Predloga:Internet Archive author

Iz Wikipedije, proste enciklopedije
Jump to navigation Jump to search
Documentation icon Dokumentacija predloge[predogled] [uredi] [zgodovina] [osveži]

Internet Archive author is a template to generate links to Internet Archive for works by or about a person. It is similar in purpose to {{Gutenberg author}} and {{Librivox author}} though not limited to authors. Musicians, film directors, etc.. any person or entity with a work on Archive.org

Uporaba[uredi kodo]

New storage arrives at Internet Archive.

The template is used like this:

           
==Zunanje povezave ==
* {{Internet Archive author |sname=Full name}}
  • Use the "Full name" without honorifics. For example:
The template can work without sname because it will default to the article title, however this is not recommended because article titles might change in the future for dab reasons. Thus, lock in the search name by hard coding sname.
  • If it's returning many false positives (works by other authors), try the sopt=t option to tighten the search. For example:
Results in many false positives.
Result has no false negatives and no false positives (good).
Tightened search has the danger of creating false negatives. Thus always compare, as above, before using sopt=t .. the goal is to have zero false negatives while creating as few false postives as possible. See section below Tips on searching.
  • If no birth-death dates are generated in the URL (ie. missing dates on Wikidata), they can be hard coded in the template using the birth-death options (see below).
  • If the name has multiple spellings use the custom search option. Example for Jakob Böhme:
More examples in the Custom search and Tips on searching sections below.
  • If the person has multiple names, such as multiple pseudonyms, use multiple instances of the template. If the names have commonalities (such as the same last name), try custom search.

Optional parameters[uredi kodo]

  • sname: Change the search name only. The name used when searching IA.
  • Useful when the article title is a shortened version of the name.
  • Or when a refined search is needed:
  • dname: Change the display name only.
  • name: Change both the display & search name. Normally defaults to the page title. Useful in a number of scenarios:
  • The page title is the real name of the person, but on Internet Archive it uses their pseudonym.
  • To search on multiple names such as multiple pseudonyms, use multiple instances of the template. Or see custom search below.
  • birth: Change the birth year. Normally defaults to the year found in Wikidata. If no dates are showing in the search string, and there is an available birth date for the article, it means Wikidata is missing. You can either 1. Update Wikidata (recommended) or 2. Hard-code the dates in the template using the birth= and death= options. For updating Wikidata, the Wikidata page for Joseph Conrad shows an example entity that has complete data. For incomplete entities, the Property for birth is P569 and Property for death is P570 and can be added by clicking "Add" at the bottom of the "Statements" section.
  • death: Change the death year. See comments for birth.
  • media: Specify a media collection(s) to search. Default is all of them: texts audio video
Dela avtorja ali dela o avtorju: Edward Elgar v Internet Archive
  • coda: Special message at end.
Dela avtorja ali dela o avtorju: Claude Hagège v Internet Archive (scanned books)
  • search: Custom search string. See section below for examples. No other options work when using custom search (except for dname), options must be hand coded in the search string.
  • sopt: Search options:
sopt=t (or sopt=tight):
This will make a more specific "tight" search reducing false positives, but at the risk of creating false negatives. Always compare results with and without a tight search.
sopt=tx:
Used when a name contains extended-ascii (ie. accented letters) such as "Stéphane Mallarmé", and sopt=t still gives too many false positives. This option is used very rarely (in 9,000 names it was needed only a dozen times). It can create false negatives when used incorrectly, always compare results without first.

BC dates[uredi kodo]

Currently the script does not automatically determine if a date is AD or BC eg. Category:63 BC births. This won't matter too much when searching Internet Archive. When setting dates via the options above, leave off AD/BC designator.

Custom search[uredi kodo]

It is possible to use a custom search. This is mainly useful when a person is known by multiple names, but also other scenarios (see examples).

A custom search is surrounded by () parenthesis when passed to the template. If Internet Archive returns an error about the 'search engine not working and try again later', it typically means the search parameter isn't enclosed in ()'s or the search string is otherwise malformed or has a syntax error.

Example custom searches:

{{Internet Archive author |search=("Abbot of Eynsham" OR "Ælfric of Eynsham" OR "Aelfric of Eynsham")}}
..but this is easier and better results:
{{Internet Archive author |search=( Eynsham AND (Abbot OR Ælfric OR Aelfric ) )}}
Searches for both "Aylmer Maude" and "Louise Maude"
  • For Saint Augustine: {{Internet Archive author|search=("Bishop of Hippo" OR "Saint Augustine" OR "Augustine, Saint" OR "Aurelius Augustine" OR "Aurelius, Augustine" OR "Saint Austin" OR "Austin, Saint")}}

Tips on searching[uredi kodo]

The goal when searching Internet Archive is to have zero missed works (false negatives or FN) and as few works by other authors (false positives or FP) as it takes to reach the goal of 0 FN. In most searches there will be some FP because it's unavoidable when searching 7+ million works. However the question is how many FP. Typically if the number of FP are 20% or less of the total results it is acceptable.

Here are some techniques to create searches that have 0 FN's and few (though not 0) FPs.

1. Many people have multiple names (birth name, pseudonyms, abbreviated names). Create template entries for each name, including a mirror template for each one using sopt=t. Preview the article - for each template entry click through to Internet Archive and note the total number of results at the top of the page. Scroll through and look for correct results and see which gives the most. Then scroll through and look for false positives (they usually appear towards the end of the list. This works best in the non-Beta site since search terms are highlighted in the search results.) Often if the sopt=t and non sopt=t entry give very different total result numbers, then sopt=t is the best choice. Once the best template entry is determined, delete the rest from the article and save. It's important to do this preview method in the article itself and not a temp page so that dates work correctly.

2. If technique #1 doesn't work you may need to build a custom search. See the custom search section for examples, but here are some general rules of thumb. Keep in mind searches are literal strings, and if the search term has multiple words it should be surrounded in quotes.

For a 3-word name where each word has multiple possibilities:
  • |search=( (first OR first} AND (middle OR middle) AND (last OR last) )
Most of the time there is usually only 1 word that multiple possibilities so something like:
  • |search=( first AND (last OR last) )
If there are still problems, consider including book titles in the search:
  • |search=( first AND (last OR last) AND ("book 1" OR "book 2")
This technique is prone to false negatives since there may be journals or anthologies not known about ahead of time.
If there are not too many false positives, simply filter them out using the negative creator switch (-creator). For example:
  • Frank Henry Mason: |search=( Frank AND (Henry OR H.) AND Mason AND (1875 OR 1876) AND -creator:(Plautus OR Holcomb OR Zephyrin OR gordjenkins) )
In this example, the "-creator:" tells search to eliminate any creators with those names from the search. Note also the inclusion of two birth dates, sometimes Internet Archive records have incorrect birth data and this will match them.

Information[uredi kodo]

  • For technical notes see the Lua source code.
  • The produced URL is Protocol Relative. See WP:PRURL.
  • Internet Archive as of 2015 has two sites, a regular and Beta. The regular and beta site use the Lucene search engine for which this module is optimized.

See also[uredi kodo]