Web Magazine for Information Professionals

Search Engines: Goodbye Alta Vista, Hallo AllTheWeb

Phil Bradley's regular column.

We all have our own particular favourites, be it football team, car or brand of tea. The same can also be said of search engines, and for many people at the moment it seems to be Google. On the courses that I run I usually ask the question ‘what’s your favourite search engine?’ and the two that are most often mentioned are Google and Yahoo! I also have a little questionnaire on my site [1], and that’s one of the questions I ask there. The current results are:

These figures also illustrate the same results, and while it’s interesting that Google is so far out in front, it also shows quite clearly that information professionals are now making less and less use of AltaVista. At the Online Conference [6] in London recently I attended several talks on search engines and workshops regarding them, and AltaVista was barely mentioned. However, even just a couple of years ago AltaVista was highly regarded as perhaps the best of the free text search engines; it had a large database which was regularly updated, it was also constantly updating and adding new features and its search syntax was very flexible. Yet now it’s being seen as an also ran and on at least one newsgroup that I take (alt.internet.search-engines) the majority of web authors say that they hardly pay any attention to it.

Is this disenchantment justified, and if so, is Google the only alternative that information professionals should move to? A quick and very unscientific series of searches produced a slightly confusing picture, as can be seen from the following table:

 

Search term

AltaVista

AllTheWeb

Google

Internet

114,353,308

55,183,553

99,100,000

mp3

15,422,080

6,309,504

14,500,000

“Everton Football Club”

2,930

4,454

5,060

+”Christmas hamper” +turkey +domain:uk

22

287

37

Pages including ‘internet’ added in the last month

16,923

112,854

88,000

AltaVista still appears to out perform both AllTheWeb and Google on a couple of the very large searches, but for a smaller, more specific phrase search, and on a very specific multi term search with a country qualifier, it was beaten into third place and significantly so by AllTheWeb. Finally, it did very poorly when looking at the number of new or updated pages in the last month. However, none of the three engines that I looked at appeared to be the out and out winner, but the result that gave me greatest cause for concern was the final one; a search engine is only as good as its results, and if these are not updated on a regular basis it really does become necessary to start considering alternatives.

I chose to look at AlltheWeb’, otherwise known as 'FAST‘rather than Google, if for no other reason than many people are already aware of what Google can offer, and I thought it would be more interesting to concentrate on a slightly lesser known search engine, but one which is increasingly being mentioned these days. AlltheWeb is owned by Fast Search & Transfer ASA (FAST), a Norwegian company. FAST claims that it has over 625,000,000 web pages indexed, which is certainly an acceptable size and is comparable to AltaVista and Northern Light, but still lagging some way behind Google (also demonstrated in my quick searches listed in the table above). However, it is making considerable claims for both the freshness of it’s data – it claims a rate of between 9-12 days, though Danny Sullivan of Search Engine Watch [7] disputes this, though not by much, which puts it way out in front of the other major engines – and for its news stories, claiming in a press release “Indexing up to 800 news stories per minute and real-time indexing of news stories from over three thousand online sources, AlltheWeb is twice as fresh as Google, with six times more news.” [8]

The main search page is very clear and uncluttered, consisting of a single screen, which makes a change from the confusing approach taken by AltaVista, while providing more immediate functionality than Google. The user has several immediate options; a choice of language to search in (almost 50 different languages), the search box itself to enter terms, a tick box to tell the engine to search for the exact phrase, and options to search for web pages, new, pictures, videos, MP3 files and FTP files.

The search box can be customised to change the default data set searched (from web pages to news for example), from languages to search type (any, all, phrase search), language settings, an offensive content reduction filter that can be turned off or on, query re-writes (the software can take the input query and re-write it to achieve more accurate results, which could prove very useful for novice searchers) highlighting search terms, and opening results in a new window. However, this is only a partial list – AllTheWeb has one of the most customisable interfaces I’ve seen in a very long time.

Advanced search features were less impressive however, although to be fair no worse than most, but I think Northern Light [9] still reigns supreme in this particular arena. Advanced search options include word and domain filters, page restrictions and display options. Many of these ‘advanced search features’ could be achieved using either the simple search function or the customised options.

The Help facilities are fairly basic as well; very little more than a guided tour of the interface, which to be fair was very clear, and would be of use to a novice, but it was not comparable to the extensive facilities one can find at AltaVista for example.

The search results page more than make up for some of the other less exciting features. The screen is clear and uncluttered, with none of those ‘featured sites’ that are becoming increasingly common with other search engines, such as AltaVista. At the very top of the list of results are a number of ‘Beta Fast Topics’ which are a dozen or so specific topics related to the results retrieved – rather like the Northern Light customised search folders, and which provide the same function – a quick way of narrowing a search down to a smaller tightly focussed group of pages. AllTheWeb provides a brief summary of the page returned, the size, and if appropriate, the opportunity of retrieving more hits from that specific site, using the same approach that AltaVista uses. Another nice feature is that even if the user does a search for web pages, a small box pops up on the right hand side giving the results of a multimedia search, with an indication of the number of hits found for images and video.

In conclusion therefore, AllTheWeb combines many of the best features of other search engines, with few of their disadvantages. That combined with the freshness of its data does make it look a very attractive alternative to Google, and worryingly for AltaVista, a very viable replacement for their own offering. I suspect that in the coming few months I shall be paying rather more visits to AllTheWeb, and rather less to AltaVista.

[1] Phil Bradley’s website http://www.philb.com
[2] AltaVista http://www.altavista.com
[3] Google http://www.google.com
[4] Yahoo! http://www.yahoo.com
[5] Ixquick http://www.ixquick.com
[6] Online Information 2001http://www.online-information.co.uk/online/
[7] Search Engine Watch http://www.searchenginewatch.com/searchday/01/sd0725-fast.html
[8] FAST press release http://www.fastsearch.com/press/press_display.asp?pr_rel=106
[9] Northern Light http://www.northernlight.com

Author Details

 
Phil Bradley
Email: philb@twilight.demon.co.uk