Lookout better than Desktop Search; Lookout can search archive files too.
As Google desktop search doesn’t install on my machine, I decided to install Windows desktop search. Most of the time, I need to search my e-mail (for which I use Microsoft Outlook). The document search is not very important for me. As it happens, Microsoft’s Desktop Search does a decent job about the search. The problem, is that for it to report results from the archives (the .PST files), it needs the archives to be loaded in Microsoft Outlook. I prefer not doing that for many reasons. One that I don’t want the .PST files to be corrupted in case Microsoft Outlook crashes for any reason. Also, the load time for Outlook increases, and I have seen some memory consumption increase too. Unfortunately LookOut – the precursor to Microsoft Desktop search is no longer available for download (Microsoft decided to let users use only the desktop search). Luckily for me, a colleague of mine had the installer on his machine (I also probably have it stashed on one of my archive media). And the best feature of LookOut is that once it finishes indexing, it doesn’t need the archive folders to be loaded into Outlook. A big relief for me. And for those of you, looking for LookOut’s installer, I have uploaded it as part of this blog post. You can find it here – LookOut 1.2.
calculate fractal dimension using logartihms
A simple google search will give you that – isn’t it. But, I found this to be the most simplistic and an elegant explanation even for a layman to understand how to calculate fractal dimensions. And once you understand how to calculate fractal dimensions, you can also understand why they are so special. Also, to tie logarithms in calculating dimensions, now that is what is appealing. Also, another nice explanation of how to create your own fractal is out at Chaotic Utopia.
Improving long tail search results
Kartal Guner, the chief architect at hakia, wonders how the long-tail of the queries can be bettered. And his ruminations, are undoubtedly right. Each of our queries, generally might fall in the long tail of the queries. Most probably, because for the rest of the information, we know exactly where to go (would you search for a site to find the latest news ? May be not ). In such a case, how does one better the long-tail queries ? A immediate thought on this follows (immediate, because, I got the thought as soon as I read the post, and thought, let me blog about it 🙂 ).
I think one of the tougher challenges in modeling the long-tail is to help the user frame the query right. For example, if the user wanted a comparison between a Sony Bravia and a Samsung LCD TV, how does the user frame the query ? Examples might be – Which is better, Sony Bravia or Samsung LCD; Feature comparison between Sony Bravia and Samsung LCD et. al. Given these kind of queries are quintessentially part of the long tail, one step to making the long tail search easy is to help the user with framing their queries. There were attempts in the dot-com era (I don’t remember the sites now), which helped the user frame their queries right. Accoona.com is also attempting that (of course, categorizing is something, that clusty.com has been working for a while now). But, neither of these two have got to the point, where the user can find what (s)he was looking for. Between these two, I prefer clusty, in being able to categorize properly.
Also, domain specific search is something that is an interesting area. Most of the long-tail searches, fall in certain domains. If the search engine can collate these queries, may be within the long-tail searches, there can be a short-tail that can be created; and applying domain specific searches within that might give a better result. There was a site back in the dot-com era, which specialized in domain search; if you remember the site, please do drop a note. The same for the portals which helped the user in creating a query (sort of a visual SQL builder !)