Charles Arthur On Technology

Searching for the perfect portal

What was the first computer virus? More importantly, since it's unlikely you'll know the correct answer to this (I certainly didn't), where would you go to look?

What was the first computer virus? More importantly, since it's unlikely you'll know the correct answer to this (I certainly didn't), where would you go to look?

Most likely your reflexive answer to the latter question is "Google". The brainchild of Sergei Brin and Lawrence Page has become the pre-eminent search engine on the net; its name has become a verb to express the act or intent of finding something out from the vast sea of information out there on the net.

But things are changing. Microsoft has made no secret of its intention to wrest Google's crown away, and there are plenty of other companies that have been around longer even than Google: notably Askjeeves.com (more easily known as Ask), which is also in the middle of an upgrade. And besides the web interfaces for searching the internet, Google has launched its Desktop Search program, available from http://desktop.google.com (for Windows machines; Mac users should not hold their breath. The next iteration of OSX, Tiger, will arrive before a Mac version of Google desktop). Microsoft intends to follow suit by the end of the year, and Ask plans a "local search" system, early next year.

So, everyone is offering everything at almost the same time. The inevitable question to ask is: which is better?

Unfortunately, that's not the sort of query that submits easily to being put into a search engine. When you're looking for answers, the best way to start is by seeing how well they find out things that you already know, and working outwards. Obviously, for most of us the place to start is with a little bit of ego-surfing. Putting one's own name into a search engine is usually fun, and in the case of people taken hostage in Iraq the results can be literally life-saving: the captors have in a couple of cases released journalists after cross-checking their claims of who they were against what's publicly available.

Thus I went ego-surfing on Google, Microsoft's search (at http://beta.search.msn.com - the "beta" is the indicator that this is not the finished product), and Askjeeves ( www.ask.co.uk). On Google and Ask, my name, typed as a phrase ("charles arthur") turned up as the top result (linking to my blog); on MSN, I was 26th, preceded by various references to Charles Arthur Floyd (a 1930s bank robber) and a journalist with exactly the same name who is an expert on Haiti. (I'm not related to either.) Which result is "right"? Obviously I find Google and Ask gratifying, but have to wonder whether they're really in proportion.

On to something more sensible. What was the first computer virus? Plug those words into our trio of engines, and let's see.

Google offers as its top result a link to faqs.org, where the answers are frankly rubbish. The second result is a BBC story about 20 years of viruses, which suggests 1983, and Fred Cohen as the author. Ask offers a link directly to a student's university homepage which states it was the "Brain" virus of 1986. MSN throws up four results which, from the extracts, clearly aren't useful, but the fifth points to the entry in the online open-source Wikipedia encyclopaedia, which gives chapter and verse (mostly chapter) about viruses - and suggests the first was one called "Elk Cloner" from 1982.

If it's knowledge you're searching for, you soon discover that using different search engines is like viewing a mountain from different vantage points. It's the same mountain, but the detail is different. All use the same basic technology: the number of links towards a site are taken as a measure of its "worth", which is then juggled across the words it contains and highlights. And as it's not a secret how Google does it, anyone can copy it. All you need is a vast index, robot programs to "crawl" the web sucking the content into that index, and a smart team to find ways of ranking it.

However, one key difference is the use of "sponsored results" - in other words, advertising. Some search engines mix advertisers' paid-for results in with the ones it has normally indexed; this is distracting. Google puts the adverts off to the side, in a coloured box. At Ask, up to 10 "advertisers" results are stacked up above the first "true" ones. MSN similarly piles up sponsored results in a clearly labelled, coloured box above the initial results, and puts another set off to the right of the main results. It's faintly distracting, and when Ask serves up 10 to scroll past, irksome too. But Tony Macklin, director of product development at Ask, argues that often when people search they are looking for something that the companies behind the sponsored results can provide.

Ask does also have some neat tweaks to the searching concept, which it needs, as it lacks the market dominance of Google and the operating-system dominance of Microsoft. For instance, when you search for "Raleigh" on Ask, you get "Related searches": Raleigh bikes and cycles, Sir Walter Raleigh, Raleigh International, Raleigh North Carolina... so that your ill-defined initial search can be channelled towards the subject you're really investigating. It's neat, and a step ahead of rivals for now.

And if you've decided to change how you search, and want to use a different search engine? The easiest way is to change the default home page of your browser (Google offers instructions at http://www.google.com/options/defaults.html; to replace it with Ask or Microsoft's beta, put their names instead of Google).

As for desktop indexing, this can be fabulously useful, if a little confusing when you see a list of files on your own machine listed alongside Google results when you do a search. But Google's desktop search isn't perfect. It will index your web-browsing history, Outlook mailbox (tough luck if you use a different e-mail program) and Microsoft Word files, AOL Instant Messenger chats, and text files from the files and folders in your user area. It won't search inside PDFs. And while it does index the files of everyone on the computer, you can only install it if you already have the "administrator" powers that would let you read those files already. In short, the Google Desktop Search is not as dramatically interesting as it might appear.

I'll return to the subject when there is more competition from Microsoft. But for now, the only conclusion I was left with from my explorations of the various search engines was that there's less between them than you might think - and that while Google has the edge in usability because it doesn't thrust sponsored searches in your face, Ask's "Related Searches" was the neatest idea I'd seen for those floundering in the sea of knowledge. Jeeves would surely have approved.

www.charlesarthur.com/blog