|
|
To print: Select File and then Print from your browser's menu
-------------------------------------------------------------- This story was printed from ZDNet Australia. --------------------------------------------------------------
|
Avert your gaze! 8 filtering packages tested September 05, 2003 URL: http://www.zdnet.com.au/reviews/software/internet/soa/Avert-your-gaze-8-filtering-packages-tested/0,139023437,120278273,00.htm
Just how good are web filtering packages? We put eight of the best head to head in our Australian review.Turn on any current affairs program, and if it's a slow day, chances are you'll hit a story decrying the internet as a haven of porn, bomb-making instructions and paedophiles. A certain quantity of this is justified -- and a fair quantity is just the TV networks making ad dollars out of hype -- but it has made people think about the kinds of access that they allow others to have to the Net, from the very top of large enterprises to the family home. We've examined here eight packages aimed primarily at home users/small PC networks, although some are capable of being implemented on larger scales. Most packages in this range focus on the concerned parent market, and within that market products need to be comprehensive in their approach, and at the same time easy to use. Different packages use different filtering approaches, and if you're pondering installing one, it's worthwhile knowing the common technology approaches used. Blacklist/Whitelist In terms of total filtering capability, it isn't possible to get better than a whitelist, which blocks everything but the sites listed within, but assembling it can be a time-consuming activity; as the Internet expands you'd need to constantly adjust your whitelist parameters. If you've got the time for that kind of approach, especially if you're thinking of using a package to protect your children from viewing inappropriate content, you may be as well served simply monitoring your children's surfing habits. The single most contentious issue surrounding blacklists is whether or not the end user has access to the master blacklist. Most packages that rely on a blacklist block access to the master list, essentially because it's what they're selling on to you, and it represents their effort and R&D funding to create in the first place. The problem with a hidden blacklist is that it removes the ability from the end user to decide what's appropriate for the end user to view; unless it's blocked you don't know if the blacklist includes sites that you might approve of, or if it misses sites that should be legitimately blocked. Only one of the packages in our roundup that includes a large blacklist, Net Nanny, allows users to see and edit the master blacklist. Word/Content Filtering The best packages in our roundup use a mix of these approaches, and even then no package passed our tests with flying colours. That's not entirely surprising, of course, given the general rate of change over the entire Net, but it does point to one basic fact; the best way to control the access of others to material is to sit directly behind their shoulders as they surf. Rhinosoft AllegroSurf 5.1AllegroSurf is the Swiss army knife of filtering products.
Webroot ChildSafe
Zecos ChildWebGuardian
Proxymis ControlKids
CyberPatrol 6.1
Solid Oak Software Cybersitter 2003
Bionet Systems Net Nanny 5
We-Web We-Blocker 1.91
How we tested
Editor's Choice
Rhinosoft AllegroSurf 5.1AllegroSurf is the Swiss army knife of filtering products; aside from its filtering functions it offers a variety of other uses, from pop-up blocking, advanced caching facilities and Net connection sharing. Those are all fine product uses, and do help to add to the value of the product, which ultimately came out with disappointing results in our filtering tests. AllegroSurf's install is easy to sit through, and with a downloadable 2.8MB executable shouldn't be too much stress on anyone's system. We were mildly annoyed with the fact that at the end of the installation it touts other RhinoSoft products, but at least that's a screen you can quickly skip out of. AllegroSurf uses black and whitelists for filtering, although when you first install the software these are blank; that gives you a lot of scope to work with, but at the same time also means that you'll have a hefty amount of work in front of you. It also offers a metafiltering option across six categories: Sex, Drugs, Hate Speech, Crime, Extreme, Sports, and that's what we tested with. Without activating a black/whitelist, AllegroSurf was clearly the weakest filtering product in our roundup. It let through three different objectionable pornographic sites, blocked an odd assortment of sites dealing with sexual education matters, and only partially blocked some of the hate sites in our test list. While it doesn't specifically state what its metafilter uses, we'd offer a guess that it's a simple word filter, albeit one that works only sporadically. Strangely enough, for a filtering package of this type, it managed to pass our filtering parody page, something that was beyond most of the other packages in our roundup. Because of its initially blank white/black list, there's no way to submit sites to a central server for review, which means (like products such as Childsafe) that you'll have to build your own filtering list to make AllegroSurf truly worthwhile, at least from a filtering perspective. On the plus side, when its metafilter did filter sites, it did so quickly and efficiently. It uses a popup message to indicate that a site is unsuitable, so users should quickly become aware what areas are allowed and what areas aren't. To be fair to AllegroSurf, content filtering is only part of the overall package on offer, but sadly within at least its filtering capabilities it's remarkably weak.
AllegroSurf 5.1
WebRoot ChildsafeWebRoot Childsafe's offering concentrates primarily on logging user activity, relying on the end user to set up a blacklist of sites. It offers no keyword filtering and little in the way of additional functionality. Given its US$29.95 asking price, it's only an acceptable choice if you've got plenty of time on your hands, and if the popup policy of the free version of We-Blocker truly annoys you. WebRoot hides itself quite well; unlike most of the other filtering packages we've looked at, it offers no system tray icon, instead relying on a keyboard combination to bring up its control panel. This is protected by a master password, but there's no facility for a reminder question, so remembering the password is quite important. It also took us some time to realise that logging/filtering has to be enabled from within the program before it will do anything at all. Childsafe's theory on filtering seems to be that users who know they're being watched will behave themselves, and as such, it concentrates more on logging user activity than on strict filtering parameters. While it does allow users to set up blacklists, when first installed it knows nothing. This placed it firmly outside the scope of our testing; without loading in the URLs of all of our test sites, we'd never be able to block anything. Childsafe does offer some options we'd love to see in other filtering programs, most notably the ability to force users to click-agree to an acceptable usage policy, as well as some very strict logging that captures pretty much every aspect of a user's activity. Activity logs can be periodically emailed out behind the scenes, making it nigh-on impossible for users to delete or modify logs. Ultimately, however, Childsafe does come up short on filtering options; the ability to set user blacklists is present in every other package we've tested, and while its logging facility is quite advanced, it doesn't offer any analysis of user pages, so you'd need to follow a user's surfing path closely in order to build up your blacklist.
WebRoot Childsafe
Zecos ChildWebGuardianInstallation of ChildWebGuardian follows the normal path for filtering programs, and includes a mandatory master password that must be created. Strangely, it presumes that the master user will be masculine, as it notes that "In the future only the man knowing this password will be able to change program settings." ChildWebGuardian's interface is easy to enough to follow, with tabs that cascade down the left hand side of the program window, covering the user log, White and Black lists, Allowed Time, Popup Killer function, Forbidden Words and standard options dialogues. ChildWebGuardian's web site claims that the use of black and white lists isn't a terribly effective way to block content, and while that's not a wholly baseless statement, we did find it odd to note that the program contains options for creating both black and white lists. The supplied blacklist is always visible, but it contains only four entries, so if you wanted to build a custom-allowable list, you'd have a fair amount of work to do. Without the use of a custom blacklist, ChildWebGuardian falls back on what it calls 'Content Analysis', and what we'd call word filtering. The list of words is easily available, but it's also quite short in every category but Sexually Explicit terms. Its Sects/Cult section, for example, only has six terms, and two of those are 'sect' and 'cult'. Another, and we're not entirely sure why, is 'Jehovah'. In general word filtering in and by itself is a highly ineffective way to filter sites; there's too much of a possibility of false positives (sites that shouldn't be blocked being blocked) and no real method to block offensive or inappropriate images. If you're using a browser that isn't Internet Explorer, then ChildWebGuardian isn't for you, although to its credit it does note this fact on its web site product pages. Another weakness in ChildWebGuardian's approach is that it doesn't have a tolerance setting for words, merely a word counter. If the blocked word appears more times than you've selected, it'll be blocked, seemingly regardless of origin or context. With the overall weakness of a word filtering mechanism in mind, we weren't expecting great things from ChildWebGuardian. The first thing that we noticed when we began testing is that the content filter isn't terribly fast, and often potentially objectionable material would begin to load before the filter kicked in. With its larger library of potentially offensive sexual terms (including some obscure foreign language terms that could be handy to have blocked depending on where you're surfing) it performed best on pornographic sites. ChildWebGuardian, along with most of the other filters in our tests, passed a couple of notably disgusting and offensive image archive sites with nary a whisper. In every other area of our testing, however, ChildWebGuardian showed the weakness of its word filtering model; with word filtering enabled it blocked the AIDS Quilt site, an anti-hatecrime site and a filtering parody site that contained no genuinely offensive material. At the same time, it had the common problem of loading part of the KKK site, and, as noted above, would often load material before blocking it, leading to partial exposure of material. To see how far the word filtering would go, we did a google search for "Essex", only to find the results blocked. Ultimately, while ChildWebGuardian's interface is quite easy to get around, it does very little to justify its US$24.95 asking price. Its word filter was overzealous in our testing, which might sound appealing until you realise that you'd spend an awful lot of time filtering out false positives, negating the value of having a filter in the first place.
Zecos ChildWebGuardian
ProxyMIS ControlKids 5.02If the name itself didn't clue you in, the cloyingly sweet picture of a small girl with her thumb held triumphant on ControlKids' Web site should make it entirely clear that this is a package aimed squarely at the concerned parent market. Like many of the packages we've tested, ControlKids offers an evaluation period with its package. ControlKids, however, has a unique twist; its evaluation period is a random number of days, between 7 and 14; we're not entirely sure what useful purpose this serves. ControlKids boasts a clean and easy to use interface, complete with cutesy icons for each of the program's function areas -- Forbidden Sites, Popups, Confidential Information, Clear Activity, Downloading, Keyboard Capture, A help file and the program's main menu. In an odd piece of GUI design, the same tabs that appear on the left hand side of the interface are duplicated exactly across the top of it as well. ControlKids does boast some features that no other package in our roundup could match, most notably the ability to keylog every character entered on the host PC. This is a common technique used by hackers, and it's a little unusual to see it in a commercial product. It does open up the possibility for small businesses to use it to log employee activity, although how appreciative said employees would be to have an application called 'ControlKids' running on their desktops is hard to say. ControlKids' main interface can be locked with the usual password method, although in the version we tested, by default it isn't, which is unusual. Like ChildWebGuardian, although it offers a site black/whitelist facility, it works primarily by filtering content based across five categories -- Sexually explicit, Violence/Intolerance/Weapons, Sects/Cults, Casino Games and Hacking/Warez/Cracks. There's no capability to view the filtering mechanism, which in our testing seemed to be word-based, or indeed to add any modifiers or additional terms to it. You can add or remove sites to the white and black lists, which work in concert with the inbuilt content filter, although by default these start off blank when the software is first installed. On the subject of all things blank, ControlKids does log internet activity, but offers no clues as to whether a site was passed or blocked, leaving that area of research up to the end-user. ControlKids is also lacking in facilities to set up multiple users, so whatever settings you enter are those that will have to be used by every user of that machine. While there are some consumers who won't need a multi-user function, it would make a handy addition. In our testing, ControlKids' results were middling to poor. The one area where it did shine in relation to other packages was its pop-up killer, which worked exactly as advertised. Sadly, that's pretty much outside the scope of this review, and simpler pop-up killers are available for free. When ControlKids finds an objectionable site, it pops up a standard graphic telling the user that the site is blocked, which is better than showing nothing, but combined with the lacklustre logs, end users will have to do a lot of digging around to ascertain if sites are in fact objectionable or not. ControlKids blocked most of our pornographic sites, although it did allow us into one particularly objectionable archive of deliberately disgusting images. Its filter proved sketchy for sites containing potentially sexual terms, as it blocked the Sydney Mardi Gras website and the AIDS quilt. Likewise with hate sites, it passed some white power sites while blocking some pro-tolerance ones, and in our general testing it blocked our filtering parody page which contained no genuinely objectionable content. At US$35, ControlKids could offer an awful lot more than simple and often incorrect content filtering, something that a package like We-Blocker does for free. As it doesn't, there's really no reason to go with this particular filtering package, unless the cutesy graphics appeal to you for some strange reason.
Proxymis ControlKids
CyberPatrol 6.1CyberPatrol uses a web-based install routine, something that may deter potential purchasers the first time a security warning pops up. To its credit, CyberPatrol's installation page walks timid users through step by step, including one of the largest downloadable installers we've seen, clocking in at 10.3MB. Once installed, the software prompts for a master password, which it refers to as the 'headquarters' password -- same thing, different term. It backs this up with a password hint; we're not sure about the overall security of this approach. Rather than having a question with a correct answer, the hint will need to relate to the real password, an approach that could inadvertently lead to unauthorised users working out the password and rendering the point of having a filtering solution moot. Once installed, CyberPatrol sits in the taskbar, and uses an interesting dual windowed approach. Any user can check current filtering status and change user via a very small window, while only those with the headquarters password can access CyberPatrol's headquarters where program parameters are set. Like every other package in our testing, CyberPatrol supports user-defined white and blacklists, as well as keyword black and whitelists. We're not quite sure what the point of a keyword whitelist is, unless you want extremely fine-grained control of a user's surfing habits. Cyberpatrol's categories list covers a wide range of topics: Adult/Sexually Explicit, Chat, Criminal Skills, Drugs, Alcohol & Tobacco, Gambling, Glamour & Intimate Apparel, Hacking, Hate Speech, Remote Proxies, Sex Education, Violence, Weapons and Multiple Category Servers. You can toggle filtering for each of these categories, but CyberPatrol offers no facility to view its master list of blocked sites. CyberPatrol also offers additional filtering options, including newsgroup and chat filtering, program execution limiting (by user name or time) and user time management. When CyberPatrol detects a site that matches its criteria, it instead loads an explanatory page detailing why the site's been blocked. This is an approach we feel best suits the needs of both system administrators and users; it allows admins to quickly categorise sites, and allows users to see why they've been blocked and potentially allow them to argue for a site's whitelisting, which somewhat (but not entirely) alleviates the problem of not being able to see CyberPatrol's master list. In our testing, CyberPatrol managed to move beyond the 'block sex sites' ethic that many of the other filters in our roundup seemed pre-occupied with, managing to filter out a reasonable quantity of material covering other topics. Given that it offers an explanation of its block, we were sometimes left wondering why it picked certain categories. We tested with CyberPatrol's default setup, and based on our testing we'd advise any potential users to carefully select categories, as CyberPatrol tends towards the conservative when choosing whether to block sites or not. While it fell into the common trap of loading one particularly offensive images site, it did manage to pass our inoffensive filtering parody page, something that many other packages fell over on. At US$39, CyberPatrol's one of the costlier packages in our roundup, and this is further exacerbated by the fact that the package includes a subscription component; your US$39 buys you one year's subscription. Let your subscription lapse, and you can opt to block everything, or nothing, but not to simply continue to use the existing blacklist.
CyberPatrol
Solid Oak Software CyberSitter 2003The subject of access to master lists is a touchy one for filtering companies, who see it as their primary R&D cost; as such they'd rather not have them visible to the public. CyberSitter perhaps takes this to an extreme; it's the only software package we've ever seen to have this rather restrictive clause in the End User Licence Agreement: "Unauthorized reverse engineering of the Software, whether for educational, fair use, or other reason is expressly forbidden." It's only available as a downloadable file, and has one notable quirk that no other package contained; it won't install without first being given the primary email address of the main system administrator. We're not sure why CyberSitter wants this information, but in an age where Spam is endemic, we suggest it's a factor that may turn many users away from the product. When CyberSitter 2003 is first launched, it gives the option to scan the hard drive for what it calls "objectionable material". It calls itself the CyberSitter System Snooper, which we'd challenge anyone to say three times fast. Its interface is easy enough to follow, but visually quite basic. CyberSitter works via a mixture of word filtering and category lists, and it's the broadest of any package we've examined. By default, it will filter for material under the following categories: Adult/Sexually Oriented, Illegal Activities/Drugs, Hate/Intolerance, Illegal Guns/Violence. Optionally, users can add the following: Gay/Lesbian Topics, Cults/Occult, Violent Games, P2P File Share, Tobacco/Alcohol, Gambling Sites, Banner Ads, Legal Guns/Weapons, Personal Ads, Tattoo/Piercing, Warez/Hacker Sites, On-Line Chat, Shareware Sites, Financial Sites, Illegal MP3 Files, Popup Ad Windows, Sports, Game Sites, On-Line Auctions, TV/Entertainment, Movie Sites, Wrestling, Job Search, Free E-Mail Sites, Pokemon Sites, Astrology/Fortune Telling, PICS Ratings Adult/Violence. Cybersitter can also block file sharing, FTP, Newsgroups and Instant Messaging clients. Users can be forced to use Cybersitter's default search engine, although strangely enough not on systems running Windows 2000 or XP. In addition to the categories list, CyberSitter also offers a word filter and whitelist option. While the word filter initially appears to be blank, it's clearly always in action; Cybersitter's own page notes that "A sophisticated "content recognition" system recognizes and blocks new objectionable web sites even before we know about them", which sounds like a marketer's way to describe a keyword filter if ever we heard one. Those of an anti-filtering package persuasion could reasonably argue that implementing all of these options could block out the entire Web. In our testing, CyberSitter sat in the upper echelon of packages, but we had strong reservations about its blocking method; rather than show a warning page, it simply loads a blank page, which gives the end user no information at all. As a result, users won't know directly if a page is forbidden, and may try again. On some test pages, it instead defaulted to showing a DNS error, which is even worse, as users will just suspect the page is down. CyberSitter's logs do show information to the administrator as to why pages are blocked, and for the most part these were informative, although we did find some hate sites classified as 'adult' for some reason. We also hit problems with certain site types, as in our testing it didn't seem to matter whether the filter was on or off. The Gay/Lesbian filter picked up pages -- including one sporting association -- whether it was activated or not, and the gambling filter gleefully let through our selection of sites regardless of it being on or not. It stumbled on our filtering parody page, which set off six different word flags.
CyberSitter 2003
Net Nanny 5Net Nanny has a fairly prominent name in the web filtering business, and it was the only package in our testing to come in an actual package; while there is a downloadable version it's also distributed locally in a box by Manaccom. Installation of Net Nanny is slightly different to many of the packages we've examined here. For a start, the software requires a reboot before it'll work properly. Once that's done you're dropped into an initial configuration screen where you can choose to create users, or just go with a generic 'anybody' account with basic blocking in place. Given that the software itself recommends giving the 'anybody' account a trial for a few days before setting up user accounts, we suspect most users will just set it that way and leave it, so that's the configuration we used in our testing. Net Nanny utilises a password to lock down the settings console, but it has a unique approach to protecting that password. Where most of the other packages we've examined use a hint question, Net Nanny requires the administrator to create two questions and answers before the settings console will open up, an extra level of security that's quite welcome in a product of this sort. Once initial configuration is complete, the package checks with Net Nanny's servers for updated configuration files and is then ready to run. While running, Net Nanny sits like so many other programs in the system tray, although it won't respond to a regular left click at all; a right click brings up the options to display the current user, change user, deactivate the software or change Net Nanny's settings. The settings display panel isn't the most attractive, or for that matter the easiest to get around, but it's quite comprehensive, and once you get used to the fact that most dialogs rotate around to different areas within the settings panel, easy to administrate. Net Nanny uses a mixture of word filtering and white/black lists to block content, with a specific focus on 'adult' (read: sexual) content. Its rather unique selling point is that it doesn't hide or encrypt its lists, which are freely available to the administrator for the purposes of review. That's a bold step -- most filtering software manufacturers tend to protect their lists on the basis of research -- although it should be noted that Net Nanny's particular focus on sexual matters means that the list of blocked sites is comparatively small. The administrator can pick whether to simply block sites or also pop up a warning notice to users when banned material is requested by the browser. Disappointingly, it offers no advice to the user as to why a site's been blocked. Net Nanny also offers a number of other functions that fall outside of our particular testing spectrum; users can block the display of objectionable words -- again, with a sexual theme being prevalent --, block users from using IRC and similar chat tools as well as file sharing applications. Net Nanny was one of the few products in our roundup that looked beyond Internet Explorer in terms of web filtering. According to the packaging it should block any version of IE above 5.01 and any Netscape product above 4.08. In our tests, it also correctly picked and detected content being loaded into Opera 7.1 and Mozilla Firebird 0.61. It did have a noticeable hit on performance, especially when checking sites with complex layouts, where presumably the Net Nanny engine has to work that much harder. In our tests Net Nanny correctly blocked all of our clearly pornographic websites, and many of our questionable adult sites. It was one of only two packages (along with Cybersitter) to correctly block two of the web's more infamous disgusting image sites. We were keen to see how well a blank installation of Net Nanny would handle sites outside its core competency of 'adult' material, and this set of tests showcased a definite weakness, as it passed all of our hate, weapons and drug sites with no blocking whatsoever, although its word filter did overwrite some objectionable terms within those sites. That's not an insurmountable problem -- as with other packages that allow user white and blacklists you could always build your own lists of allowed/disallowed sites based on the user logs. Net Nanny's logs are quite easy to read and export quickly into .txt format files. Net Nanny is a very effective pornography blocker that's noticeably weak in other areas, although the ease of adding sites does mean that it could be built into an all-round filtering tool, as long as you've got the time to administer it properly.
Net Nanny 5
We-Blocker 1.91We-Blocker might have an unfortunate and somewhat urological-sounding name, but it does score brownie points for one feature it has over all the rest of the competition; it's free to install and use. As such, it poses a strong challenge to the rest of the software pack, as there's little reason to put down money when a free alternative is available. Installation of We-Blocker couldn't be much simpler; the downloadable application clocks in at a little under 2.5MB, and installation follows the usual Windows path. Once installed, the software requests that you create a master password and password secret question, with the usual warnings about configuration and uninstallation problems if you forget both. We-Blocker uses a mixture of blacklists and word filtering to work out whether or not to block pages. The word filter is quite robust, as words are assigned a severity, and it's possible to set one of three different severity levels before a page can be blocked. The software does warn you that there's likely to be a performance hit for having word filtering switched on. Its blacklist categories cover Pornography, Adult Subjects, Violence/Criminal Activity, Hate/Persecution, Drugs/Alcohol, Gambling, Weaponry and a single 'User Defined Category' field, although we were unable to locate anywhere in the software where this user defined category could be created. It's also got user-defined white and blacklists for fine-tuning applications. We-Blocker doesn't go as far in revealing its blacklists as Net Nanny, but it does offer a semi-compromise; you can search on the We-Blocker site for individual URLs to ascertain if they're in the blocked sites list. In operation, we did hit a few snags. Firstly, We-Blocker only works with Internet Explorer; switching over to alternate browsers such as Opera or Mozilla Firebird allowed us free and easy access to whatever content we wanted. Blocked sites are accompanied by an audio file that says 'blocked' for every infringement it discovers, which can become a touch cacophonous. By default, We-Blocker will head for its own site rather than give a full explanation of why a site was blocked. It points to the master user's logs for reference, but we often found those logs to give little indication either. Remember how we mentioned that We-Blocker is free? Well, it is and it isn't. You can use it for as long as you like, as long as you're happy with intermittent pop-up ads requesting that you donate towards its further development. Donate, and the package disables the pop-up self-advertising; your tolerance for this 'feature' may vary, but we're not keen on it in supposedly 'free' software. We-Blocker's filtering technology, like many of the filters we've reviewed here, performed best with sites containing pornographic material. Like most of the rest of our filters, it failed to block one particularly disturbing image site, but otherwise picked matters sexual with a fair degree of accuracy. It initially blocked access to the AIDS quilt information site, which may worry some users; turning the word filter off was enough to gain us access to the site. In other areas that it purportedly blocks, We-Blocker was much less effective, especially if the word filtering aspect of the software is disabled. Like most other filters, it had problems with the use of frames in the Klu Klux Klan's web site, and with the word filter off it completely failed all of our hate site tests. It fared slightly better for gambling sites, again performing best when it could fall back onto its word filter. We-Blocker also demonstrated beautifully the problem with needing to fall back on a word filter, as a site we tested that contained many filtering 'bad words' (but in reality is harmless) was blocked every time by the package, but passed once word filtering was switched off. Ultimately, We-Blocker scores well with us for configuration flexibility and price, but is somewhat hobbled by the pop-up ad support, IE-only status and over-reliance on word filtering to keep itself safe. Like most filters that allow for user-defined black and white lists, it would be possible to build up We-Blocker into a very effective tool, given enough time.
We-Blocker 1.91
Web Filtering: How we tested
We tested web filtering packages against three different criteria; how well they stood up to manufacturer claims, how easy they were to use, and ultimately how effective they were at blocking out content. In order to test the effectiveness and claims of each product, they were tested against a master list of sites likely to trigger filtering software. This is an imprecise science, but then so is the whole filtering software approach, as what each individual thinks of as objectionable varies so widely. All of our test sites were selected via very simple Google searches in order to make them as accessible as possible to both end-users and filtering software companies. Our list was broken across eleven categories: Pornographic Sites Possibly Pornographic Sites Sexual Matters Sites Bomb Recipes Sites of a non-right wing nature Hate sites Religious/Anti-Religious Sites Gambling sites ZDNet's own sites 'Naughty' content sites Filtering critical sites
Editor's Choice: Web Filtering
Editor's Choice: None Highly Commended: Net Nanny, CyberPatrol, We-Blocker This was a highly contentious issue, but in the end we feel we can't offer the Editor's Choice award to any package, because in our opinion no package from our test group performed sufficiently well enough. There will always be an issue with the growth of the Net and keeping filtering products up to scratch, but even within our group of test sample pages, each package had enough basic weaknesses to disqualify it from Editor's Choice contention. We suspect it will take someone hitting upon a new filtering method to revitalise the industry and bring some true effectiveness, in much the same way as Google changed the search engine world with the use of its pagerank technology. So, what's the end consumer to do? Our highly commended packages got the nod for individual approaches that we'd love to see in one single package. Net Nanny's the only large-scale package to offer up its blacklist for all to peruse, something we think should be mandatory in any package. CyberPatrol's filtering was mostly effective, and it offered sensible reasons for why pages were blocked to end users, which should make communication between users and administrators that much easier. We-Blocker challenges the notion that filtering products need to be expensive packages, although for a package that relies on community support, we can't fathom why its lists aren't yet open. We should note that while we like the individual approaches of each package, it would be a very bad idea indeed to try running all three at once. The best approach at the home/SOHO level, from the experience in our testing, would be to watch your users very carefully. One of the packages we've tested can be a useful adjunct to that monitoring, but none of them truly match up to what you'd expect from a software product in any other category.
Copyright © 2009 CBS Interactive, a CBS Company. All Rights Reserved. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
