Now blogging at diego's weblog. See you over there!

google desktop search: not yet for me

This is one of those things that has already burned its way through the blogsphere, but anyway...

One of the few things I did (was able to do) yesterday was install the recently-released Google Desktop Search.

Why? Two reasons.

Number one, I wanted to try the latest new new thing.

Number two, I wanted to find a particular document based on a particular set of keywords, and I was hoping GDS would make that easier. I could wait for a few hours for it to index my hard drive.

The download was quick. The installation was a snap. It apparently integrated with Firefox automatically (it wanted to close it before installing), but that wasn't mentioned anywhere. Whatever. Fine. I could see that it was a personal web server. Good solution, nice and seamless integration with Google web.

But I couldn't use it yet, because then there was the wait for Google to index some 60,000 "documents" it eventually identified in my machine. I honestly don't know how long it took--I left it running all day, and when coming back in the evening it was done, so 6-7 hours max.

Before searching, I tried closing it to see what would happen. It complained that if I closed it, any new files created or viewed (web) during that time would not be indexed at all. "Really?" I thought. "That seems kind of harsh. Anyway...". I didn't close it.

Finally I was ready. Double click on the taskbar icon. Search comes up. Type in keywords.

Scan the results.

Garbage. Images, mixed in with documents. Some things given precedence over others because, apparently, the keywords where in the directory name (I'm not sure though). A wave of dissapointment.

Hm. Of course that makes sense, you know, without hyperlinks to provide rank, it becomes more difficult to find what you want.

Okay, so I started trying more accurate keyword sequences. Different combinations.

Time after time, garbage. After a while, I started getting confused at the data I thought I knew. Hm.

I had to reply to email, etc. so I kept working on other things. I had to do web searches, which now, suddenly showed up with a first result that pointed to my own hard drive search results for that term. I mistakenly clicked on them. I did this a couple of times without looking--it took me a while to figure out what the hell was happening, and every time I ended up being left in the mess of my local result list.

It didn't take long before I simply changed my default Firefox search engine to A9.

Early this morning I uninstalled GDS. But guess what. I just realized that A9 is still my default search engine, and I'm starting to get used to the features I was mentioning the other day. Plus, I was browsing through Amazon and I discovered that using A9 gives you a small discount when ordering. Hm. Suddenly I might switch over to A9. I'm still not sure. I've switched back and forth before, but A9 doesn't seem to be as annoying as Yahoo! was with its ads.

So I started off yesterday as a Google user trying Desktop Search, and a day later I'm neither a user of GDS or of the web search? I thought: What the hell...?

I ordered the neuron to do some thinking on the subject and after a short argument, this is what I got to.

Clearly, this is, in part, because of the data I've got. I have many "subtopics" that I work on which contain references, ideas, texts, drafts... it becomes difficult to separate the tree from the forest (or whatever). But other search tools I've used on my data, such as Enfish, have never presented me with the seemingly chaotic results that GDS was showing.

So my neuron came to the conclusion that what was happening was that GDS was too adept at finding information, but, unable to discern proper ordering, it was actually making things worse. Ok, Google isn't God. We all make mistakes. No problem.

But why did I leave Google Web so quickly, and worse, almost without realizing it?

It seems to me that the answer is simple: integration. Google has gone to great lengths to make the local search experience be a seamless continuum with the web search experience, and viceversa. On principle, it's a great idea. However, I trust Google Web to deliver good results. When GDS provides the same experience, I expect the same results. But that doesn't happen (and it's doubtful that it ever can). So suddenly I don't trust Google Search, the experience in general. The web, which is now integrated with desktop results, gets dragged down into the mud by the crappiness of the desktop search results.

Consequence: Google loses a user for GDS. And then, because of the seamless integration, for all its properties. Even if that's not the case, the web results are now tarnished by sharing being equated with the local search results.

It is important to note that this is how I interpret my own actions since yesterday, things I did more or less "subconsciously". I'm not saying this would be a conscious thought process...

I also note that the GDS design integrates with the web design in a way the previous web design might not have been able to do--not cleanly at least. (Am I wrong?). This would prove that GDS has been on the works for some time---my point being that those journalists that say that GDS was "rushed" are wrong.


Keep in mind my only functional neuron is heavily congested, so it's possible I'm missing something, but I think it may be a mistake on Google's part to take the fight to obviously and directly to the desktop. That said, everything Google does is so scrutinized that it's probably impossible to make this a "low key" release.

The desktop is Microsoft's turf. The Web is Google's. I was kind of hoping that Google would behave differently than others in the past and just keep on moving into other markets, rather than retreat (at least partially) to fight into Microsoft's. GDS, together with Picasa (and the ever-present rumors of a Google Browser), seem to indicate that it's going to be a mixed thing at best, with Google leaning on its web side to pull in desktop functionality and users with integrated "seamless navigation" features (e.g., Picasa is to Google Image search what GDS is to Google Web Search, and so on).

Anyway. It was an interesting experience. Given that it was a smooth install/uninstall cycle, I'm certainly willing to give it another try when a new release is out.

PS: My experience was similar in some respects to Don's, but it never got that bad. He noted: "I don't enjoy writing code inside a jet engine". LOL. Also, Jon discusses Firefox integration and points to alternatives. Dave makes a good point about user formats (since one would expect that other major formats, such as OpenOffice, would eventually be supported), and exposing an API (which Jon would points out could be made by "reflecting" content in a format the indexer understands, in this case XHTML). Kevin ponders a lucene-based answer, and Om compares it to Blinkx as well as pointing to other reviews.

Categories: technology
Posted by diego on October 20, 2004 at 2:30 PM


Another cool piece of software that was released while I was away is Erik's GooglME, which he updated to 0.2 beta last week, is a cool J2ME app that provides a front end to Google's SMS service.

The MIDP 2.0 version installs and launches successfully on my P900 but it throws an error--to be expected since the P900 isn't yet on the list of tested devices. (btw, If you have tried the app on another MIDP-enabled phone, let Erik know the results as he notes here).

Categories: technology
Posted by diego on October 20, 2004 at 11:57 AM

removing images from google's image index

Something that has increased load on my blog (and therefore server) beyond its already-high levels is the sudden appearance of images from my weblog as the first hit on many queries in Google Image search (example: "the lord of the rings").

This leads to people linking to the image from profiles in web boards or homepages, without hosting the image themselves. I ended up removing external linking for images, but the hits still arrive since people tend to copy the link without verifying that it won't work for them.

So that's a pain, and tempted as I am to replace the images with ads (heh) or something (such as an H2G2 'Don't Panic' image), I think a better solution is to remove them from the Google Image index altogether, which also means Googlebot won't be reindexing them all the time.

So, how to do it? The first search provided the answer. Done. Now to wait until the next google bot pass and google dance, and see if it works.

Posted by diego on October 20, 2004 at 10:12 AM

adblock (and other Firefox extensions)

A must-have Firefox extension: adblock. Not only improves reading content online -- it also reduces bandwidth usage. I use it with sites that I pay a subscription for and that still insist on bothering me with ads, or when the ads have become so large that they cover a quarter of a browser's window (otherwise, I allow ads, since it's clear that it's how a site pays for its bills, something we all need to do).

Other Firefox extensions I use regularly: EditCSS, JSLib, Venkman and WebDeveloper.

Update: Luke notes, via email, that adblock does not save bandwidth--it only hides or removes the elements from the display layer. Good point, and thanks for the correction! I wonder how hard it would be to add that feature... it would probably still require downloading in some cases, to be able to calculate the width and/or height of what you're not including (if the size is not available in the tag) to maintain the layout...

Categories: technology
Posted by diego on October 20, 2004 at 9:30 AM

Copyright © Diego Doval 2002-2011.