General discussion

Locked

Archiving of Knowledge

By Hockeyist ·
It could be said that the value and success of a company is measured by the amount of knowledge it has. Gone are the days when companies can be mostly brain dead and still be successful pumping out low-tech widgets (although some still exist).
Archiving of data is not given a second thought by most business/IT people. Data is archived to a media, generally off-line, and never accessed again. Being off-line it's out of the reach of knowledge mining solutions (even simple solutions such as Isys).
I thought about this problem several years ago when searching archived data for our staff scientists/engineers. I would spend several hours per week retrieving documents that were used for their knowledge content. I had plenty of space available so I moved all 10 years of archives (nearly all documents and spreadsheets) back on-line to an old server which was an instant hit with the staff. A lot of time was saved by staff because previous knowledge was used/referred to.
Key employees leave companies taking their expertise with them and leaving behind valuable knowledge in the form of documents etc.
This poses a question. If companies accumulate knowledge and it is stored in the form of data and this data is archived because no-one gives a rats about data content, only last accessed date, then what is happening to the knowledge of your average company? Is knowledge being archived along with useless data?

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Comments

Collapse -

Not just archived, thrown away

by Tony Hopkinson In reply to Archiving of Knowledge

I've seen a large increase in data archives as well. Principally this was diven by legislation though. Manufacturing traceablility requires you to store production information for a least ten years for instance.. Also as in your case, people have got big benefits from being to analyse past performance over longer time spans to improve future. My last archiving project enabled a scheme to save about 700k a year. This saving was always there, but no one could prove it so the investment in more IT could not be justified to the bean counters.
Knowledgewise, you can only go for standard procedures, help, troubleshooting sheets etc. Or a full blown expert system, most of that tends to be in people's heads. Lose them you lose it. Trying to get management to invest the pathetically small amount to start collecting and storing the data is still difficult though.
The only useless data is data that you can't use, not data that you don't use at the moment.

Collapse -

definately

by Jaqui In reply to Not just archived, thrown ...

I always try to keep all files, in an accessable form.

thinking about building a db app specifically to be able to just keep it online, and increase searchability.

picture having all the contents in any given file in a db that can be searched for and displayed, as the content is stored in a db, rather than just the file name / location and a summary.

Collapse -

Mere data

by deepsand In reply to Archiving of Knowledge

Without the means to effectively extract information, there is no knowledge, merely data.

Collapse -

True but

by Tony Hopkinson In reply to Mere data

You can always knock something up to provide information from data, you can't get the data back once you've thrown it away though.
A linux/mysql solution and a population routine is incredibly cheap. Just keep adding population routines and keep all the data, there's always a use for it.

Collapse -

Sure, but ...

by deepsand In reply to True but

more often than not, the search & retrieval step gets short shrift.

Collapse -

Not in my experience

by Tony Hopkinson In reply to Sure, but ...

Usually unlinked data islands in varying formats and platforms with no real links between related items and then they want an app to pull it all together. It's certainly true one without the other is a waste of time, unless you've got the data and some very capable power users.

Just sorted one manufacturing unit out, loads of separate time stamped data about various phases of production, no product id !!!!
Huh ?

Collapse -

A lot of cheap options

by Hockeyist In reply to Mere data

I looked at Isis s/w and found it to be more than suitable.

Collapse -

Comment on causes

by LandoTek In reply to Archiving of Knowledge

Two problems contribute to this- 1)the exponential increase in the amount of data generated, regardless of its worth, and 2)the decrease in a relational style of working.
Although organization gurus like David Allen ("Getting Things Done") would have us reduce the amount of information we retain, Google seems to be addressing the information glut by letting us search it more quickly. As storage costs go down, we can keep more available.
The second cause is a result of the first. Workers are increasingly "componentized," that is, moved in and out of positions with little thought to their increased value due to acquired knowledge and skills. Management thinking seems to look at a factory model even when dealing with people. The amount of information stored electronically makes it easier to fall into the assumption that all of an employees knowledge is captured for the benefit of the organization.

Collapse -

Two categories here

by Tony Hopkinson In reply to Comment on causes

Performance and Knowledge.
Performance data, collect as much as you can for as long as you can. Just make sure the structure is right. Timestamp everything and make sure that there is a way of cross-referencing it later on.
Knowledge that's a hard one. A procedure how to restart system x is a wonderfully useful thing, but it always assumes a very high level of knowledge in the reader, and the bulk of that will be in the reader's head, unless you are going to document their entire education and work experience to date.
A good article, how to, what I did when database is always useful but to paraphrase deepsand only if you are capable of gleaning information from the collection of words.
Unless we can come up with a way of writing an expert system that will cope in a dynamic environment as opposed to simply hiring one and then asking him/her to sort out the problem.

Related Discussions

Related Forums