The opinion of the court was delivered by: Gregory G. Hollows United States Magistrate Judge
FINDINGS & RECOMMENDATIONS
The trial judge, Honorable Lawrence K. Karlton, referred the matter of discovery, and specifically the production of voluminous amounts of electronic and hard copy discovery to the undersigned, to advise on the impact of such discovery to motion and trial dates set, or to be modified, in the above criminal case. Unlike the usual discovery dispute -- not enough produced -- the dispute between the parties involves too much produced, in too many formats, and whether the defense has been given a fair opportunity within the parameters of an adversary system of criminal justice to make use of that discovery.
The undersigned first analyzes, as best he can, the history which led up to the present situation, the real problems faced by the defense, the solutions which have been launched at this time to alleviate any problems, the matter of a web-based common data base equally accessible to both parties, and finally the undersigned's evaluation of the defense readiness for motions and trial in light of or despite discovery/resource problems. Case Background*fn1
The United States believes that defendant Salyer, the one-time head of a large food processing company(s) involving, to a large extent, tomato products, and assisted by persons both within and outside his company, committed several commercial crimes over an extended period of time. Specifically, Salyer is charged with racketeering activity, falsification of records and antitrust violations. In general, Salyer's company, SK Foods, acquired tomatoes and other vegetables from growing enterprises, and through brokers or otherwise, sold processed products to large companies engaged in using the wholesale food products in retail/consumer products, e.g., catsup, salsa, spaghetti sauce, and the like. The sixty-five page superseding indictment alleges that Salyer: preserv[ed, protect[ed] and enhanc[ed] SK Foods' profits and customer base through acts of mail fraud, wire fraud and bribery; caus[ed] purchasing agents and other employed by SK Foods' customers to breach a duty of loyalty and fidelity which they owed to their respective employers; increas[ed] SK Foods' profits by fraudulently inducing certain of SK Foods' customers to pay for adulterated and misbranded processed tomato products by causing the falsification of mold counts and other grading factors and data contained on quality control documents that accompanied customer-bound shipments of processed tomato products that were produced, purchased, and sold by SK Foods, and by actually shipping the adulterated and misbranded tomato products to customers along with the altered documentation;
Superseding Indictment at 10.
SK Foods and its related companies were not a small operation. It sold
multi-million dollars worth of product to companies located throughout
the nation. It may well be (or was) international in scope, but the
focus of this prosecution is United States based. As is the case with
all large companies with many customers involving large quantities of
product, immense amount of records, paper and electronic, are
involved. The many years long scope of
the investigation (1998-2008) exacerbates the volume of records
potentially relevant. The multitudinous number of companies with which
SK Foods did business again increases the volume of pertinent records
in that SK Foods included documentation with or about its shipments to
the companies, and the companies themselves kept their own records. SK
Foods use of "independent" brokers again multiplies the near mountain
of records. Added to this vast array of documentation are the records
created by the investigative agencies regarding its investigation
including wire recordings/transcripts, subpoenaed documents, and
documents acquired by search warrant. There exists much duplication in
the records acquired, but they must be examined to some extent to see
if discrepancies are extant. It is probably no exaggeration to state
that 1-2 terabytes of information are involved. A very visual
definition of the amounts of data involved appears in
www.whatsabyte.com and is set forth in the footnote below.*fn2
when one consider the terabytes of information in this case are
comprised of thousands and thousands of individual records, the mass
of documentation acquired in the investigation and turned over in
discovery is extreme.
As set forth in the government brief to the undersigned, the categories of documents produced to Salyer in discovery from different sources, and which nearly encompass the waterfront of produced materials, are a useful way to view the conceptual problems involved in production of massive amounts of electronic information:
1. The first category consists of documents such as the government[']s case files, interview reports, photographs and items received from the informant. These were OCR'd and scanned into .pdf form and produced almost all with accompanying text searchable files.[*fn3 ]
2. The second category consists of third party records [understood by the undersigned to be non-SK Foods documents, e.g., Kraft Foods, Frito-Lay, Antitrust Division, and so forth] received by the government in electronic form. These were copied for the defense in the same fashion received. Almost all are text searchable [government included a footnote saying that all would be text searchable soon] and came with load files for Summation and Concordance.[*fn4 ]
3. The third category consists of third-party records received by the government in paper form. These were scanned into electronic form and produced in text-searchable [format] with Summation and Concordance load files.
4. The fourth category consists of all paper records seized pursuant to search warrants. These have been meticulously organized in boxes by item number in the search warrant and are easily accessible to the defendant. [undersigned's explanation][On Tuesday, April 5, the undersigned visited the FBI location where these documents were kept. Many boxes of documents existed in pods and in two rooms inside an office annex. However, the bulk of the documents were well organized and labeled shipping documents. These documents reflected a particular product shipped at a particular time along with product specification information. While there were many, many of these boxes, each containing voluminous records, if one had a particular shipment in mind, it would be relatively easy to acquire the documents for that shipment as the boxes were well labeled and organized. The remaining hard copy documents (some of which had been scanned and delivered to the defense already) were a pot pourri of records seized as a result of various search warrants and subpoenas. While the boxes were labeled, it would be difficult to understand what had been placed in each box without looking through it. Handwritten search returns were available for review, but as is generally the case, the descriptions on the return are fairly general. Arrangements were made to supply the defense with a master index (assumed to be fairly accurate) of all the boxes contained at the FBI facility. The FBI facility is a very secure setting. Any visitor is under escort at all times for chain of custody purposes and other reasons. This makes it difficult for the defense to engage in a review of documents in that conversations amongst the defense team are kept to a minimum. The fact that the boxes are in different locations adds to the difficulty; however, if one knows what he is looking for, the difficulty is manageable. The pods setting also makes it difficult to review documents when the weather is inclement. There are no desks or chairs in pods setting effectively limiting the time review can be made due to discomfort. Any copying to be done must be done by a copy service brought to the FBI facility for that purpose. ]
5. The fifth category is a subset of category 4 and consists of selected shipping records tracing [according to the government] the shipping of moldy and/or mislabeled paste with altered documentation. These were culled by hand over a period of months in 2008. They have been scanned and the hard copies organized into about thirty binders which have been presented to the Defense. [undesigned's note][These binders constitute the heart of the government's fraud case. However, defense counsel refused to concede that this organization of the government's case aided the defense discovery efforts; this dispute is referenced further below.]
6. The sixth category consists of the SK Foods inventory and quality control system, known as the "AAS" system. This was produced to the defense as a forensic image along with an explanation of how to recreate it as it ran at SK Foods at the time of the search. [undersigned's note][The forensic image referenced was actually the entire inventory and quality control system database as it appeared on the main SK corporate computer/server as of the date of the search. This database approaches a 500 gigabyte size (very large). The forensic copy (initially requested by Salyer) was only of limited value. Although it could be made searchable by means of special software, it is more difficult to find what one desires to find than when one was looking for the same information on the actual SK computer system housing the information and using SK systems/software. The United States, through sophisticated software manipulation, and the assistance of present or former SK employees, was able to actually "re-create" the SK computer database, and run it as if that database were being queried by an SK employee at the time of the search. The re-creation has the look of what the SK employee would see at the time as well. Salyer was told how to do this re-creation, but his experts did not have access to the SK employees who aided the government computer personnel when those personnel were re-creating the database, and Salyer's experts may have been unfamiliar with the proprietary software that SK was using at the time. At hearing on March 30, the government offered to "give" Salyer (lend for the duration of the proceedings) an up-and- running re-created database. At the April 5 "hard copy" site review, two of these re-created systems were given to the defense so that Salyer could utilize this re-creation himself in Monterey County. A separate order will reflect this "gift." This re-creation is very useful in attempting to correlate input/modified data on the system with initial handwritten data recordations. A large part of the government's case involves its belief that the computer data was unlawfully modified, and the government plans to prove this by reference to the raw, handwritten data.]
7. The seventh category consists of all of the other computers seized from SK Foods and the other subjects in the case. These were produced as forensic images because that is what the defendant specifically requested. The defendant does not contend that these are not searchable [however, special software must be employed to make forensic images searchable], only that the 25 or so hard drives are not searchable at once. This is not a capability the government has either.
After consideration of all the submissions of the parties, the in-court demonstration, and the visit to the FBI hard copy document storage, the court finds that the following problems/issues have complicated expeditious pre-trial review of the massive amounts of information. But first, the undersigned emphasizes that no party to this criminal action has acted in bad faith. The government has attempted to mitigate the electronic discovery problems encountered by Salyer; indeed, the government has gone the extra mile in supplying Salyer part of its work product in terms of exhibits to be used at trial and re-creation of SK computer systems. Salyer has been initially, legitimately overwhelmed by the amount of information to review and organize, although the extent to which this initial state remains is not completely clear.
1. Too Much Information From Multiple Sources As is the case with many criminal cases in which much of the evidence will be found on electronic storage devices, the present method of operation is seize first, analyze later. The undersigned is not asserting that such is unreasonable per se, as the government has little practical choice but to perform the computer forensics to discover information pertinent to a search warrant or subpoena in a separate laboratory situation. The end result, however, is that much information seized will not be utilized in the case. The defendant, generally and in this case, desires to see everything that the government has collected, evidently not trusting the government's relevance analysis. The problem is, however, that as the information to be reviewed explodes exponentially, not much thought is given by the government during the acquiring phase to the effect on the defendant of having so much information to review, and the effect that such massive disclosure will have on the pretrial process.
And, the "too much" problem here, though not created by the government, is exacerbated because the information comes from various sources. As presently constituted, defendant has to maintain a separate data base, more or less, for each batch of discovery which arrives from a different source. While some issues in the case should not require the querying of every data base, as a particular issue may be associated with only one source, general issues, e.g, the giving of bribes/gifts, others may require looking into numerous sources. Separate hard drives with differing formatting requirements have to be loaded, or one CD replaced by another, has to be viewed. The problem here is analogous to having massive paper discovery organized by source all placed into separate rooms of scattered locale, all with somewhat differing organization. Trying to pin down all the differing responses to an e-mail discussing adulterated food products, e.g., the e-mail "thread," in-house memorandums, corresponding testing results, captured ...