite a lot of papers, because it is a survey or the authors did an unusually thorough assessment of related work.
• Types of searches
  • backwards bibliography search
    • Portal, CiteSeer?
  • Forwards bibliography search
    • Portal, CiteSeer? , Google scholar
  • By researcher
    • lookup homepages, DBLP
• For gestalt, look for concepts and big ideas, ignore details (skimming)
  • http://www.ucc.vt.edu/stdysk/skimming.html
  • http://web2.uvcs.uvic.ca/elc/studyzone/570/pulp/hemp1.htm
  • http://helios.hampshire.edu/~apmNS/design/RESOURCES/HOW_READ.html
• Go back to papers you couldn't understand later.
• Suggested format for BibTex? tags: FirstAuthor? :Venue:Year