Well, the rumored Google Desktop search app is now reality

Danny, with some help from yours truly, has written two articles about the Google Desktop Search.

+ Google Desktop Search Launched (http://searchenginewatch.com/searchday/article.php/3421651)
+ Privacy & Security Issues With Google Desktop Search (http://searchenginewatch.com/sereport/article.phpr/3421621)

I've been using GDS for the past two days and I'm VERY impressed. However, as Danny correctly points out in the "Google Desktop Search Launched" article, other desktop search apps like Copernic's (http://www.copernic.com) recently launched product and NEO (http://www.neo.com) are still valuable tools.

Let us know what you think.

Nifty. This is going to be a tremendous assistance to my memory-challenged existance. I like the local cache feature.

When I see things like this, the single thought that comes to mind is that this is all spyware dressed up as candy.

I'm not sure how others feel about it, but I'm personally uncomfortable in the way that these devices are essentially set-up and distrubuted solely for collecting marketing data.

However, as marketing is nothing more than manipulating consumers, I expect to see this wonderful new marketing device shepherd people sheep-like into using it, so as to empower billion-dollar corporate giants with free research.

My cynical British 2c.

On the privacy side, there's an opt out (although it could be an opt in...)

A good move for Google, but how many of the public will install it compared to the amount who use Google? A very low % I'm sure you will agree - esp. as they may not be able to install at work.

BUT if they could get it onto the public's radar another way...but how?

Well, I think MS and Yahoo have an advantage in doing this, because they have IM software. These apps now have web search boxes, how long until the next upgrade is a LookOut-powered local search feature in MSN Instant Messenger? With the ability to save pages and search histroy, of course, like Furl and company.

Of course, Yahoo don't have a local machine search app as such - but Yahoo! people + AV people + FAST people + Overture people...will it be long? Built into Yahoo! Messenger? Or install as part of the toolbar?

And then we have this situation: millions of IM users get local machine search. They use it because it makes tasks easier and they may see contextual ads (PPC probably) from Overture. Inevitably, even for local-machine searches, this is more inventory for Overture to sell.

SO Google needs a way of making installing the desktop search app a no-brainer, no hassle thing to do (after the beta test). Via the Toolbar? That's one way. By persuading IT depts to install it and train staff to use it to increase productivity? Another, privacy concerns aside.

And then we have lots of users with local-search, maybe via IM software, which have user profiles. With location and personal interest data.

Can you see where I am going here?

Yup, search engine results served based on the location specified in your Yahoo/MS Passport profile AND behaviourly targeted ads based on this profiel on content sites etc. Some if you type "car dealers" into Yahoo!, it knows you live in Des Moines from your Yahoo profile, and shows car dealers in Des Moines. This will target search inventory more geographically AND provide more releant results, IMO.

Of course this is speculation with some leaps of faith thrown in (privacy laws? will users want to connect their profiles up to these services? what about searching for info on a different area? what will the conversion rates be like for this traffic? etc.).

This is a good move from Google - but like LookOut and other local search tools, they need to find a way to get them into the public's sweaty hands, being used every day.

So maybe we'll see the GIM (Google Instant Messenger!) - a combined Google Instant Messenger/Send SMS/local search app. If it could work with all the IM networks...Yahoo IM, MSN IM, ICQ, AOL... now that would be a USP for the public...

</speculation ends>

It seems to me that Danny Sullivan has failed to explore a couple of important privacy issues.

1. Okay, so the desktop search software phones home with statistical and reliability information that is non-personally identifiable. You can opt out of this feature. Has anyone put a packet sniffer on this phone home? And I beg to differ that it is not identifiable. Even if it doesn't offer up your Google cookie ID when it phones home, it certainly offers up your separate, unique desktop ID, as well as your IP address. From what I can tell, every time you do a search from the desktop that includes results from both the Google index and from your desktop, Google gets your cookie ID from reading your cookie, as well as your desktop ID from a "s=" number in the URL. These two numbers are already associated with each other at the Googleplex, so even if Google gets just one or the other, that's all Google needs.

2. It looks to me like the desktop search software updates automatically when new versions are available, just like the toolbar. Are we supposed to keep checking the version number, and keep checking the privacy policy, and keep checking our packet sniffers?

3. Here's the big problem I have with privacy. The integration of the desktop search with the Google search means that perhaps some 90 percent of all of those who install Google desktop search will get into the habit of searching in a manner that does both the Google web search and the local desktop search at the same time. This means that when they are looking for something on their desktop, they will forget to specify the desktop search alone, and use the combined search instead. Why is this important? Because it means that Google will be able to build up their list of search terms you've been using to locate items on your hard drive. Even though Google won't know the results of this local search, they will know what you are looking for. This is valuable for profiling.

A colleague installed the desktop search and is now unable to open Outlook. She uninstalled it and is still unable to open Outlook. Anyone else had an issue with it? I'm not going to install it until I hear from others....

It seems to me that Danny Sullivan has failed to explore a couple of important privacy issues.
I did do an entire sidebar on privacy issues: A Closer Look At Privacy & Desktop Search (http://searchenginewatch.com/sereport/article.php/3421621).

Because it means that Google will be able to build up their list of search terms you've been using to locate items on your hard drive. Even though Google won't know the results of this local search, they will know what you are looking for. This is valuable for profiling.
If you go to Google.com, do a "desktop" search, Google has no way of knowing that was "desktop" meant, as far as I can tell. That's because you're simply doing a web search -- and it's also showing you any desktop matches. Similarly, it might also show you news, local, book, Froogle matches -- but it doesn't "know" that you meant these.

It could try to guess if you did a search, then actually clicked on a desktop result also shown within web search results. But it still wouldn't know for certain. For all it knows, you did a search, then the desktop integration suggested there was something you might find useful there -- and a very big chance that the data will be web pages you've already seen.

In short:

This is valuable for profiling.
If that's the goal, I'd expect they'd simply monitor and track everything you do when you expressly do a desktop search. They say they won't; it's obviously up to people to believe this or not.

The auto-updating issue is a good point, by the way. Sorry I didn't get that part in. It was mainly because the sidebar was raising some entirely new things that desktop search from anyone raises in terms of privacy.

Have been using this tool for the past 3-4 weeks... was one of the beta testers... it is fast and handy... especially when you lose that file and don't want to go through every thing on your system to find where it may have ended up.

While I'm making no comment here, you should read THIS (http://ap.tbo.com/ap/breaking/MGBTMNRHB0E.html)

Nobody got problems with; .htm (not .html) files not crawl, 101k limit, metatag indexing, and finally the worst for me: all the same date of your computer files (not web history ones) are the same if you dont open it after the crawl ???

Here my first impression (http://inlogicalbearer.blogspot.com/2004/10/google-desktop-first-impression.html) and after a good night and another few hours testing this morning the second one (http://inlogicalbearer.blogspot.com/2004/10/google-desktop-second-impression-too.html).

My thoughts on improving it after my first plays with it:

Emails: often I want to search one or more email folders in Outlook. Lookout does this, Google needs to add this.

In fact I'd say the whole interface for email results is clunky and needs work; in LookOut I can sort the results by date/recipient/subject etc - a core piece of functionality.

Hopefully on their to do list....the way users refine seraches within indiviaul apps like Outlook/LookOut or for certain media formats varies by format, compare searcheing for emails to JPEGs to Word files.

Images: Inegrate Picasa / at least thumbnails of images...Windows Explorer can do that!

Zip files: I've read it can't spider these. Oh dear.

MP3 and all common audio and video files: add these and the ability to search their file names or the tags with artist details etc.

Configuration: for privacy reasons, make the feedback an opt in.

Instead of having to manaully type the paths of folders I don't want spdiered, it should have a directory treee with tickboxes per drive - like eMule, Kazza etc. when you opt out of sharing directories.

OK, that's my 2p's worth, I acknowledge it's a new-born beta...

UPDATE: you can send Google feedback here: http://desktop.google.com/support/bin/request.py

I understand the privacy issues raised by some. You should probably not use this tool - or most tools like this that are expected to come. It's a free choice and I really do respect some peoples need to a higher degree of provacy than others.

Some of my best friends go to extremes to preserve their privacy only paying in cash for everything they buy, never registering products and not even using saving cards such as a Safe Way card (allthough, I DID actually find a way to get the discounts from the card without actually giving ANY personal information to Safe Way, but thats another story ...)

Personally, I did not like when Google added AdWords to e-mail because they actually have to read and "understand" the content of my files to do that. So far, this system only build that knowladge on my desktop for my use and is not used to filter in advertising. However, if they start doing that then I am not so sure I will like it ...

So, I have my limits too - I just think they are a little less extreme than some - but definately more restricted than others.

As long as Google operates within the intent of the law then I think they are free to do what they want. We, on the other hand, as users, are free to use their services and software - or not.

I understand the privacy issues raised by some. You should probably not use this tool - or most tools like this that are expected to come. It's a free choice and I really do respect some peoples need to a higher degree of provacy than others.

Well said, Mikkel. Nothing's free in life. You want a cool tool to search your hard drive? Then there's always a trade off.

This warning on the Google desktop search from Allergic at Cre8asite (http://www.cre8asiteforums.com/viewtopic.php?t=16794) -- Google Desktop : Security Warning (http://inlogicalbearer.blogspot.com/2004/10/google-desktop-security-warning.html)

McF

Another thing to know about Googles desktop search is that it searches all user information on the specified machine. If you have multiple users on the same machine this program is probably not for you. This could bring up some major privacy issues. http://forums.searchenginewatch.com/images/icons/icon9.gif

Google's New Tool Brings Search Home
Free Download Scans PC, Web or Both
By Leslie Walker and David A. Vise
Washington Post, October 15, 2004; Page E01

Some excerpts:

_____________

David M. Burns, chief executive of Copernic Technologies Inc., which recently released its own free desktop search product, said his company spurned the idea of combining Web and computer searches for privacy reasons. When people are looking for private material, they may unwittingly choose the unified search option, he said, which will send their keyword over the Internet to run a Web search at Google. "I don't think people will like having their private keyword sent over the public Internet," Burns said.

...

[Marissa] Mayer said early testing shows that most people will use the new software to search the Internet and their personal computers simultaneously, which will add to the number of Internet searches done through Google. "As a result, we will serve more Web results pages and more ads, and those ads have more chances of getting clicked on. So there will be incremental Web search revenue from this product," Mayer said.

_____________ End of Wash Post quotes


Here are some other points:

The toolbar, if you have that installed, will default to the dual simultaneous search. The deskbar, if you have that installed, will not. However, Google goes out of their way (http://desktop.google.com/support/bin/answer.py?answer=10236&topic=111) to tell you how to do this.

Google says (http://desktop.google.com/support/bin/answer.py?answer=10123&topic=111) that "We realize that many of our users might like us to add a Desktop Search button for the Toolbar, and we may include one in a later version of Desktop Search." How much would you like to bet that they aren't in a hurry to do this?

Clearly, Google is banking on the fact that almost everyone will be using the dual search.

But I'm not even sure it's a privacy issue at this point. When people put in things like private numbers and details into the Google search box, and see their hard disk results in a Google box integrated into the web results, I'm sure they'll get the picture.

Even though Google won't get the private information, it will appear to the unsophisticated that the Googleplex has access to it. And it will be obvious that if they use a dual search to look for something on their hard disk, it will still trigger AdWords from Google because the web search is happening too. Obviously, Google gets your private search terms.

If users who experience this have no objections on privacy grounds, and cannot figure out that Google is at least building up a nice collection of extra search terms under your cookie ID and your desktop installation ID (disabling your Google cookie is superfluous if you have desktop search installed), then I have nothing to say to them.

Privacy issues aside for a second, I'm glad to see Google do something about the horrendous Microsoft search.

Another thing to know about Googles desktop search is that it searches all user information on the specified machine. If you have multiple users on the same machine this program is probably not for you. This could bring up some major privacy issues. http://forums.searchenginewatch.com/images/icons/icon9.gif

Multiple users on the same system have a couple of choices.

1) encryption for sensitive files/folders. I learned a long time ago that if I want anything private on the computer then I need to use some sort of strong encryption and keep the key very, very safe.
2) use the option in the Google Desktop preferences: Do not search list. This option is there to block indexing of directories or domains and it will keep the Google Desktop away from indexing things users don't want it to index.

Another important issue, IMO, is the other applications being used that create log files and cached data of important information that is not encrypted.

A lot of inexpensive and cheap software doesn't encrypt important/sensitive info ... plain text is fodder for all types of applications -- benign or not. This is something that users should remain dilligent in following.

Why would anyone use this instead of Copernic Desktop Search?

Unless I am missing something, its hardly ground-breaking.

Hey Danny,

Just heard your Google Desktop Search Tool interview with John Gordon on NPR Future Tense. Nice job!

Also, I've been using blinkx.com desktop search for a few months and it works very well. Has anyone else used blinkx?

Also, I've been using blinkx.com desktop search for a few months and it works very well. Has anyone else used blinkx?
I tried it for a month or so and found I didn't use it very much. So far using Google desktop search much more. Blinkx is a cool idea though.

Dennis O'Reilly has penned a new article (http://www.pcworld.com/reviews/article/0,aid,117809,00.asp) that will appear in the November issue of PC World Magazine which has an accounting of ten desktop search tools for your hard drive and inbox. Some of the software is well known, and not so well known. Google Desktop was not one of them.

While this is a more interesting read than the Open Sourcery's list (http://www.theopensourcery.com/wordp1/index.php?p=116) and Amit's detailed comparison (http://labnol.blogspot.com/2004/10/detailed-comparison-of-desktop-search.html), it only accentuates the fact that everyone's view of what Desktop Search should be and how it should function are widely varied. All three had different choices for top dog in desktop search. (Btw, Amit mentions Filehand Search (http://www.filehand.com/) which was not on either of the other two lists.)

The only thing they could agree on is that Windows Search sucks and Google Desktop is an interesting novelty, albeit a hot topic garnering a lot of buzz.

Commonalities in Desktop Search
All of the software have something in common though, they do not support too many non-Window file types. If they do, it is made to order.

The fact is some of these programs leverage a portion of the Windows operating system known as the Catalog Index Server which drops a file called index.dat onto your hard-drive. The aforementioned Filehand Search states that the program will not function properly if you turn the Index Server off.

Carlos Perez says (http://www.manageability.org/blog/stuff/what-files-does-google-desktop-index) that Google Desktop uses the index.dat file to locate URLs you visit, the email you receive and send, AIM chats, your cookies and apparently everything you open with the file manager. He also has a stark warning about spyware that already exploit this file:

"Google Desktop could make it even easier to exploit, allowing for easier discovery of confidential data on your computers! Now you have to make a call. Do you trade convenience for a potentially massive privacy hole? The problem is compounded when you have no option but to use Internet Explorer, the primary entry point of most Spyware programs."

I do not hold a PHD in Computer Science, so I will take this statement with a grain of salt. I will concede however that with all the choices that we now have out there ... are any of them really that safe to use?

Faster Search using Windows Index Server
So why is it that Windows Search does not utilize this Index Server as efficiently as these programs can? Good question, of which I do not have an answer for.

But Windows does have another file search available to the user, if you look in the right place. This search is lightening quick and searches a lot more than just email and html pages.

Located in the Control Panel's Administrative Tools is the Computer Management snap-in. This snap-in allows you to configure and create custom catalogs that can be queried within the snap-in itself.

The catalog that you will want to query is located in the Computer Management snap-in under Services and Applications > Indexing Service > System. You will see the "Query the Catalog" link, which will open up your search form. I tried a couple of searches and the results were instantaneous.

Another feature is that you can create custom snap-ins that will hold your searchable catalogs. This allows an easy one-click access to your custom catalogs -- all of which you choose what gets searched and what directories will be included to be indexed.

Here is a complete guide on Creating/Configuring a Catalog for Indexing (http://support.microsoft.com/default.aspx?scid=kb;en-us;308202&FR=1&PA=1&SD=HSCH) at Microsoft Support.

I am exploring this (http://www.webstractions.com/news/2004/10/desktop-search-closer-look.html) more in detail. It is not as pretty as the other alternatives, but it is just as effective. This method does give you far more control than any of the other alternatives, namely you are in control without the use of a third-party software and therefore no security issues.

Hi Dodger,

I try this morning the Microsoft Catalog Index with XP and was quite disapointed by this. Yes is fast as Copernic, but that's it. Less files found than regular XP searches and Copernic and a lot less options and flexibility.

But using Slogger + MSCI give better results than GDS and found even words in CSS files but not in PDF.

BTW : Did you found a way to get a shortcut to go rapidly to MSCI or a Alias ?

Yes I agree. I said in the end that "it is not pretty". :D

I just wanted to show that Microsoft does have a search hooked into the system that is actually fast. Why that cannot be taken an used to their advantage in the first place is a mystery.

You can create a shortcut for this snap-in and put it on your desktop if you want. Also, there is a way to create your own custom snap-in module for just the Services & Applications portion of the Computer Management module by using the Microsoft Management Console.

I noticed that there were some files that are not in there as well. Specifically, I do not see any email files. It might be that the Indexing Service is not configured to scan them (I am almost positive it isn't). But this thing does scan a lot of stuff that you would not even think of, like text inside of .dll, .ini, .hlp, etc. In some ways, if you are a techie -- this thing can locate a lot of good stuff for you.

I also noticed that the index is not set to scan the Internet cached pages. It can be configured to scan them though.

The form they use for search -- I believe it is a DHTML vbscripted form. I have found some technical information in the SDK at MSDN about it. Evidently there is a sample application on their SDK somewhere. It would be nice to get a hold of that and see if this could be used to make a more "user friendly" interface. Still digging for that.

I have found the Google Desktop Search tool really cuts down my wading through endless sub-folders time. I generally agree with the "Search Don't Sort" philosophy espoused by Google.