Ordinary Nothing

I never used any serious desktop search tool until now, not considering the basic OS supplied ones, or a simple dir, or ls, command.

The main reason up until a while ago was that everybody wanted money for their tools. This is perfectly legitimate and fair. I’m actually quite capable of paying money for software I want and use, and have certainly done so in the past. But since I’m fairly organized in the way I keep files, I never saw such tools as more than a potentially minor convenience for me, and so not worth me spending my money.

Recently, however, some has been released for free. The first, or at least the one that got the most hype, was the Google desktop search tool. I considered giving it a spin, but decided not to due to several simple reasons (Which from the first reports I read also pretty much apply to the MSN desktop search tool):

1. If it’s a real local program, and is indexing local files, I want a real interface and not an HTML one. No matter how well made, a browser rendered HTML page is not the same, and has its limitations. For a web service that runs on a different computer, that’s fine. But if I already have a program running on my computer, why cripple the interface?
2. I don’t want to run a web server, and talk to a computer on my own computer via HTTP. Why would I need to allow a search program to provide web interface? Why should I need to deal with networking/firewall/communication settings for a program that’s fully local?
3. I want a clear distinction between what’s on my computer and what’s not. This is apparently different from the vision all those search companies are having, where information could be accessed similarly whether it’s stored on the local computer, or on some server on the Internet. What’s on my computer is mine, and I chose to put it there. What’s on the Internet someone else put there, and most of what’s there doesn’t interest me. Even searching for the same keywords, I’d choose these options totally separately and independently. But those search tools try to make everything more integrated and seamless.

The Yahoo! desktop search program is a program with a regular user interface, doesn’t require to install a web server, and doesn’t get confused between file searches and web searches. So it seemed worth a try. In addition, it’s supposed to be based on the X1 program, which exists for quite some time and has gotten pretty good reviews.

So I decided to download and try it. Now Yahoo! did set up a forum/message board for comments. I tried getting in there numerous times, with hours of difference, but it didn’t respond. I did succeed in entering the message board just one time, but after a very initial browse, to see what’s already reported, it stopped responding again.
So I’ll comment here. I assume nothing I have to say is totally new or unique (How much can I notice from a short use?), so it shouldn’t be a problem if nobody notices it.

The good bits:

1. Indexing is fairly fast.  Of course I didn’t have anything else to compare it too, so I can’t vouch for how it is in relation, but it progressed in a nice pace and it didn’t take too long for it to index (and re-index when I made some changes) everything.
2. The built-in preview is nice, and supported the basic file types I tried to search for. Not having to open each file in an external program just to see if it’s the one I’m looking for is convenient.
3. It’s easy to run additional filters based on file name, type, directory, dates and so on. The main options are on the same window, but arranged sensibly and without cluttering the display.

The bad bits:

1. Changing parameters caused the program to re-index everything instead of trying to sensibly find the differences (If I add a single file extension to index, why not simply search and add those files?). Actually, it’s far worse, simply entering the configuration dialog and then closing it, without chancing anything, caused the program to re-index everything.
2. The search edit box, where the search query is entered, doesn’t match the default system code-page. Meaning to say, the test system was a Win2k computer, with an English interface, but also support for Hebrew. But I couldn’t search for Hebrew characters since when I tried to change the input language I got the "western European" characters instead of the Hebrew ones. I assume this means no UNICODE support, also.
3. The installation placed files in places that were different from where the search interface was looking for them. When I selected a file that it could not preview, it tried to show in the preview pane an HTML file that came with the program, containing the message that the file cannot be previewed. Instead I got an error message that it cannot find the HTML file it tried to show. The files were installed into the program directory, but it looked for them under the user profile directory.
4. When it could not find a file that was indexed, the error came up in the form of a message box. This means that whatever I was doing was interrupted until I pressed the OK button. This is really really bad. It means that when I started to enter a search term, it would cut me off in the middle to show this box I need to confirm. And when I scrolled through several files, again it would occasionally just cut me off and break the flow. Errors like that should be presented without interrupting the user. You can’t find the file? Then let me go on to choose another one instead of interrupting me. Take the same text you’re showing on the message box, and show it in the preview pane instead of a fancy HTML. Due to the previous problem, this came up a lot.
5. Sometime it couldn’t find and open files for no good reason. The files were real, and exactly where it listed them as being. But no, I got an error message stating that the file cannot be found. Many of these files contained spaces in the file name or directory name, but I think not all of them.
6. When the files did contain a space, it was listed in the messages as %20, not as a space. This is fine when sending it to a browser, but not for a regular message box.

The would-be-a-nice-addition bits:

1. Index also email messages from Thunderbird/Mozilla and other mail clients. Not just Outlook or Outlook Express.
2. When rendering HTML files in the preview, maybe allow to use the Mozilla/Firefox engine instead of IE? I don’t think this is a trivial thing to do, and since these are only local files it’s not as critical as when rendering content from unknown web server, but it would be nice.
3. Try to determine file content/type for extensions that were not selected to index. Maybe this would require a separate do-not-index list for people that want this for security/privacy reason. It’s good to have control, and I could (and did) add my extra extensions myself, but I was surprised not to get full-text indexing for text files, even if their extensions were not on the list. On some level I expected that things like .log files, or .cpp and .h files, will be indexed and searchable. They’re just text after all.

That’s either about it, or about what I recall now.

What mostly bothers me about the bugs I found isn’t so much that they’re there (since that can, and probably will, be fixed), but that they’re the things that very basic beta testing should have found. Since it’s based on an existing product, and since Yahoo! are certainly capable of affording a QA team, I’m not sure how these things got through…

Overall YDS is really nice, does a good job, and is much better than the internal windows search option. I still don’t think I need any file search tool enough to pay for it, but for a free version I’ll probably keep using YDS once they get rid of those constantly popping message boxes.

January 19th, 2005 at 1:32 EET | Categories: Reviews, Web | Track comments. You can skip to the end and leave a response. Pinging is currently not allowed.