Anatoly Vorobey (avva) wrote,
Anatoly Vorobey
avva

как искать информацию?

Написал подробный ответ на форуме lesswrong.com в дискуссии о том, как люди ищут информацию в сети. Скопирую сюда, но учтите, что это по-английски и в первую очередь касается поиска информации по-английски. Это мое представление о том, как примерно работает мой "алгоритм поиска"; весьма вероятно, что я не вспомнил какие-то важные пункты или источники, но в основном, думаю, не погрешил против истины. Если хотите поделиться своими "алгоритмами" поиска полезной информации или ответов на конкретные вопросы, добавить к моему, спросить или покритиковать - пожалуйста.

I think I do very different things based on what it is that I want to find out. Some of those things are (as a grab-bag of items with no claim to generality):



  • an initial Google search biased towards quickly landing on the relevant Wikipedia page. Whatever I want to find out, there's probably a Wikipedia page to give me rough initial context, and I have a vague idea what it should be called. If it's not on the first page of Google results, that by itself is a signal. Do I misunderstand the domain? Is it possible I'm misspelling or misremembering the key search terms, or they're ambiguous in ways I didn't consider? Maybe the domain is more obscure than I imagine and genuinely doesn't have a wiki page? Is my query too narrow? - try removing a word or two. Still nothing useful from wikipedia? - try adding site:wikipedia.org.

  • Assuming I found a useful WP page: scan it quickly, taking in familiar signals that hint at its quality. Does it look controversial, edit-warred? Is it a placeholder or a well-developed page? If something hints at controversy, scan the ToC of the Talk page (don't read through it all). Are there obviously useful outgoing links? - follow them now (I always open all links in new tabs with middle-click). Does it cite academic papers that look like they might answer my question? - open those links now.

  • if I had a wiki page to orient myself but still don't have an answer, I probably have a much better idea what to search for, and an understanding of how contentious the subject is (if at all). Things that are especially helpful to know at this point: 1) is there a rare word (usually a technical term) that very likely co-occurs with the answer I'm looking for? 2) is there a phrase that very likely occurs in the answer I'm looking for? 3) is it something likely to be in a printed book? 4) is it something likely to be in an academic paper?

  • go back to Google. Make a better query, using narrower, rarer words or phrases. Phrase search (with quotation marks) is especially helpful as it cuts away the junk results like nothing else. Often when I'm unsure how my ideal result would phrase my answer I'll quickly try several phrase searches in a row. Possibly do a Google Scholar search. Possibly do a Google Books search. In those two cases, again, phrase searches are very helpful. If the first result page doesn't have useful links, or even if it does but it looks like there may be more, keep in mind the possibility of paging more. I sometimes scan through 5-10 search pages quickly (just looking at titles/snippets) before giving up on a particular search query; that often helps with phrase searches.

  • if I don't have enough context: add "FAQ", "tutorial" etc. to search terms. Search for FAQs/tutorials in the field, if there're many, open a few, quickly scan to assess quality, find the best one, read it. Always keep an eye out for too-biased-PV signals.

  • suppose I found an academic article or a book search results that look like they might answer my question. See if I can find the actual texts of articles/links. Books may be free/largely previewable on Google Books. If they're really old, could also be in Gutenberg or on archive.org. Adding site:archive.org often helps (Google indexes ASCII versions of books stored there, but often doesn't rank them high). Academic articles: based on field, perhaps look in one of large repositories (arxiv pubmed etc.). If author likely to be alive, Google them, find their homepage, look for downloadable articles.

  • suppose I have an academic article(s)/book(s). Quickly scan the article or relevant chapters in the book. It doesn't have my answer and/or talks about something related but not quite what I need. Look into references for review articles. Make snap judgements based on titles of other articles in references whether they'll bring me closer. For every title I'm interested in, copy the title and open a tab with Google search on it to remember I want to find/scan it. Found a review article? Read it slowly and closely, find the right references, follow those; at the very least update my idea of the right search query.

  • am I looking for something I might've seen on some site or was probably discussed on some site? Use site: operator. E.g. if it's at all related to memes or discussions of everyday culture, search on site:reddit.com, make up likely phrase searches.

  • remember special repos of knowledge in particular fields that are better than WP. Stanford encyclopedia of philosophy. PlanetMath or Wolfram Alpha for math. Arxiv. Stackoverflow and its satellite network of other like sites. Probably lots more I'll remember in the time of need but can't think of right now.

  • is my question likely to be answered in a textbook and I know what kind of textbook or even the exact title? Use Google Books or [omitted] pirate sources to find textbooks and quickly scan them. Not sure if textbook is at all good? Scan its Amazon reviews, though remember it's alright for a good academic textbook to have just a handful of them. Possibly "search inside this book" on Amazon (rarely helped me). If looking for the right book in a field, use "also bought/also viewed" to quickly amass a list of candidates on Amazon, scan through reviews looking for the really helpful ones, they're easy to spot.

  • is my question likely to be answered/answerable on a forum? Use Google Groups search to mine Usenet (again, phrase search helps). Use really good forums you know and site:-search them (e.g. physicsforums.com for physics questions, pianoworld.com for anything piano-related. Consider actually posting a query to a forum if you know it's frequented by knowledegable friendly people.

  • What else? Google Images if you need to see how something might look. m-w.com as the first stop for English vocabulary, urbandictionary.com for modern slang, OED for serious dictionary-digging, analogues of all these in other languages. Google News search for anything might have been covered by mass media recently. NYTimes archive search for anything historical NYTimes might have written about.
Subscribe
  • Post a new comment

    Error

    default userpic

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.
  • 10 comments