One of the documents, DJBCJF-2, subsequently known as ``the paper,'' was not comprehensible to computers, not something that most people would call ``software.'' Two of the documents were software in the C language. The other two documents, DJBCJF-5 and DJBCJF-6, were intermediate forms of information, challenging any naive attempt to draw a line around ``software.'' The government didn't draw a line; it clearly labelled all of the documents as ``defense articles.''
I sued in February 1995. In June 1995, the government sent me a letter stating that DJBCJF-5 and DJBCJF-6 appeared to be ``technical data,'' not ``defense articles,'' and that DJBCJF-2 did not appear to be subject to the regulations. In August 1995, the government lawyers said to the District Court that my ``theory'' that the government had imposed a prior restraint on DJBCJF-2, DJBCJF-5, and DJBCJF-6 was ``unfounded.''
In its April 1996 decision, the District Court reviewed the facts: ``On October 5, 1993 the ODTC notified Bernstein that all of the referenced items were defense articles ... As the court has noted, plaintiff had every reason to believe his paper had been determined to be a defense article until defendants' clarifying letter of June 29, 1995. ... [The government lawyers] conclude summarily that both the definition of cryptographic software and the exemptions from this definition are clear to a person of ordinary intelligence. This seems to be a bit of dissimulation, unless it is a confession, since the ODTC itself mistakenly classified Bernstein's academic paper as a defense article under Category XIII.''
In July 1996, the government continued to deny that it had ever censored the paper: ``As defendants have set forth, the two commodity jurisdiction determinations pertained to Snuffle software, and not the other explanatory items submitted.''
In its December 1996 decision, the District Court again reviewed the facts: ``On October 5, 1993 the ODTC notified Bernstein that all of the referenced items were defense articles under Category XIII(b)(1) ... This court noted, in considering defendants' motion to dismiss, that Bernstein had every reason to believe his paper was determined to be on the USML until June 29, 1995.''
In December 1997 oral arguments, an Appeals Court judge asked the government lawyer about the paper: ``At some earlier point, the government was contending that Bernstein's papers on this subject were subject to licensing. Now they've abandoned that position.'' The government lawyer tried to focus on the software, although he admitted that I had submitted several separate documents. The judge asked again about the paper. The government lawyer dodged again, and tried to pretend that there was no issue: ``I think at this point ... there's no debate between the parties as to ... what the government thinks is subject to licensing and what the government thinks is not.''