altavista.com

This one's a bit techy - and partly adapted from an essay of mine. But I suppose it might interest somebody!

A few years back, around the time of the 'internet boom,' it seemed that there were two clear leaders in the searching arena. Yahoo was the choice of the masses, while anybody wanting to do some serious searching was pointed towards the mighty AltaVista. Yet between 2000 and 2002, the average number of users per month plummeted ...
...BBC News Online article). Now Google seems to be the first choice for all. The question is how does AltaVista measure up these days – did it deserve to be eclipsed by Google?

First of all, going by the size of the database, AltaVista would still appear to be a good bet. At the end of 2002, it was thought to index 1,689 million web pages. This is less than Google’s 3,033 million but more than Hotbot and MSN. (Figures: http://www.searchengineshowdown.com/) ... more

This one's a bit techy - and partly adapted from an essay of mine. But I suppose it might interest somebody!

A few years back, around the time of the 'internet boom,' it seemed that there were two clear leaders in the searching arena. Yahoo was the choice of the masses, while anybody wanting to do some serious searching was pointed towards the mighty AltaVista. Yet between 2000 and 2002, the average number of users per month plummeted from 16 million to just 6 million. (Stats from a BBC News Online article). Now Google seems to be the first choice for all. The question is how does AltaVista measure up these days – did it deserve to be eclipsed by Google?

First of all, going by the size of the database, AltaVista would still appear to be a good bet. At the end of 2002, it was thought to index 1,689 million web pages. This is less than Google’s 3,033 million but more than Hotbot and MSN. (Figures: http://www.searchengineshowdown.com/) AltaVista is unlikely to have the best range of results, but it is still a broad, useful resource with a global database.

Most AltaVista users will head straight for the basic search. This is simple and intuitive. Search terms are entered in the box, and then it's just a simple case of hitting the 'Find' button. Simple options exist for searching the UK only or searching in various languages. You will find that this works quickly, and is usually quite effective.

By default, websites containing all search terms are returned. These are displayed ten at a time, after a selection of "sponsored links". Each entry can be enhanced with a variety of extra information. As with most search engines, the page title and a brief description based on the first few lines of the document are returned to help the user decide between websites.

Extra help is given to the user trying to decide between results. The web address is displayed, which can be useful to identify likely authorship or authority (for example, health information may be more credible if on a .gov.uk or .nhs.uk domain). Helpfully, recently updated records are also highlighted with a comment in green such as "Refreshed in past 48 hours."

Some useful tools are also accessible:
++ "Related pages" seems successful. Clicking it below the entry for National Museums and Galleries of Wales (NMGW) brought up entries including National Museums of Scotland, the National Gallery and Merseyside Museums. Strangely though, it also decided that the BBC Sport Wales homepage was related. So, this tool is useful but not guaranteed to work!
++ "Translate" takes you to AltaVista’s long-running Babel Fish service from which can translate web pages into different languages. The standard of translation is probably far from perfect, and the number of languages is limited – but it is still a useful tool for extracting information especially as although the page is translated, the original formatting is retained.
++ "More pages from..." returns similar pages from the same domain. This is useful where further details are required from an already trusted source. Looking again at the NMGW example, the "more pages" option suggests the National Museum in Cardiff, the Big Pit Museum, Swansea’s Waterfront Museum and other major sub-sites of www.nmgw.ac.uk. This is a good way of discovering resources buried within large institutional sites.

The search results pages also have a useful addition that known as "AltaVista Prisma". This prompts you to refocus vague searches on suggested search topics. For example, "walking holiday" brings up many valid hits, but Prisma also offers the chance to re-search on topics such as "activity holidays", "walking tours" or "hill walking". It also throws in a few complementary categories such as "accommodation" and "Highlands".

Hidden behind the front page is also a quite sophisticated advanced search options. Confusingly, there are two ways of accessing this. The most obvious is the "more precision" link next to the Find button. Clicking this brings up four possible search fields:
++ "All of these words"
++ "This exact phrase"
++ "Any of these words"
++ "And none of these words"
This is an excellent way to exploit the capabilities of the searching software without having to formulate complex search strings.

The full advanced function is not instantly apparent, but is in fact found just below the search terms box on the homepage. Clicking the "Advanced" link allows searching using the methods mentioned in the previous paragraph, or by using Boolean expressions. This allows complex queries to be constructed, through the use of AND, OR, AND NOT and NEAR. It is also possible to use parenthesis and truncation, which Google does not seem to support. For example, searching Europ* will find Europe, European, Europa.

Further filtering is possible:
++ "Date" specifies what age pages should be returned
++ "File type" allows the user to search for web pages or PDF files
++ "Location" can narrow down results to a specific web server, or to domains such as .com or .uk
++ "Site collapse" will restrict the number of results returned from any one site. This can be very useful. For example, a full UK only search for "Aberystwyth" results in 63,542 results. Of these, 19,951 of these are hosted on the University of Wales, Aberystwyth domain .aber.ac.uk. A user not interested in UWA information when searching for Aberystwyth could avoid being swamped with .aber.ac.uk results with this tool!
++ It is possible to vary the number of results returned per page – in this case up to 50. This can be useful for users wanting to quickly scan a large range of pages.

AltaVista also features an editor-compiled hierarchical directory of sites - though it is hidden away, only referenced to by a small tab above the search box. This can be useful for searching vague subjects or if you're not sure of your search terms, but it can be frustrating to use and is rather time consuming. I would much rather use the search facility, especially if information is likely to be scattered over several branches of the directory.

Today’s Alta Vista devotes much of its front page to deceptively-titled "search centres". Initial thoughts would suggest that these would take you into the AltaVista directory. Disappointingly, they simply link through to advertising. Thus, clicking on "Mobile" doesn’t bring useful mobile phone links. Instead, it brings up three large graphical advert links mobile phone retailers. The 1999 version of AltaVista (viewable at www.archive.org) gave plenty of front page space to listing directory categories. The old site also featured more valuable complementary information such as international news and popular searches.

One other area in which AltaVista once led the field was in the option to allow searching for non-textual information. Searches can be conducted for images, video or MP3 and other audio files. This is fairly effective, concentrating on finding search terms in file names or close to image files in web pages.

So how does AltaVista measure up? Some simple searches returned the following, without restricting to UK only results:

--> "chocolate making"
AltaVista: 3,570 hits. 9 of the first 10 were commercial enterprises who only provided very brief information about their own techniques.
Google: 6,260 hits. More useful than AV. The Hershey Visitor Centre appeared, as did a useful overview page. Still some obscure sites, such a statistics site.
MSN: 4,247 hits. Several shopping sites in the top 10, but some useful information sites.

--> "walking holiday" AND Wales AND NOT (Ceredigion OR Powys OR “Mid Wales”)
(A confusing, complicated search term, that I hoped would find me Welsh walking holidays that weren't based in Mid Wales)
AltaVista: 596 hits. Excellent results – all of the first 10 were in some way relevant, with helpful sites about holidays in Swansea and Pembrokeshire among others.
Google: Couldn’t cope with this complex query. Returned lots of Mid Wales sites. That said, reformulating the search as "walking holiday" Wales -Ceredigion -Powys -"Mid Wales") gave Results comparable to AltaVista.
MSN: Only 702 hits, but the top results were very good again.

--> Image search for London
AltaVista: 336,153 images. Of the first 15, all were relevant. A good mix: Tower Bridge, Buckingham Palace, a Tube map....
Google: 519,000 images. Of the first 15, there were some very strange inclusions such as a graph and a bald man. Still, there were several useful images and my search was rather vague!

In my searches with AltaVista, I was pleased not to find many broken links at all. Most results were live web sites, even if not always 100% relevant.

In general, AltaVista is still a very good search engine. It consistently provides a good spread of results from a variety of sources. It also has the potential to be more powerful than Google for advanced users. It does suffer from a bias towards commercial sites, but from my own experience it seems to have improved from say, 2001, when a lot more irrelevant commercial sites were returned when searching.

I would have to admit that I will probably continue to default to Google myself, due to the speed of the service and the huge database. However, I am sufficiently impressed by AltaVista to consider giving it a few more tries in the future, and am open to changing my searching habits if it proves useful!

 

...less