The desktop indexing

| | bookmark | email | 1 comments
For a long time I've been looking for the best solution for desktop indexing. In the past I have tried different solutions, but now I thought the market is much more stable and mature, so I have given it a new try. And to clarify an important aspect: when I am saying the best solution I am refering to a solution that best fits my needs, and not about what some generic solution. My usage scenarios are quite limitted and are in the following lines:
  • search Thunderbird mails (I am active on about 10 mailing lists with medium to high traffic, and I am watching another 20-30 mailing lists)
  • search some specific locations on my computer for different file names
  • (rarely) search for specific text in those specific locations
I know there are quite a lot of entries on the blogosphere about this subject, and I have based my initial choices on some of these readings. I have finally chosen and used the following 3 solutions: The way I've tested these tools were quite simple: index my Thunderbird email and some specific folders on my hardrive. Next, move around files and emails, and see if the search results are correct.
Here are my opinions about the following aspects:

UI

Ranking:
  1. Google Desktop Search: Google continues to amaze me with their simple and clean UI. I've been mostly using the Quick Search box which provide the best experience so far. Indeed, without the power behind the query criteria, you would probably not be able to benefit of this neat feature. Presenting the results in a normal web browser page is another excellent decission Google took.
  2. Both Copernic Desktop Search and Yahoo Desktop Search are coming on the second place, mostly because they use the same type of UI: normal desktop application window, with multiple tabs per search type (mail, file, etc). Probably, I should give a plus to Copernic Desktop Search that comes with Deskbar.

Query language

Ranking:
  1. Google Desktop Search is again on the top position. The power behind the search criteria is huge and till this moment the only missing one would probably be the possibility to specify a search criteria by the filename. I guess this can be very easily added in the future, because I am almost sure that this information is present in the index files.
  2. Once again, both Copernic Desktop Search and Yahoo Desktop Search are coming on the second place because they both have chosen to provide a classical approach with multiple filters boxes.
I know that this ranking is very subjective, but in my case writting down a single query in one box, without needing to move around and select different other subfilters, represents a huge advantage.

General functionality

Configuration

  1. Copernic Desktop Search and Yahoo Desktop Search are both providing a very intuitive and easy to configure interface. You are allowed to choose what folders to be indexed and what mail folders must be indexed. These are kept under continous observation and the indexed are updated. Only the Thunderbird are not indexed on the fly.
  2. Unfortunately, Google Desktop Search preferences page is far from being intuitive. I am still not sure what is happening with the excluded folders, or the reason why its logic is working through exclusion rules and not through inclusion. There is impossible to specify which mail folders to be indexed. The only positive point is that Google Desktop Search can index mails on the fly.
I have to say that this represented a big issue for me (the time to exclude everything just to set a few folders to be indexed) and had weighted quite a lot in the final decission.

Index size

I know that nowadays the storage is not anymore a problem. Still, I don't think you are gonna jump to change the laptop harddrive any time soon, because the index of your files is too big. As I mentioned, the tools were used to index all Thunderbird mails (probably around 50000 mails) and a few harddrive locations containing mainly pdf-s, doc-s, html-s and a lot of zip files (mounting to about 20Gb).
Ranking:
  1. Yahoo Desktop Search indexed everything using about 450M.
  2. Copernic Desktop Search indexed everything using about 700M
  3. Google Desktop Search indexed only emails in about 650M (I stopped it while indexing the folders because it was taking too long and too much space: more than 2Gb).
While Yahoo Desktop Search and Copernic Desktop Search seem to be quite good, I was astonished to see Google Desktop Search results.

Indexing features/plugin

All of them provide quite the same features so I couldn't make a difference between them. I will add a plus to Google Desktop Search because there is a Trillian plugin available.

Reindexing

Initially I haven't thought about such a category. Still, after playing with Google Desktop Search I had to include it, because Google Desktop Search is not able to monitor file changes (move, delete), nor email changes (move, delete). This is leading to situations where the search results are pointing to phantom data. I've been searching through Google Desktop Search Group and even if I understand that sometimes keeping the old data may represent a nice feature, for me having stale results is not an option. Unfortunately, adding to this the impossibility to trigger index updates (the recommended way is: uninstall/reinstall, which is quite an unpleasant way to do it, or through TweakGDS Plugin which is erasing the whole index, so once again not a very pleasant way to do it), finally rendered Google Desktop Search as completely unusable for me. Even if I love the UI, the responsiveness and the way results are presented, missing the possibility to restrict the area of indexes and missing the possibility to at least refresh the indexes transformed it from the top solution to an almost unusable one (I really hope things will get better and I will be able to find in the next versions the features I need, considering that my requirements are not even absurd).

Conclusions

As a conclusion I would say that both Copernic Desktop Search and Yahoo Desktop Search are quite well doing their job. I would definitely like to see simplified UIs and a more complete/simplified query language (and probably a Trillian plugin), but other than these I am quite happy with either of them. I still need to monitor how the performance of my laptop is influenced by these tools on-the-fly indexing, but till this moment I haven't noticed any bad impact. Unfortunately, in my opinion, Google Desktop Search which has the nicest query language and the cleanest/most intuitive UI, is still missing some fundamental features.