Articles on search by OrcaTec’s Herb Roitblat are rare, but worth waiting for. I would much rather point you to his article, and to the articles by Ralph Losey to which they refer, than try and write them myself.
I have just been invited to contribute to a book about a subject which falls within my area of interest and competence, which would involve research and analysis of the kind that my training fits me for, and which would add lustre to my CV. I have declined the kind invitation with some regret. There is one overriding practical reason for this – there are only seven days in the week and I keep meeting the dog queueing for his breakfast on the stairs as I make my way to bed, so I am not sure where the extra hours would come from.
There is also a matter of writing style to be considered – the relatively free and easy prose which I use here and which is my “normal” writing style is not appropriate for a learned or academic work; self-publishing gives me the luxury of not having an editor peering over my shoulder, imposing deadlines, drawing attention to my omissions, and “correcting” my punctuation.
My niche is carefully chosen. I am not a journalist, so I do not generally have to work to somebody else’s timetable. I am not an analyst, so am spared the obligation to research and analyse primary material. The relevant case law involves practical things like breaches of what are, in truth, easily-understood rules rather than complex matters of contractual interpretation or tax statutes. I do not do system specifications, nor am I a user, so I do not purport to undertake the comparative analysis of one software application over another beyond a broad understanding of what each of them does. I am not a computer scientist or an expert in linguisitcs or statistics. My role is to pick out the essentials of all these things and try to dish them up in palatable form to a broad range of interest groups and skill levels, acting as a translator between people whose possibly deep skills in one area might exclude any knowledge of another.
I also have the significant advantage of belonging in a jurisdiction which does not make a religion out of the minutiae of eDiscovery procedure and the relevant technology. One of the reasons why English lawyers can be reasonably sure that the software they use, and the techniques which accompany it, are adequate for their obligations is that most of it has been through the fire of the US Federal Rules of Civil Procedure or the no less onerous requirements of US regulators. One of the consequences of the US burdens is an upward spiral of technological sophistication and related thinking, as technology first meets the challenges and then, because it exists, raises the bar set by courts and regulators and by those who make discovery demands of others. That spawns a high level of jurisprudential and scientifically-based thinking and writing, the existence of which gives comfort to the journeyman lawyer who may not understand it all but who is glad to know that somebody else does.
All that is by way of introduction to an article by Herb Roitblat of discovery and information management software company OrcaTec. Herb’s recent post On Some Selected Search Secrets runs to more than 3,000 words of closely-argued comment, mainly on three long articles by Ralph Losey about the difficulties of validating search results, of comparing different ways of arriving at results and of attempts to arrive at some objective standards. Some of it amounts to respectful disagreement with Ralph Losey’s conclusions, which does not make either of them right, but shows that there is more than one viewpoint.
You won’t necessarily thank me for this, but I think you should read all three of Ralph’s articles (Herb’s first line provides links to them) and Herb’s own analysis. I am not suggesting that you need to understand it all and I do not invite you to reach any definitive conclusion at the end of the exercise. I think you should read them, though.
This is a high-risk strategy, I know. One of the demons with which we grapple (in the US as well as in the UK) is the litigation lawyer who airily admits that he understands nothing of electronic discovery / edisclosure, and I see the danger that steering such people towards heavyweight articles might reinforce the feeling that it is all too complex. Well, quite a lot of things in the law are complex. Dealing with complexity is what the clients are (still) willing to pay for where their commercial objectives require it. I am not suggesting that the lawyers need to understand the science, but they do need to be aware of the issues before blandly asserting, for example, that there is no substitute for human review by their own associates and that their cases and their clients are “not ready” for “black box” technology.
The briefest skim of Herb Roitblat’s article will, if nothing else, show the lawyer that these are deep waters and that there is more than one way of navigating them. It is one thing to be up S**t Creek without a paddle, but quite another not even to realise that that is where you are when a client, an opponent, a regulator or a judge starts asking questions about your discovery, or when you realise, too late, that someone else has run rings round you. These articles will give you a good idea both of what the issues are and of what is at stake.