Wednesday, August 6, 2008

Searching : Finding Files Easily On Intranets/Portals

"Miracles aren't real . Poor understanding is the reality"

Today, I spent nearly an hour trying to find something on the internal network at IBM to install and to my disgust, I was just not able to find it. I was looking for an Executable file that was needed to install an application (for reasons of confidentiality breach I can't say what). That's what this blog post is all about ....

Everyone, at some point or the other, has had the extreme misfortune of looking for something they cannot find. This can happen due to many reasons.

  1. They don't understand the taxonomy with which the data is organized
  2. They don't know what the keywords of the product/information are
  3. The search engine was not able to give them the options they 'thought' were necessary to find what they were looking for
  4. The information was genuinely not available but the search engine still gave them the best possible results to the nearest match ( Kind of like when your parents pat you on the back and say 'You gave it all your best' even though you really didn't)
  5. They didn't really know what they were looking for
I am only going to stick to searching for files in this post because if we opened up the discussion for what we can't find on an intranet, we could be here for days. I had a general idea of what I was looking for but could not find it because no one had tagged the page correctly. I finally did end up asking someone from the team that developed the product about where I needed to go and find the product and lo-and-behold, it actually had a whole website dedicated to it.

Now, I'd like to think that I do know what I am looking for. Infact, when I did finally browse through the webpage, I found the same keywords that I was looking for and it wasn't a new page but a page that has been around since last October. So where did it all go wrong? I could go ahead and blame the search engine (which is really basic but does its job as designed) but I think it might be better to take this case scenario and talk about how searches should work.

File repositories
-----------------
Every big company should really invest in a single repository (maybe a cluster environment in the background) for delivering a search option to users for files they need to find. I will say that in some intranets, I have found the search option for document file types like PDF, XLS, DOC, etc. However, there may be a need to look for stuff like ZIP, EXE, BAT, COM files as well because someone will need to install something and if they can't find it within the first two pages or even the first page, they may just walk away discouraged. I know that if I had an option to set the search type and if the browser was allowed to remember my preferences (a simple cookie issue) then I could safetly improve my search by atleast 70%.

Ideal Scenario
---------------
If I chose to search file repositories and asked for a product, eg. Google Talk. Then I would like to see just the files that came up along with a short description maybe two lines explaining what the file was. Instead of a link to a website only and then have to hunt for the file. The downside to this is that any disclaimers may be bypassed but that can simply be fixed by having the link point to the disclaimer/information page first and then a link to the file either at the top in a frame or at the end of the page.

At the end of the day there will be loads of problems with this structure and many companies would have to change the way their pages are made and deployed. People want the most easiest way to deploy their webpage and content. This can of course change over time if the team that is responsible for managing these deployments makes a good template for how they want the information and how the final page will be deployed. Lots of hurdles on the way but I will write a test case to analyze a scenario and see how it can best be tackled.