Searching the LAN

While finding information on the Internet is free and easy, the same doesn’t hold true for the corporate LAN. In fact, locating important documents on the company network has never been harder, with increasing numbers of client PCs, servers, databases and documents making it a painful and time-consuming exercise.

That’s compounded by the fact that no network manager wants their company’s valuable information to be accessible to every employee, no financial director wants to waste money on search tools, and no employee wants to spend hours locating documents they filed away months ago.

Enterprise search has become a hot topic, with Google raising the profile of the technology. But it isn’t just the 500-plus employee companies that are benefiting from the new software and hardware that promises to save businesses time and money.

So which option is best for your company? It’s not just about comparing licensing costs – you need to decide exactly what you and your employees need from a search tool.

The company librarian

Searching for things is one of those bizarre philosophical borders: the vast tribe of the organised stare in disgust at small roving bands of disorganised searchers. There’s a huge divide in the way people handle knowledge, and perhaps this explains why it’s taken a good two decades for add-on search software to make an appearance in the LAN market.

Perhaps the delay in the emergence of products to search the network has been because networks tend inherently to be a group endeavour, and it’s impossible to work as a group without some organising principle. Every file server I’ve seen represents the workflow of its users through the structure of directories, shares and the filenames it stores.

Every company tends to have at least one ‘librarian’, even if that isn’t their title or role. It’s inherent to most business activities that the flow of obligations through a team means you’re better off asking the librarian than diving into the file structure – plus, they get to feel good about being asked: karma is preserved.

However, there’s a pretty sharp limitation to that kind of structure in companies, and it’s a beacon of hope for all of us disorganised creatives. There are some searches you’ll need to carry out that just don’t follow the structure of your workflow or your storage.

Numerical disadvantage

Shortly after 11 September 2001, all UK banks were asked to declare formally that they hadn’t had dealings with a list of individuals. This list was delivered from the UN, and a good two-thirds of the names included the word ‘mullah’. However, of those, quite a few weren’t typed ‘mullah’ but ‘mu-1-1-ah’. This is a style of representation that isn’t unusual in Arab to English translations, and which was hidden in the case of this list by being presented in plain text (which everyone represents as Courier font, in which the numeral 1 and the letter l look very similar).

Here we had a classic search killer. Finding proof-of-absence of any mullahs in the list, by going to a librarian type and asking them to flick back through the files, was a non-starter. Finding the names with a fairly simple search query or one-off program had a hidden gotcha. What was needed was a much smarter method of looking at the data.

There are languages and technologies to get around this problem: regular expressions, GREP and PERL, and even SQL in an abstract sort of way, are all methods of expressing ‘find all Word documents containing “payment” within the same sentence as “Mullah”‘. But the fact remains, most LANs don’t have in-house PERL gurus waiting for worldwide terrorist atrocities just so they can show off their coding skills.