Tuesday, January 11, 2005

Desktop search comparison

With Yahoo! releasing their desktop search tool, I thought it was time to take another good look at all the major desktop search tools. Here is my comparison:

Introduction -

Effective and easy access to information on a computer is a basic need. For long, this hasn't been an important issue due to the limited amount of information that could be stored on personal computers. But, with the ballooning size of hard disks and widespread use of digital devices that are generating tons of content, the demand for efficient methods to retrieve and search for this data is slowly growing and becoming more popular. This demand has grown big enough to attract the attention of the big players in the industry. As a result, what we now have is a glut of desktop search applications in the Windows environment, with many more expected, which are spoiling consumers for choice and making the entire experience of using a personal computer less frustrating. In this comparison I'll examine the different desktop search applications and recommend the one that provides the best overall experience.

To begin with, what does 'desktop search' mean? It's a little like having a search engine exclusively for your desktop. Just like searching the internet for information, desktop search allows searching of hard drives, at similar speeds. Yes, that's the interesting part. The first time a desktop search application is installed, it indexes the contents of your hard drive. Indexing is the process of reading through all supported file formats and storing that information in a way that can be quickly accessed when queried. This leads to extremely fast results, like in an online search engine. Indexing of the entire hard drive is a one time process, and as files are added, deleted or modified on the hard drive, these indexes are also updated. The advantage of this is that, one no longer needs to precisely remember the location of the thousands of files on a hard drive. Only a general idea about the file is sufficient. But, the index takes up some hard drive space. The space required depends on the amount of data present on the drive. But, this is usually not more than a few hundred megabytes in most cases and the resulting convenience is well worth the lost space. Compared to the soporific searches consumers had to endure at the hands of Windows' default search, this technology is a godsend. Actually, Windows XP has an indexing service which aims to do something similar, but, is badly implemented and is better left unused.

This comparison includes desktop search applications from Google, Copernic, Microsoft, Yahoo, blinkx and Filehand. This comparison is based on user interface, features, file format support, search results, resource utilization when running etc. Since Yahoo desktop search uses X1 technology, they are similar in many aspects and differ only in some features.


Installation and indexing -

While Google, Copernic and Filehand search were pretty small downloads, Yahoo and blinkx were larger than 6 MB. MSN desktop search is a small download by itself, by PDF file support requires an additional download from Adobe which is around 10 MB. The applications from Copernic and blinkx support Windows 98 onwards, but GDS, YDS, MSN desktop search and Filehand search only support Windows 2000 and XP. While the Google, Yahoo, MSN, Copernic and blinkx applications were free downloads, Filehand search is free but requires the user to register the product with a free registration key provided by them.

The installation of all the applications was pretty straightforward and there were no surprises.

After installation, each of them started their initial indexing. MSN desktop search, Filehand search and blinkx do not, by default index the entire hard drive. Changes have to be made in the application settings to ensure that the entire hard drive is indexed.

The time required for the initial indexing becomes important if the hard drive contains a very large amount of data. YDS took the least amount of time to index my hard drive and was extremely fast. This is impressive considering that YDS supports many more file formats than the other applications. More on file formats later. Copernic, Google, MSN and blinkx desktop searches took about the same time, but a little longer than YDS. Filehand search, though, took the longest; almost four times as long as YDS. The exact time taken by each application is not relevant as it depends on various factors such as the amount of data, the rotational speed of the hard disk, the speed of the processor etc. One gripe I have about MSN desktop search is that even though it indicated that indexing was complete, all I had to do was leave my computer idle for a few seconds, and it would pounce in and start indexing again. I don't know if these were efforts at incremental indexing or initial indexing itself.

If you are a user who keeps the computer on only when you are working, then you might want to leave it on for some time to let the desktop search application finish its indexing, as they index only when the computer isn't used for some time. This is to ensure that the system resources hungry process of indexing does not interfere with your normal work. Filehand search and YDS are exceptions to this. They index in fixed intervals, which can be defined by the user, or between a user specified period of time during the day. blinkx gives the user total control over the indexing, which is not the case with the other applications.

If the user tries to perform a search when the indexing is incomplete or in progress, then, the results are not comprehensive and do not include all the content on the hard drive.


User Interface -


Google desktop search -

Google Desktop Search home page

Google desktop search (GDS) is the only application to use a browser as the interface with the user. All the other applications have their own dedicated applications. The interface is very similar to the Google web search interface. Using a browser as the interface has both pros and cons. The advantage is that Google can easily integrate search results from the web and the desktop. The disadvantage is that Google cannot leverage the inherent power of desktop applications, such as a preview of search results, instant results as the query is being typed etc.

Web results page

The above image shows how GDS integrates search results from the web and desktop.


Copernic desktop search -

Email Search

Copernic desktop search (CDS) has by far the best interface of the lot. It has a familiar e-mail client look to it and is very convenient and easy to use. The user can select what kind of file he wants to search for in the toolbar at the top. Searches can be refined to narrow down the results. The results can be instantly previewed in the 'Preview' pane. This is extremely convenient and saves the trouble of actually opening each result in its native application. Details of the selected result are also displayed.


blinkx -

Blinkx

blinkx has a well designed user interface, but sorely lacks a preview pane. As a saving grace, it does show a snippet of the document in a pop up box when the mouse hovers over it in the results list box, but, only the first few lines of the document and not the portion where the search term is.

blinkx provides a separate button for each of the different file types that it supports, so, searching among a particular file type takes one lesser step.


Filehand search -

The Filehand search UI is very simple, without any bells and whistles. All the important controls fall easily to hand. Advanced search options are available on clicking a button. Though there is no preview pane, Filehand search uses the web search engine trick of displaying a portion of the document where the search term is present.


MSN desktop search -

In typical Microsoft fashion, MSN desktop search is all eye candy and looks great. It consists of different toolbars for the desktop, IE and Outlook. In the desktop toolbar, a small window pops up providing instant feedback as a query is typed. Searches from Outlook and IE open in an application window, but, there is no preview pane. Only a small snippet of relevant text is displayed below the title/subject.


Yahoo desktop search -

YDS always displays all the files in the index. The search results are narrowed down as the user types the query. It has separate tabs to search for e-mail messages and also attachments that none of the applications have.

The UI of YDS is based on the version 5.0 beta of X1 and it is a huge improvement over X1's earlier version.


File format support -

The utility of any desktop search application lies in the number of different file formats that it can support. Obviously, the more the better.

All the applications support text, HTML, MS Word, Excel and Powerpoint, MS Outlook/Outlook Express and the popular music formats. Only Filehand search does not support image and video files. Corel Wordperfect files are also supported by YDS, CDS and Filehand search.

Only GDS does not support the all important PDF format. Considering the popularity of PDF and the number of documents the average user has, this is a glaring miss from Google. GDS redeems itself by being the only one to support searching of chat logs, but, supports only AOL instant messenger.

CDS is the only application in the comparison to support Firefox. With the rising popularity of Firefox, this will definitely be a plus for its users.

The winner in this department is YDS. It supports a whopping 225+ file formats which is impossible to list here. The only applications it does not support are chat tools and Firefox.


Features -

Here I'll highlight only the standout features of each application, the basic search functionality being the same in every application.


Google desktop search -

Google Desktop Search results page

The results in GDS are listed in the familiar Google format. The number of search results for each file format is displayed on the top and the results can be narrowed down with a single click.

As mentioned earlier, Google's decision to stick with the browser pays off here, with a seamless integration of web and desktop searches. If GDS is installed, searching on the Google website will return results which include data on the desktop.

GDS also allows the display of search results sorted according to date or relevance as determined by the Google algorithms.

Supported file formats need to be opened in their native applications, but, chat logs, e-mail conversations and cached web pages can be opened in the browser itself. Also, GDS does not allow searching through attachments in mail.

By using the browser, GDS can't provide the instant feedback experience, but, using the recently unveiled 'Suggest' feature, they may make up for it by suggesting keywords from the index as the user types them.

The stand out feature of GDS is its ability to index the IE browser cache and history. Whether it is a security risk or not is for the user to decide. On single user computers this can be a very useful feature.


Copernic desktop search -

Deskbar

CDS puts a search bar in the Windows task bar which can be used to quickly access the application. This is pretty handy, but at the cost of some system resources. The preview pane is very useful, but, is not available for the multimedia content. It also has a few problems displaying Powerpoint files. Also, although CDS displays that portion of the file where the search term first occurs, subsequent occurrences will have to searched for manually. This is very inconvenient when documents have hundreds of pages. On contacting Copernic Support for a solution to this problem, the impressively swift reply was that the feature was still not implemented and it could be accomplished by using the Ctrl + F key combination.

Another nifty little feature of CDS is that every time it's indexing the drive the system tray icon changes to reflect it. While it doesn't seem to be very important, its value becomes apparent during day to day usage. Every time the application starts indexing it takes up a lot of system resources and the user can immediately see that CDS is the culprit. A small move of the mouse and CDS will quietly run back to its corner. But, with other applications it takes some time to realize the cause of the system slowdown.


blinkx -

The latest version of blinkx has a number of new innovative features. This version also indexes the metadata of JPEG images and MP3 files, a welcome inclusion since multimedia search shouldn't be confined to searching for just the filenames.

There is also something called SIS or 'Stuff I've seen'. Searching under this category allows the user to search only through what he has been looking at online.

The 'Smart folder' feature can be used to create folders which are topical and over time blinkx will add content to that folder from the web and local drive when it thinks that specific topics are similar to the ones in that folder.

Another new feature is the inclusion of the ability to search on peer to peer networks. blinkx uses the Gnutella protocol and searches can be performed and files downloaded directly from within the application. The user is forewarned that blinkx is not responsible for content downloaded from P2P networks. P2P networks being extremely dynamic in nature, file availability is never consistent.

blinkx also provides a way to search the web in a number of categories directly just with the click of a button. The default search engine is blinkx's own, but this can be changed to one that the user desires.

The blinkx toolbar appears on the title bar of the active window and based on the contents of that window provides the user links to related websites, blogs, products and other content on the local machine. The buttons on the toolbar are initially dim, but become coloured when relevant content is found.

The toolbar is supported for the Internet explorer, MS word. Blinkx claims MS Outlook is also supported, but did not work with Outlook 2003 on my computer.


Filehand search -

The best feature of Filehand is the small toolbar that appears above each search result when the mouse hovers over it. The most useful feature lacking in all the other applications is the ability to move from one search term to the next through the entire document. This prevents scrolling through the entire document looking for the search term. The amount of extract displayed can also be increased or decreased. Quick navigation to the beginning or end of the file, or even scrolling through it is possible.

Filehand search also indexes the metadata of MP3's, but does not support audio or video files. Searching through attachments in mail is also not supported.

Similar to GDS, Filehand search also provides relevance ranking.


MSN desktop search -

MSN desktop search tool's strength is its tight integration with Outlook. It can search through the Outlook calendar, tasks and notes. MS Onenote is supported too. Certain files can be modified to enable a large number of search shortcuts which can be quite handy. It does not search through the IE browser cache and history though.


Yahoo desktop search -

YDS gives extra importance to e-mail search and it works like a charm. A separate tab is provided to search through attachments. The one area where YDS beats the competition hollow is in the speed of its searches. Results are provided as the user is typing his query and is blazing fast. Although similar instant feedback systems are present in blinkx and CDS, they are nowhere as quick as YDS.

Another killer feature of YDS is it's ability to preview any file format in exactly the same way as it would appear in it's native application. Even multimedia files can be previewed within YDS itself! This file viewer support for all the formats it supports is really what sets YDS apart.

YDS does integrate online search, but, it's not integrated into the application like in CDS. YDS also provides a search tool bars for Outlook and the Windows taskbar.

Searches can also be saved to be used at a later date, a feature unique to YDS.


Search results -

All the applications provide a means to sort the search results according to the modified date, file type etc. but only GDS and Filehandle search use their own algorithms to determine which result is most likely to match the user query. The other applications simply display the matches from their database of keywords. So, the quality of searches is not an issue in the keyword matching applications and in the absence of a well constructed search query, relevance ranking is more of a hit or miss game.


Resource utilization -

Resource utilization is always an important factor to be considered in case of applications which always run in the background. Among the desktop search applications, Filehand search was the most frugal in this regard. It required just more than 1 MB. CDS, GDS and MSN desktop serach were within the 10 MB mark, and YDS needed around 20 MB, but, the biggest shock was blinkx, which needed more than 40 MB of precious RAM space. This is a price the user will have to pay for the many features that blinkx supports. Features like providing relevant search results constantly based on what the user is viewing is bound to be resources heavy. Blinkx is definitely not for someone who is already short of RAM.

Indexing is naturally resource hogging in all cases and is best scheduled for when the computer is idle. CDS, blinkx and GDS gain extra points here as they wait to see when the system isn't busy and finish their job.


Privacy -

With the growing popularity of desktop search applications, privacy fears were bound to arise. But, by taking some simple precautions everyone can enjoy the benefits of these wonderful productivity enhancers. YDS, CDS, blinkx and GDS do not support searching of network drives. MSN desktop search and Filehand search do, so, special attention has to be paid to their configuration on networks, to prevent the spread of confidential data.

On desktops, users can take simple precautions like turning off a desktop search application if they do not want their activities to be indexed for later retrieval, like on a public computer. Extra care can also be taken by specifying that folders with private information should not be indexed.

Google has been at the receiving end of allegations about how GDS can be used to breach a user's privacy. This is partly because GDS allows the caching and searching of browser history, especially secure web pages, the kind used in online banking, shopping etc. But, caching of secure pages can be disabled if the user wants to. Also, there are doubts that Google sends personal data to its servers to integrate web and local drive searches. This isn't true. GDS sits in between the user's computer and Google's servers. When the user enters a search query, the GDS application receives the web related results from Google, adds its own locally and displays both in the browser. No data ever leaves the user's computer.


And the winner is....

Google desktop search is a very good application that has a lot going for it. But, the lack of PDF support is something that can't be overlooked. Instant feedback during searches is also missed. Like some of the other applications, GDS is still in beta, and I'm sure it'll be hard to beat when it is ready.

Filehandle search is a minimalist and does the things it is designed to do extremely well. If I had to pick the best feature among all the applications, then it would have to be the ability to scroll to each occurrence in the results. But, it is too slow in its indexing and also takes up a lot of space for its index compared to the other applications. It also does not support searching of attachments in mail, photos or video, which is quite important. With digital cameras becoming as popular as they are, image search is a must.

MSN desktop search is a tool that works for the avid Outlook user. Though it has a polished feel to it, a preview pane is sorely missed. And, I think it will never support Mozilla applications! So, I'll give it a miss for now.

The sheer breadth of features offered by blinkx is unmatched. 'Smart folders' and SIS are extremely innovative features and indexing of JPEG and MP3 metadata is invaluable. If you can get over the lack of a preview pane and the enormous system requirements then, blinkx is definitely for you.

Copernic desktop search will always have many users simply because of its marvelous user interface. Everything falls easily to hand and the search bar is also very handy. It also supports most of the file formats normally used at home and all in all is an excellent all rounder. If you use Firefox then CDS is a must have as no other application supports it.

The application I will not be uninstalling at the end of this comparison though will be Yahoo desktop search. All the amazing features of X1 at no cost is nothing but a steal. For the vast majority of users who don't use Mozilla or Eudora and don't need to index remote drives, YDS fits the bill perfectly.


The future....

The desktop search space is really heating up. As I write this, Mamma.com, a search engine is in the process of acquiring Copernic. AOL has signed up to use Copernic technology in its desktop search application. Microsoft might as well use its monopoly with Windows and do another Netscape.

So, what can we expect from desktop search applications in the years to come? Well, I can only speculate.

Google is in a wonderful position to revolutionize search all over again. Right now, web search results only include content on public servers. But, with GDS, Google could give people the ability to search every hard drive on the planet with its application installed. Privacy and security will be an issue, but not something that can't be taken care of. Blinkx already has implemented a rudimentary form of this in its P2P component, but Google has the ability to make the whole process transparent to the user.

With the exploding popularity of digital content generation devices like still and video cameras, the ability to search through this content will become critical. Otherwise, it will be difficult to cope with the sheer amount of data that is present on our hard drives. Search utilities must be integrated with the already available technology of pattern matching in images and videos. The dream of on demand information at ones fingertips will truly be realized in the years to come and that need will have to be fulfilled on the desktop by these very applications.


Some applications that I did not include in this comparison are -

Ask Jeeves desktop search

HotBot Desktop


Note -
All screenshots were obtained from the respective websites.


Tags -
, , , , , , ,

7 Comments:

Blogger Marc said...

Excellent overview - very comprehensive and balanced. The only techical error I saw was the omission of the fact that the Yahoo version of x1 does not support Tunderbird, Mozilla, or Eudora mail (which the x1 version does).

1/12/2005 02:27:00 AM  
Anonymous Anonymous said...

The fuck up is that blogspot wants you to register to post a comment, anyway I will leave my name though it said "Post Anonymously". Yes, this is a comprehensive comparison of desktop search tools, pretty neat, but I guess most of the people just hit F3 and I have it, may not be great shit like others(YDS, blinkx, etc), but who cares? Anyway many may do use the tools, not me though.

But sure is a great blog.

Lakkad

1/13/2005 11:19:00 AM  
Anonymous Anonymous said...

Definitely a long interview although after using MSN for a while now it feels the MSN coverage was lacking a bit. The reason you see MSN indexing even after it initially says its done is because it keeps up with changes through notifications, a welcome feature not found on all of the products you cover. It also indexes metadata in pictures and music files and supports advanced queries plus has the richest ability to work with the results. It does lack a preview pane which is a shame but overall it seems like you covered the other applications a lot more in depth than MSN's. I tried Yahoo and it ate up my machine while indexing and doesn't index as files change. I like the preview pane but I don't think it's a clear winner vs MSN.

1/14/2005 01:53:00 PM  
Anonymous Anonymous said...

Yeah Wordpress rocks, I recently switched.. I'm trying to find a cool template that I like.. so far i like the one i have now - anyway, my blog's url is www.theblahbrain.com I've got a page rank of 4 [if you know anything about page ranks and search engine optimization] and 309 pages indexed in google.. unfortunately it isn't more than that... anyway...you've got a great blog here! -Paul

11/04/2005 05:26:00 PM  
Anonymous florida mortgage said...

mortgage articles for refinance mortgage home loans

11/09/2005 02:21:00 AM  
Anonymous home equity line of credit said...

A

11/28/2005 01:19:00 PM  
Blogger My VoIp Solutions said...

Hi Thanks for your interesting blog. I also have a blog/site, covering flyfone voip related stuff. Feel free to visit my flyfone voip site.

12/18/2005 04:31:00 AM  

Post a Comment

<< Home