There have been several additions to the world of search engines over the summer, and I thought I’d do a quick round-up of them to see how they perform.
ZapMeta  is a multi-/meta-search engine that has been around since 2002, but I must confess that it has only just come to my notice. The home page is very spartan and reminds me of the way in which Google used to look when it first came out. There is a search box, an option to search the Web or shopping, links to preferences and advanced searches and that’s just about it.
Let’s start by taking a look at the preferences to see what we can do with this creature. You can set a timeout of anything between 1-20 seconds, which is very reasonable - many multi-search engines don’t give you that option at all, or you’re limited to a smaller set of options. Results can be grouped, or they can be displayed - this means that results will remove duplicates by grouping pages with the same URL under one match. The number of results per page can be a number of options from 8 to All, which is nicely flexible. Results can be sorted by relevance, which you’d expect, but also by title and source, and either in ascending or descending order. I like this - it’s rather unusual, I can see ways in which this could allow a searcher to jump much more quickly to particular pages or sites. Keywords can be highlighted or not; quite why someone wouldn’t want keywords highlighted is a bit of a mystery to me, but if you’re one of those folk, it’s an option for you. Pages can be opened in the same, or a new window; a useful feature, but nothing startling. You can choose to display description, sources, URL and page preview. This again is a nice feature, giving you more flexibility than you’ll find with many other engines.
Finally we have a choice of nine search engines ZapMeta will use to search from. Oddly enough, not all of these are checked. Of the nine AltaVista, Open Directory and GigaBlast are unchecked while Yahoo, AlltheWeb, Wisenut, AOL MSN and HotBot are. I do find this a little strange since the strength of a multi-/meta-search engine is surely to search as broadly as possible as a default, with an option of narrowing down to specific engines afterwards? Of course, you can easily set the default to search all the available choices, but I suspect that a lot of people won’t think of doing this at once. The other major point to make here is that Google is obviously missing from the list, but since AOL is included (and is powered by Google) it’s not quite the omission that one might at first assume. One slight annoyance is that once you’ve set your preferences it’s not immediately clear how you can get back to the search screen, (by clicking on the ZapMeta logo), which may cause a degree of confusion or frustration.
Advanced Search Function
The Advanced Web Search functions are not terribly exciting, and they’re what you’d expect to find. We have the usual boxes for ‘Must contain the words’, ‘Match the exact phrase’ and ‘Exclude the words’ with the options of ‘Anywhere’ ‘In the title’ ‘In the URL’ which are fine - a couple of years ago this would have been worth particular mention, but this is fairly pedestrian now. There’s a Domain filter, with options to limit by any one of 8 regions (it’s a shame there isn’t an option to choose a couple of regions though), by Domain (.com, co.uk, org and so on) or by Host, (phib.com or dell.com). Nice, but nothing fancy. The next option allows users to sort by relevance, URL, Popularity, Title and Source. This is an interesting option, and one that I’ll come back to later when I’m talking about the results that we get from the search engine. The final options are again uninspiring - results per page, display options, timeout and sources used. So, nothing special - it would have been nice to have seen search criteria based on media, news, images and so on.
Moving onto the search function, I did a search on ‘Everton’ and got several sponsored listings before the ‘proper’ results. These listings weren’t really appropriate at all, though to the right was a small box of ‘related searches’ which did include ‘Everton football’ and ‘Everton fc’, so I’ll give it marks for getting that right - or rather, picking up those options from AlltheWeb. I had the main body of results ranked by order of relevance, and ZapMeta used the standard approach of ranking by number of search engines and position in the top ten. A nice function is the ‘Quick View’ of the page, which an increasing number of search engines are offering, providing you with an opportunity to view the page without having to actually visit it. The last section of the page offered me some Related categories, which mainly seemed to have been culled from Dmoz.
I re-ranked the search results by clicking on the ‘Popularity’ option, and ZapMeta managed this very quickly, re-ranking the results to give me the official club Web site as top of the listing. Quite what it based that on I don’t know - a failing of the site is that it’s not actually that forthcoming about such things, but as long as it works, I suppose that’s the main thing. Re-ranking by title was less useful, since I then had 1UpTravel.com as first in the list, which is logical, but of little interest to me. The Source option again put the official site back at the top of the listing, while the Domain option once again gave me 1UpTravel. Several other searches confirmed what I was already thinking - and by no means alone - relevance and popularity are probably the most useful options, followed by title. However, it’s a refreshing change to be given the option! Another useful option was ‘search within these results’ which, (although I seldom use it myself), was good to have available.
The layout of the results page was quite clear, although I did have one minor gripe, related to moving to the next page of results. The options given are:
1 2 3 4 5 6 7 8 9 10 next > next 10
Clicking on ‘next 10’ did not take me to results 20-29, but to 101-110. Clearly, the ‘next 10’ option refers to pages 11 12 13 and so on. Logical when you think about it, but until you do, I can guarantee a quick double-take when you look at the results!
I also ran some comparative searches with one of the other multi search engines that I use a lot - ez2www . On all eight searches ZapMeta performed better than ez2www, and although it’s obviously a subjective viewpoint, the results that I was getting seemed rather better. Theoretically all the results should have been the same, since the same engines were being used for the same amount of time. It’s possible that there were problems with timeouts, since I did all the ez2www searches first, and then all the Zapmeta searches next, but for all the results to come out in favour of one search engine does tend to imply that it is doing its job rather better. I would however encourage anyone to run some searches themselves - it’s more than possible that you’ll get a totally different set of results.
My conclusion is fairly short and to the point - ZapMeta has just taken over as my favourite multi-/meta-search engine. I’m happy with the features, the display and the results that I’m getting. I like the options for re-ranking and it just seems to be that bit sharper than the other engines that I use in this category. That of course doesn’t mean that I won’t continue to use them - far from it - but I think they’re going to have to do some upgrading pretty quickly in order to compete.
The first thing to be aware of with the Netnose search engine  is that it’s offering users the opportunity of ranking or rating sites. I’m not particularly convinced that this option works terribly well, and even after having long discussions with their staff I’m still questioning this. However, let’s look at it as a search engine first, and deal with the ranking element later.
Netnose uses a non-tandard syntax; there isn’t an option to search for a phrase; to search for two or more terms it’s necessary to use and: term1 term2 . A phrase search for Phil Bradley is therefore and: Phil Bradley (which of course isn’t a ‘real’ phrase search since both terms could be found at different places on the page). The options to include or exclude are however the same - a plus/minus sign. There is an interesting option which has not yet been fully implemented (and in fairness I should say that Netnose is only in Beta testing at the moment). This allows searchers to limit to particular types of site, such as commercial or adult. Consequently, to search just shopping sites the syntax would be buy: car stereo Seven more categories are planned for the future, including Kid safe, Research and Entertainment. Netnose appears to work with a fairly small database - searching on ‘internet’ only returned 26,000 references in comparison to over 188,000,000 that AlltheWeb returned.
The displayed results allow the user to interact however, which is the interesting aspect of this search engine - it’s possible to vote for a category to place a particular site into, or to indicate to Netnose that the site failed to load or doesn’t belong in the category assigned to it. Unfortunately this last option doesn’t have an ‘Are you sure?’ option, so my apologies to the site that I maligned when I was trying it out! Of course in order to do this properly it’s going to be necessary to visit the site first and probably spend some time there in order to do the site justice. Netnose are of the opinion that their users will do this, but perhaps I’m just being cynical, but I really don’t see it happening myself.
Netnose is worth looking at if for no other reason than they’re trying something new. As well as the option of voting for sites that are returned from your search, there is also an option ‘Rate the Web’. People who volunteer to take part are given a random Web page and some words to describe that page. They then match the terms to the content from ‘Wrong’ to ‘Best’. That is of course a very simplistic overview, and Netnose does explain it in greater detail on their site . Netnose isn’t a search engine that I’ll be using myself, (not yet at least), but it is one that’s going to be worth keeping an eye on in the future.
According to their Web site, Wotbox  has been in development for almost a year, and it’s still in alpha testing, so it only has a database of some six million pages, though they hope to reach 100 million pages by the end of the year. Although it’s still only in test, it has a reasonable array of functions - in fact, more than some engines of longer standing. All the usual search options are there - phrase, all words, any words, exclude, a country filter and option to search a specific site (although these last two use a non-standard syntax which may cause confusion). Interestingly, however, Wotbox also allows the use of the wildcard to replace either a single character or multiple characters, and it’s refreshing to see a new engine embracing these options at the outset. Another function offered by Wotbox is ‘related searches’ which appear with most sets of search results. Unsurprisingly these are links to searches that are related to the one that has just been completed.
Wotbox is another search engine that I won’t be making any great use of myself at the moment, though again, that’s not a criticism, just based on practicality. I will however be keeping an eye on its development over the coming months to see how it gets on. In case readers are confused, until recently Wotbox was called Wotbot, but the owners came under pressure from Lycos and Hotbot to change the name, which they did in mid-September 2003.
Stop Press!: Almost as this column goes live I’ve been informed of another new search engine - this is called Mooter  and it’s a little bit like Kartoo , in that it’s a graphical search engine. I’ve not had a chance to play with it yet, but if you have five minutes to spare, take a quick look!
- zapmeta http://www.zapmeta.com
- ez2find http://ez2www.com
- NetNose http://www.netnose.com
- Netnose Search Technology http://www.netnose.com/nnst.htm
- Wotbox http://www.wotbot.com
- Mooter http://mooter.com:8080/moot
- Kartoo http://www.kartoo.com
5 Walton Gardens
Phil Bradley is an Independent Internet Consultant.
Article Title: “Would All New Search Engines Take One Step Forward!”
Author: Phil Bradley
Publication Date: 30-October-2003
Publication: Ariadne Issue 37
Originating URL: http://www.ariadne.ac.uk/issue37/search-engines/