HCI Bibliography Search Help
Search Example ⇒ What it means
user interface ⇒ both USER and INTERFACE
error | mistake ⇒ either ERROR or MISTAKE
artificial intelligence | ai ⇒ (ARTIFICIAL INTELLIGENCE) or AI
color | colour & display | screen ⇒ alternative 'and' has lower precedence
internationali* | locali* | translat* ⇒ wildcards only at ends of words
pemberton ~ interactions ⇒ PEMBERTON, but not INTERACTIONS
librar* -hyper* -multi* ⇒ alternative syntax for negation
(error* | mistake*) & (loss* | cost*) ⇒ parentheses for grouping
buxton_b* | buxton_w* ⇒ special author indexing, wildcards
Notes:
- All searches are case-insensitive.
- Terms like AND, OR, and NOT are searched literally.
- Double-quotes, used by some search systems for phrase searches, are ignored.
There is no support for phrase searching at this time.
- Wildcards (*) are only allowed at the ends of words.
- Hyphenated terms, like user-centered, are double-indexed as a hyphenated term
and as individual terms.
- Common terms match large, random subsets of the database, and should be avoided:
- Common English terms such as: the of and an
- Common terms in HCI, such as: human computer interaction.
A search for "hci human computer interaction user interface"
matches a few hundred records -- less than 1% of the bibliography --
but it has no specificity. It is better to use specific terms
and avoid common terms. Try a few specific terms and use
the subject or other limits provided in the results.
- International variations of spelling include:
- color/colour - try: (color*|colour*)
- center/centre try: (center*|centre*)
- visualize/visualise - try: visuali*
- usable/useable - try: (usab*|useab*)
Example (converting failure to success):
- questionnairs for software evaluation
- Spelling error, common words, result in no matches.
- questionnaires for software evaluation
- Correcting the spelling error finds some records,
but the common words reduce the number.
- questionnaires software evaluation
- Just removing a low-content word like "for" improves results.
- questionnaires evaluation
- The term "software" matches thousands of records, but only about 10% of the HCI
bibliography; words like human, computer, interaction, software, etc. apply
to almost every record, whether they are present or not,
so they are best left out of searches.
The correctly-spelled reduced query gets good results, but wildcards would find more.
- questionnaire* evaluat*
-
To find records with "questionnaires" or "questionnaire", use "questionnaire*".
To find records with "evaluation" or "evaluations" or "evaluating" or "evaluate", or "evaluated", use "evaluat*".
The search with wildcards matches more than twice as many records.
- questionnaire* evaluat* -(questionnaires evaluation)
- Subtract the records that match without wildcards
to see what would have been missed.
- questionnaire* evaluat* [journal articles]
-
Reduce the number of matches by limiting the results to journal articles..
Search Features
- Search Box: Enter terms separated by spaces.
- Clear Button: Clears all your options.
- Clear Limits: If limits are used,
clears those without clearing your options.
- Limit:
Shows frequent terms, authors, dates, and publication sources.
Choose a value and the results will be updated.
- Type:
Special limit based on record type.
Popular types include books and websites (internet).
Unpopular types include journals
(these are about journals, not articles in them),
volumes (about a volume of a journal, not articles in them),
and proceedings
(about conferences, not the papers in them).
Display Features
- Sort:
Sort results by publication year, first author, or title.
- Highlight:
Shows search terms in the results in different colors
to help you see why a record was matched.
- Record (number):
Shows which record number you are viewing at the top of the page.
- Pagesize:
Controls the number of records per page.
- Format:
- Tabular: One record per line. Column headers sort records.
Hover over hotlinks for more information about what they do.
- Brief:
Publication information. No abstracts or keywords.
- Detailed:
All record information, formatted with HTML.
- Raw:
All record information, in modified Refer format.
- Hidden:
No record information is shown, but results summary is.
- Debug:
Not for everyone. Maybe not for anyone.
Information used in the information retrieval process is shown.
Slightly more interesting for wildcards because all
matching index terms are shown.
- Number Records:
Numbers are placed next to records.
- xyz:
- Pagination Navigation:
The usual
first, previous, next, last, and an indication of which record numbers are shown
if there are enough records.
The "Jump to" shows years if records are sorted by years,
letters if records are sorted by author or title.
Privacy Policy
Current settings (search terms, limits, options) are saved in a cookie.
These can be cleared by clicking the Clear button.
Records are kept of searches (date, time, IP address, search-types, and limit usage),
but not of the specific terms or limits used with one exception:
searches matching no records are compiled to detect bugs and improve results.
Accessibility
- All text can be resized.
- Links to skip to the search form and search results are shown at the top of pages.
- Label tags are used to associate text with form elements.
- Title attributes are used to label record parts
- Accesskeys include:
- s - Search
- c - Clear
- h - Help
- f - First page
- n - Next page
- p - Previous page
- l - Last page
- Images, although not used, would have alt text.