OK; so right now I should be connecting my laptop to work’s VPN and checking up on a database I’m supposed to be diagnosing. Or, failing that, I should be tidying the flat, taking the rubbish out, or thinking about cooking my dinner. Or, failing that, I have some server maintenance, home directory housekeeping and CD ripping that I’m going to have to do eventually. But I’ve had an arse of a day and I’m feeling double-plus-lazy, so I’m going to write something instead. A quick look back in time tells me that I’d promised a second opinion on Mahalo once I’d had time to try it out, and since I’ve had time to try it out; here it is.

The first thing to say is that it’s no google, and I’m guessing they’re not even going for that market. I’ve found it to be close to useless for any search I carry out in the course of my working (as a software engineer) day. Pasting error codes into the search form just throws out a random selection of (usually biographical) pages. The google results at the bottom of the page are as good as you’d expect, but they’re slower to load and less obvious on the page than if I’d just used google in the first place. Searching for information on unix commands leaves it scratching it’s head in bewilderment, and, perhaps more seriously, even relatively well known computer software can leave it similarly speechless.

And it’s not just technical searches, it’s anything … overly specific. If I search for “cat lifespan” it gives me a few options for pages about cats, but nothing that is obviously going to tell me what I want to know. Google, by comparison, gives me the answer in it’s first link (although it’s second one is to a corporate website.) I guess expecting answers to specific questions is just too much to ask of the site’s human indexing engine.

That use of humans to build the entire index is both Mahalo’s biggest weakness and it’s greatest strength, since while I’ve been fairly disappointed in the breadth of subjects it is knowledgeable on, I’ve been hugely impressed with the quality of the results it does return. It’s hard to describe the joy of entering a search term and having the results come back, free from spam or noise, in a neatly categorised list of high quality links. I don’t know, re-reading that last sentence makes me wonder if I’m overreacting to this, but I can’t help it; it just reminds me of what the internet should be; a place where I can get access to information quickly and easily, without having to know the tricks of avoiding spammers and scammers, and without having to trawl through a mountain of ‘sponsored links’ or uninformed twaddle to get to the actual information. Of course, there’s a place for the uninformed twaddle as well (else I’d have to shut up shop,) but I don’t necessarily want a whole load of conspiracy theories polluting my screen when ‘m trying to research the moon landing.

This leads me to believe that the term “search engine” might be something of a misnomer. Certainly, you can search Mahalo, but it’s hardly it’s strongest suit; constructing a search term for a specific piece of information is rarely satisfactory, and that’s what people expect to be able to do with something called a “search engine”; it’s sort-of implied by the name. Where Mahalo really shines, conversely, is on sufficiently broad searches for well defined subjects (say, ‘Evolution’, or ‘Egypt’,) in those cases it’s hand-built nature and high quality put me more in mind of an index, or a guide than of a blind search algorithm. It’s like going back to Yahoo in the really old days, but with a somewhat-useful search feature, and links you actually want to follow. I guess there’s value in the term “search engine”; people know what it means, but I can’t help wondering if they might help differentiate themselves, as well as communicate their unique strengths, if they used a different term.

So, regardless of what it is, would I recommend Mahalo? Yes, absolutely, but not unreservedly. It’s not complete yet; that much is painfully clear in daily use, and I end up falling through to wikipediaor google at least as often as I get the answer from Mahalo, and there are types of searches that I don’t think it will ever be any good for (but then, I think there’s an argument for specialised search engines for many of those anyway.) For the searches it’s designed to handle, though, I think the value of it is well worth the initial inconvenience of having to contribute to the index myself. Suggesting links is quick and easy, and the more people that use it regularly and add the pages that it didn’t find for them, the sooner it’ll be complete and really useful.

So, go on - check it out. If you use firefox, you can configure it to use Mahalo as the default search engine, if not then it might be a little less convenient, but I think it’ll be worth the effort.

I’ve just been pointed at Mahalo (thanks Phil,) a new search engine. What sets it aside from the mainstream engines, is that it’s maintained by humans, rather than robots. Obviously, that raises concerns about it’s completeness, but it also means that you can trust its results not to be SEO-spam. In my limited testing, I’ve found it very good, with nicely presented, categorised and, most importantly, relevant results. It also provides results from the major robot-backed engines as well, in case they don’t have what you need indexed yet.

I doubt it’s going to unseat google any time soon, but I’m going to trial using it as my search engine of choice for a week or so, and see how it goes. I’ll report back when I’ve actually used it in anger.