IDGNS: How about enterprise search? Do you have any new developments in the pipeline?
Sikka: This is an ongoing development. What you have seen from us so far is early thinking in this area. A lot more is to come. You will see enterprise search evolving around different fronts, such as contextual searching. We own the application so we know who the user is, what the user’s role is and what the user is authorized to see.
So called “enterprise search” is a fascinating topic. The issue is a classic traffic-jam between simplicity as a benefit vs. complexity as a valuable feature. Simple search that mimics the search available on the public web — Google — will solve the majority of an enterprises’ search problems. Intranets are littered with mounds of information that simply can’t be found because the company isn’t running simple search.
Of late, there are plenty of solutions from the for-pay Google Appliances to the free OmniFind edition. Savy enterprisey types out there will ask about enforcing regulations and security: how can I enforce Chinese walls and track violations of those walls with “simple search engine”? Good question, maybe the answer is you don’t and you need something more complex. Or, the answer is that snarky response of questioning your question instead of answering it: why not get hip with the 2.0, bro, screw what The Mainstream Government tells you is “the law” ;>
Seriously, though, the thing that makes me queasy about enterprise search is that “enterprise” part. My fear is that all businesses are waiting on enterprise vendors solving the “enterprise search” problem when those businesses should just deploy “simple search.”
Part of the problem is that “simple search” is hard to sell at a high price. There aren’t that many enterprise software bucks in simply providing search over an Intranet. Now, there’s hardware bucks for sure — perhaps some good motivation for IBM’s free OmniFind edition, eh? — but in the age of Google, search seems like it should be cheap and plentiful.
Effectively, this means that vendors need to bake enough value (or, “extra features”) into simple search to make selling search worth their while. Once you start baking these things in, that’s when simple search becomes enterprise search. More importantly, it becomes “we’re spending real money on this so we should think about it” search. Then, next thing you know, you have a committee deliberating for months on end about which enterprise search platform to use or wait to use once it’s released.
In the meantime, employees are left trying to find their 2007 holiday schedule across different content silos. Do we get MLK day off this year, or is it Palm SundaySpring Holidy?
What would happen if those enterprises instead just installed simple search?
Now, to be really dangerous (and American in my thinking), I’d suggest: most of the time, to get their job done, a user needs more than just what they’re “authorized” to see.
Disclaimer: IBM is a client.
I've been playing with Omnifind over our Intranet, and while it will tell me a document exists, if it is secured and I don't have access to it, I can't open it. This seems to satisfy the required security model in a 'simple' search facility. Longer term I'm with you – perhaps we should just start trusting people more, rather than make it difficult to do their jobs …
I've come close to being frustrated enough with my own inability to find useful documentation on my company's intranet to deploying my *own* Simple Search — letting Google Desktop run rampant overnight(s) on the company servers.
Perhaps that's the right way to solve the problem — use the standard IT-management techniques to push Google Desktop or its moral equivalent onto every machine, and let people index only what they already have access to…