Patent searchers need to be mindful to branch out past English-language patents during a prior art search. Relavent prior art may come in any format and in any language, so searchers need to utilize techniques and resources to access prior art in any and all available languages and formats. Sobotong is a new bilingual search engine that can help searchers identify non-patent literature prior art in both multiple languages and formats. A recent post from the Beyond Search blog gave a brief review of the Sobotong bilingual search engine, which was created by Qiang Wang. Although the Sobotong website offers some unique search options that may help the user identify multilingual relavent prior art, the site also suffers from one main problem: Sobotong utilizes machine translation when conducting the multilingual searches. Professional patent searchers who fluently speak multiple languages will always be able to conduct more thorough searches than a system utilizing free machine translation services to conduct a multilingual search.
After the jump, learn about the unique features of Sobotong that allow the user to identify prior art in multiple languages and formats, and find out the down side of using a multi-lingual search platform!
The Sobotong homepage appears to indicate that the search and translation functions for the system are powered by Bing and Google. On the Sobotong homepage, users can enter a search term (in English) and select from two drop-down menus which languages to search in. The user can select from 64 available languages. In the screenshot below, the two selected languages are English and Japanese.
Search results will display in English, with the search term translated into the two other selected languages displayed above the English search results. The user can select the translated terms to view search results in the selected language. If the user has searched for a term in English and selected two non-English languages, then results will initially appear in English. However, if the user selects one of the non-English equivalent search terms, then the non-English search term will be automatically entered in the search form. The user will then be unable to return to the English search results unless they re-do the search.
It is therefore recommended that users choose one of the two languages to search in simultaneously as English, so that they can always select to return to the English-language search results through the English search term link.
In the left side menu, the user can select to view a variety of search result types:
- Web – Some result listings include thumbnail images. Individual result listings include a site title, an excerpt, and a URL.
- Image – Only a thumbnail image is displayed in the result listings.
- News – An icon indicating the news source, article title, URL, and excerpt are included in the result listings.
- Music – The site title, URL, and excerpt are included in the result listings. An option to filter results by website is included by each search result.
- Video – A video title and thumbnail image are included for each search result.
- Movie – The site title, URL, and excerpt are included in the result listings. An option to filter results by website is included by each search result.
- RSS – The site title, URL, and excerpt are included in the result listings. An option to filter results by website is included by each search result.
- Document – The site title, URL, and excerpt are included in the result listings. An option to filter results by website is included by each search result. Users can also choose to filter results by document format (.pdf, .doc, .ppt, .rtf, .txt, .xls).
- Software – The site title, URL, and excerpt are included in the result listings. An option to filter results by website is included by each search result.
- Google – The search term is transfered directly into the Google search form.
Within the result listings, the search term is bolded in the result title, URL, and excerpts. The user can scroll to the bottom of the page to navigate to the next page of search results. Selecting a search result will open the link in a new tab.
On the right side of each result page, a list of related search terms is displayed (only for some terms, mostly in English). The user can select a related search term to automatically enter that term in the search form.
- Search terms are automatically translated and searched in up to two languages, so the user can enter their original search term in English.
- Related search terms are suggested for some search terms.
- Multiple result types are available, and the user can view each result type for each set of language results.
- The user can filter document results by format type.
- Google ads are displayed at the bottom the the search page.
- No “help” section is available to give users tips on utilizng Boolean/truncation operators and correct search syntax.
- It is unclear whether the search/translation functions are powered by Bing or Google.
- Few non-English results seem to be available under certain result types (News, Music, Movie, RSS, Document, Software) during some test searches (for terms like “centrifuge” and “iron”).
- Machine translation is used to translate search terms, and as previous blog posts have discussed, machine translation is not as accurate as human translation.
A professional patent searcher who can fluently search in multiple languages will always produce more accurate and thorough prior art results then a machine-translated multilingual search system. Sobotong still provides a useful interface for English-language searchers who want to conduct an initial search for non-English prior art, and the system has the added bonus of allowing the user to search through multiple document formats simultaneously. Using Sobotong for a multilingual search will certainly be quicker than searching Google in English, translating the term using Google Translate, and searching again in the new language. Sobotong provides no help literature, though, so the user is on their own when figuring out how to correctly format queries.
Do you know of any other search systems that let the user search in multiple languages simultaneously? Let us know in the comments!
This post was contributed by Joelle Mornini. The Intellogist blog is provided for free by Intellogist’s parent company Landon IP, a major provider of patent searches, trademark searches, technical translations, and information retrieval services.