Enquire within on everything: getting questions answered on the Internet

Online Currents 2003 – 18(2) 22-24

---

This article compares several question-answering services on the Internet: the paid Google Answers service, the free Usenet Newsgroup system, and AskNow!, a free collaborative reference system run by Australian state and national government librarians.

Introduction

During its period of massive growth, from about 1992 to 2000, the Internet was largely supported by unpaid volunteers: hundreds of thousands of people gave up millions of hours of free time to support and encourage others. Although this still continues today, there are indications that the tide is turning and that a cash-for-service expectation is developing. Where a free service and a paid service offer the same results, the economic pressure on the paid service has often resulted in its closure. Paid service providers may mount an organised attack on free services through lobbying; conversely, free service providers may look to find ways of charging for their efforts.

The Google search engine system is a case in point; beginning as a free service, it has now begun to sell placements, (although these still remain distinct from the free results). It has also branched out into a range of different services which relate to its core business in interesting and innovative ways. In this article I will look at Google Answers, the paid question-answering service it now provides. I will compare this with a ‘traditional’ Internet approach to asking and answering questions through the Usenet Newsgroup system, and with AskNow!, a new service provided free by Australian librarians from State libraries and the National Library of Australia

Google Answers

The basic premise of Google Answers (answers.google.com) is simple: people post questions, and indicate how much they are willing to pay for an answer. Anyone who wishes to answer the question for that price can do so, and if the answer is satisfactory they receive the payment, less a 25% fee which goes to Google for managing the service.

Where money is involved it’s important to develop and maintain trust. Google attempts to do this by establishing and maintaining the system as a ‘community’. Questioners are required to log in and provide details about themselves. Answerers (known as ‘researchers’) must do the same, plus successfully completing a test to establish their credentials. Researchers receive a weekly email newsletter discussing the system and any changes, and have access to a busy chat room. They are gently reminded to be courteous and comprehensive in their answers, and warned against using ‘bots’ to reserve questions. Participants who seriously violate house rules are banned from the system (although as is usually the case on the Web, there is nothing to stop them rejoining under another name).

Other aspects of management include ‘locking’ questions for two hours at a time while research is being carried out to prevent researchers gazumping each other; and allowing for the posting of comments and requests for clarification by both questioners and researchers. Each question is assigned a category in the Google directory so that researchers have a starting point; although this is not necessarily going to help since most Google questions cannot be answered by simple directory searches. Questions can be given an expiry date which they need to be answered by. Google arbitrates any disputes between questioners and researchers, and screens answers and comments to preserve the anonymity of the participants and enforce house rules.

Google has opted to make the whole system open, so that questioners and researchers can examine questions that have been asked in the past, and all the comments and requests that have been made. Questions, comments and answers can also be searched for keywords or viewed by category, so before posting a question (or answer) of your own you can see what other people have had to say on the topic.

The Questions

Three things strike the answerer viewing the current list of unanswered questions: firstly its size; secondly the diversity of the questions; and thirdly the sheer difficulty of most of them. At the time of writing (November 2002) there were over 4,000 questions listed, some 1500 of which still required an answer. Most of these had attracted at least one comment from researchers and some had eight or more. The oldest unanswered question on the system (about the artist Edith Harrington) dated from April 22 and was thus more than eight months old, but the comments indicated that some work was still being done on it as late as October.

The motives behind the questions are a mixed bag. There are:

• purely commercial questions (‘Cheapest and reliable way to transfer money & stocks from Germany to US’)
• technical queries (‘Outlook 2000 – Error emailing to Contact’)
• personal issues and enquiries (‘Help with non-identifying adoption record info in Phoenix Arizona’)
• enquiries about books, movies, songs, etc, of personal interest (‘Soundtrack to “Noises Off” film’)
• enquiries about where or how to obtain goods and services (‘Looking for a compass used indoors’)
• homework or exam questions (‘George Washington’s War Style’)
• puzzles from other sources (‘A difficult cricket poser’)
• miscellaneous questions that may result from bets or disputes (‘Did the Supreme Court of Florida mock the US Constitution?’)
• self-referential questions to test the system and challenge its participants (‘What is the answer to this question?’)

As many researchers have discovered, asking successful questions requires nearly as much training and practice as answering them. A large number of the questions as originally posed are poorly structured and difficult to understand, although this seems to have improved over time. Luckily there is an option for questions to be clarified by the questioner. The questioner may also rate the answer to the question, awarding stars for one that meets their requirements extremely well.

Finally, most of the questions are very difficult; in fact Google Answers seems to function as a court of last resort for people who have attempted to find answers themselves using search engines and other means but without success. The odd easy questions are snapped up and answered almost immediately, so anything that’s been on the system more than an hour or two is almost guaranteed to be a tough nut to crack. Specialist knowledge may help in some cases, but the primary skills required by researchers appear to be web searching abilities, lateral thinking, an organised mind and endless patience.

Some question types are more likely to be answered than others: I noticed that questions involving lists were less popular with researchers than ones that could be answered with a paragraph or two of text. One question I posed myself, asking for a list of online detective fiction, went unanswered altogether.

Prices and tips

Note: figures given in this section are all in $US.

Google has chosen to set the minimum question fee at $2, making it open to nearly everyone and providing a strong challenge to the competing free systems like those described below. At present there is little correlation between the fee offered and the amount of work required to answer a question, though this may be changing over time; there are signs that researchers are beginning to bypass questions where the fee is wildly inadequate for the work involved. Payments so far have remained fairly constant, rising from an average of $14.72 for 25 questions in April to $15.88 for the most recent 25 questions. Questioners can repost a question with a higher fee if it is not answered; alternatively, in a recent innovation, they can choose to give cash ‘tips’ to researchers who provide first-rate answers. The maximum question fee is $200.

Some experienced researchers claim to be making over $1000 per month. A more credible estimate that I obtained first-hand was about $350 for 100 hours’ work. This is well below the average wage in developed countries and – given the expertise and talent demonstrated by most researchers – much less than most of them would be making in a regular job. But for those who are unable or unwilling to find regular work, who prefer to work from home and who have a passion for research, the money may make an added incentive for research that they might otherwise be doing for free. Some researchers are living in less developed countries where the payment represents a reasonable income for the work involved. However, to qualify as a researcher requires a fairly high standard of English literacy, reducing the opportunities for researchers who don’t have English as a first language.

Comments and clarifications

Virtually all questions that are unanswered for more than a few hours attract comments. Anyone on the system may comment on a question, whether they are a researcher or just another questioner. Combined with requests for clarification and clarifications themselves, the comments provide a detailed and often fascinating picture of the research process in action, sometimes extending over hundreds of lines of text and months of time. Although there is no formal method for collaboration among researchers, this is often done informally through the comment system, with one researcher offering suggestions for lines of enquiry that another will follow up. Since comments are unpaid, this suggests that the researchers are driven by personal interest and altruism rather than financial motives.

A researcher may take on informal ownership of a particularly hard question and worry at it until it is answered or shown to be insoluble. Even when a question has expired or been withdrawn by the original questioner, it may still elicit comments and attempted solutions. The largest number of comments for a question approaches 100, for a difficult puzzle relating to US states. A more typical question, about where to get chewy root beer candy, elicited 8 comments and a total of 316 lines of text.

The future for Google Questions

Collaborative projects on the Internet often attract an enormous amount of enthusiasm at the beginning which tapers off over time. Whether Google Questions becomes an exception to this rule will depend on how well it can balance the cost to questioners and the returns to researchers. Keeping the cost to questioners low may eventually drive away skilled and talented researchers to more rewarding work elsewhere; but raising returns to researchers could cut the flow of questions down to untenable levels, particularly in the face of competition from free research systems such as the ones discussed below.

Usenet Newsgroups

The Usenet Newsgroup system is a long-established means of providing access to discussion groups. The groups are circulated world-wide, through a large number of news servers, usually associated with internet service providers. Each of the discussion groups, of which there are currently over 100,000, is ostensibly related to a particular topic, where questions and comments about that topic can be posted for others to read and respond to.  Unfortunately newsgroups are often given arbitrary names that may bear little relation to the topics discussed in them. (For instance, the ‘alt.rec.crosswords’ group was originally set up to discuss Scrabble™, but crossword puzzle fans who were misled by the name eventually came to dominate the group and its discussions.) And since newsgroups are categorized only by name there is no guaranteed method for finding a newsgroup on a particular topic or even establishing that such a newsgroup exists.

Newsgroup names arise by fission; starting with a relatively small number of groups covering a few major topics (‘rec’ for recreation, ‘soc’ for social issues, etc), new groups are added by tacking on a suffix preceded by a full stop: thus rec has spawned  rec.arts, rec.books, rec.music, rec.games, etc; rec.arts in turn has spawned rec.arts.comics, rec.arts.movies and rec.arts.theatre, and so on. Some active groups are now at the fourth or fifth level of fission. Since anyone can create a newsgroup, and there is no incentive for anyone to remove them, there is a constant flow of new groups, several dozen per week, most of which attract little or no interest and only clutter up the system.

The host who maintains a news server decides which groups to make accessible through that server. This is usually done by guesswork, although most hosts will be happy to add groups at a user’s request.

Being open to the public, newsgroups attract a lot of junk mail, irrelevant comments and scatology. Using the newsgroup system therefore requires time and patience. It can, however, be very rewarding, since the larger, busier, longer-established groups do tend to attract experts in their field who are usually happy to answer questions.

Some companies – notably IBM and Microsoft – have adopted the newsgroup system as a means of software support, and established their own newsgroups relating to their products, with full-time staff responding to postings; it’s quicker and cheaper for them to deal with questions in this way rather than on an individual basis.

Posting a question

I decided to pose a genuine question to the newsgroup system: ‘How can I get files in the Windows Explorer to sort in alphabetical order?’ Using Outlook Express as my newsgroup software and OptusNet as my ISP, I set up a news account and downloaded a list of all the newsgroups currently available to me.

This list I displayed on the screen and filtered by keywords. A search on ‘explorer’ revealed about fifty groups relating to the Internet Explorer, including many in languages other than English, but none on the Windows Explorer. Searching on ‘windows’ and ‘win98’ was more productive, revealing several groups in a variety of languages including Czech, German, Russian and Spanish. I picked out those that were in English and looked most relevant, and checked them for activity:

• alt.windows – 63 postings
• alt.windows98 – 5343 postings
• comp.archives.mswindows.discuss – 4 postings – one of many defunct groups.
• comp.os.ms-windows.misc – 1046 postings
• comp.windows.misc – 143 postings
• microsoft.public.win98 – 84 postings
• microsoft.public.win98.display.general – 135 postings
• microsoft.public.win98.setup – 291 postings

I have italicised the folders which looked the most promising, although they had nothing in them directly relevant to my problem. I addressed and sent a new posting to all four groups as follows:

Whenever I open a file folder in Windows Explorer or My Computer, the files and folders are listed in chronological order. I want them in alpha order, but although I’ve tried to set this several times through View/Folder Options/View/Like Current Folder, this only seems to apply to that folder in that particular session; all other folders are still in reverse chronological order and the folder I’ve changed reverts to reverse chrono order next time I switch on. Help!!

The question was posted at 10:41. Within three hours there was a response in the microsoft.public.win98 group from a Microsoft employee (hold down the control key when closing a sorted file window), and this actually solved the problem. At 2:04 I received another response which had been cross-posted to all the groups in which I had asked the question, but unfortunately this response was not helpful. After twenty-four hours I hadn’t received any further responses. For specific queries within clearly-defined areas, Usenet newsgroups are ideal.

AskNow!

AskNow! is an Australian question-answering service, currently on trial at www.asknow.gov.au. It enlists State and national government librarians from around the country to provide the on-line equivalent of a reference desk. AskNow! first requests  that users provide some personal details – without any real indication why this is necessary, or how these will be used. The user then enters a question, and are told that a librarian will respond ‘shortly’. There’s no indication of what is meant by ‘shortly’, and users accustomed to quick response times may find the term misleading. Another drawback is that questions can only be asked between 9am and 8pm (Eastern Summer Time) on a weekday; obviously it takes more than just a technological revolution to make public servants work nights and weekends!

The first question I posted was: What companies in Australia offer paid or unpaid news alerting services? When there was no response from a librarian within five minutes I gave up and chose to send my question in to the National Library by email – a separate process requiring the question to be entered all over again. Most email queries are answered within five working days; more complex ones can take up to four weeks.

My second question, Why doesn’t the government use cloud seeding to cause rain? evoked a quick response from a librarian in Tasmania called Jane. Our dialogue took the form of a chat session in a small panel at the right of the screen, leaving the majority of the display unused. Confusingly, the dialogue ran from bottom to top. Each of Jane’s responses took several minutes. I posted the question at 4:30 and at 4:40 I was provided with the address of a NSW Government page with links to water resources, but which made no mention of clouds, rain or seeding. At 4:43 I was provided with a link to a cloud seeding page from CSIRO. The process had taken up nearly fifteen minutes of my time and that of a trained librarian.

For comparison I did my own Google web search on the terms ‘cloud seeding Australia’. This produced as its first hit a more comprehensive paper on cloud seeding – also from CSIRO – but failed to find the one I had been directed to by Jane until I added the word ‘experiments’. Time taken to find both pages; about thirty seconds.

I have no doubt that the librarians offering their skills are competent and dedicated, but it seems very inefficient to use them in this way as surrogate Google searchers. If someone is Internet-literate enough to find and use the AskNow! page and patient enough to sit through the dialogue process, they are surely capable of doing their own web searching.

Measured up against Google Answers, AskNow! didn’t impress me. It is slow, it may not be available when you need it, and for my own query I could have obtained the same results myself in a fraction of the time. It uses frames and it crams the active area into a small portion of the screen; it requires the user’s name and address for no apparent reason; and it has an annoying exclamation mark in its name. The taxpayers’ money that is being spent on this could fund an awful lot of question-answering at Google. On the other hand it is free and it may provide encouragement and social support for beginners who can be directed on to the system – perhaps by the staff at their own local library.