What's Discoverable and the Value of Where

Let’s suppose that a user’s local hard-drive (or shared file system) contains the following directories with the file entries listed below:

>>directory1>projects>ibm>file1.doc (note, these are the same file)
>>directory1>projects>ibm>file2.doc
>>directory1>projects>ibm>file3.doc

>>directory2>projects>oracle>file1.doc (note, these are the same file)
>>directory2>projects>oracle>file4.doc
>>directory2>projects>oracle>file5.doc

Now, your comapny is joined in a legal suit and recieved a discovery request for relevant documents to the legal matter. You are now on the receiving end of having to find and produce relevant documents in a very important legal matter. On the surface the request seems clear, "all relevant documents", well after many rounds of legal back and forth (the discovery order is to broad, etc.) the discovery request is now understood and amended to specify "documents" containing the word "IBM"

Now comes the conundrum.

Let’s cover the simple scenarios first, if file2 or file3 contain the word "IBM" they are clearly discoverable pursuant to the request and should be turned over.

But, what if file1 does not contain the word "IBM", though it is clearly in a folder named "IBM" is this document discoverable?

If the supplier of the information withholds the document pursuant to the agreement that only "documents" containing the word "IBM" are relevant, is the supplier within his rights to withhold the document?

Let’s suppose that this document is withheld, but later discovered under a broader discovery request, will the opposing counsel claim foul and will a judge potentially find that while the letter of the discovery request was met that the company’s action was in bad faith and although the "document" did not contain the word "IBM" that it was obviously relevant, having been stored in a folder named "IBM" and should have been included - and the courts has no choice but to find that it was excluded intentionally and impose sanctions or even worse instruct the jury that there was tricky business afoot by the company and this should be considered as they deliberate, (ouch).

What about the fact that the same document was in a folder clearly marked "Oracle" wouldn’t this be enough to prove that this document was relevant to "Oracle" and not IBM ? By the way, the contents of the document did contain the word "Oracle" and not "IBM".

Ah, the never ending complexities of discovery, a lawyer with an engineering background in systems can probably have a field day.

And, the point is exactly that - these are all issues, questions, suppositions, assumptions for legal counsel, many with answers that rely on their ability to make (or not) successful arguments in front of a judge as well as dig up any relevant case law and or legal precedents.

These are not questions suitable for software vendors providing the underlying technologies that support responding to legal discovery requests.

However, there are some very pertinent quetions that the above scenario(s) do present for software vendors that are more important to your companies ability to support any number of scenarios; I’ve addressed the importance of flexibility before and here we go again; a system of record, namely a durable active archival system must support documents, e-mails, instant messages, other electronic data in the form of XML messages (as an example) that represent transactions in various states usually traversing through a series of queues within a message bus, end of day or intra-day report out-put such as FX tickets, trade-confirms and end of day confirms with cancel corrects; all of this data must be recorded as states of time, categorized, and marked for expiration based on regulatory and internal policy.

A vendors ability to handle the sheer volume associated with the above is critical.

Now, if a record must be maintained for 7 years, an organization may take the approach of expiring the document on the 7 year anniversary to the second, or go 7 + 1 year, etc.

Systems should provide event based flexibility to support liability and risk factors, for example the logic and reasoning that goes into expiring a tax document after 7 years is pretty simple, that’s what the law allows.

But, are complex business processes for global enterprises ever that simple cut and dry, well not in my experience.

Expiration logic which can be expressed as *expire* tax documents in 7 years, if account numbers beginning with "Z" ("Z" equals international) have passed all external audits (maintained in another system) and if company names contained within the documents (as a indexed field) are not part of any legal matter irregardless if the tax documents are on legal hold, also based on the previous triggers send these documents to the legal department for review and purge or retain with a recorded attestation.

This example is probably more in keeping with complex business processes;

So, back to the matter of :where: obviously where documents originate from, reside, or are moved to can have consequences to their admissibility.

So vendors should record this information as part of a "complete" record, referring back to another post, we have discussed the differences between complete and correct, they are distinct and both critical.

How a vendor tracks where and to what extent, such as the physical asset is important, see that your vendor understands and incorporates "where" as part of their overall systems.