h a l f b a k e r yWhat's a nice idea like yours doing in a place like this?
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
In the rather old-school manufacturing community, we have accumulated piles upon piles of disjointed, antiquated, and largely useless software applications, which through the wonders of huge corporate beaurocracy, imperfect or lost information, tend to create all sorts of incredible waste in repeated
efforts and time loss. Consequently, companies end up doing the same dumb things over and over again.
Of course each of these databases has their important-sounding acronym, and somebody somewhere will require that they be filled in, although they aren't sure why. Thus, the unhappy release engineer is forced to wade through stacks of systems which are not much better than the stacks of forms they were intended to replace, and take years to learn. While at DaimlerChrysler I counted no less than 11 corporate-wide issue management systems, each essentially doing bits and pieces of the same thing from different directions.
Secondly, retaining corporate knowledge in an organized manner is yet to be well-implemented. Specific groups create intranet websites with their information, but very little ties it all together corporate-wide, and efforts to do so have thus far created more havoc than they solve.
The 'Multiply' website has a feature which directs targeted ads to appear according to the topics raised on the specific user's page; for instance, the ads on my recipe page, once I mentioned 'cakes,' had a half-dozen sidebar ads concerning everything cake-related.
So now, what would happen if the targeted ad technology that drives Google and Multiply and such were thrown at internal engineering databases? Start talking about topic X, and the links for everything to do with topic X suddenly appear on-page. If a problem arises with the foam for car seats which needs addressed and cataloged, the database should go find every prior art problem and solution having to deal with carseat foam that exists in the company.
[link]
|
|
This would also be useful for other fields. "The patient is exhibiting red skin welts along with severe itching and..." *have you considered chicken pox?* |
|
|
This exists in a basic form as expert systems, but these systems are generally built by hand and don't usually learn as you use them. |
|
|
It wouldn't have to be nearly as smart as Cyc. Given a reasonable description of the system in question, all it would need to do is parse the text, and use network associations to assign probabilities of what's useful based on what's popular, like Google currently does. |
|
|
Don't forget a thesaurus filled with custom content. For example, what was called the XYZ project typically gets renamed to the X++YK project. |
|
|
Good luck, though. I've never seen a good internal knowledge-management system in valuable use. There's just tons of cr@p. |
|
| |