0

I wrote:

De-identification differs from anonymization in that the latter is supposed to be irreversible while the former allows a trusted party to re-identify the data.

To my non-native ears, "in that" sounds like basic English. When facing a "X differs from Y in that Z" structure, is there a more stylized to phrase it?

I thought of using whereby:

De-identification differs from anonymization whereby data is supposed to be irreversibly anonymized while the former allows a trusted party to re-identify the data.

1
  • 1
    No, they mean different things. In the first, the "in that..." clause applies to the difference. (The difference lies in the way that...) In the second, the "whereby..." clause applies to the anonymization. (Anonymization is the means by which...) – Tim Pederick Jul 12 '14 at 7:44
3

In that is fine; in fact, it is formal rather than 'basic'.

Whereby, on the other hand, is impossible here: it means by which, and therefore claims that the fact that de-identification differs from anonymization the means by which the difference you describe comes into being: which makes no sense at all.

The problems with your sentence lie in what follows the subordinator.

  1. Properly, what follows in that should be structured as a statement which tells the reader how de-identification differs from anonymization rather than merely describing the two processes and leaving it to the reader to discern how they are different. Specifically, the subject of the that clause should be de-identification.

  2. Latter and former have no place in contemporary writing. Your use of them here is time-honoured, its logic is impeccable; but in a complicated sentence like this it puts a wholly unnecessary processing burden on the reader, especially when you flip the terms out of their original order. Readers are not computers, and texts are not programs: find a way of expressing your meaning that doesn't make the reader look up the refrences.

Here are couple of ways to avoid these problems:

De-identification differs from anonymization in allowing [or in that it allows] a trusted party to re-identify the data, where anonymization is supposedly irreversible.

The difference between anonymization and de-identification is that anonymization is supposed to be irreversible but de-identification allows a trusted party to re-identify the data.

These are a little bit longer than your sentence, but they're a lot easier to follow.

I have no idea what that supposedly means, so you might want to clean that up, too.

2
  • Thanks a lot for the thorough answer! (I used supposedly to lay emphasis on the fact that in many instances anonymization turned out to be reversible, e.g. in the $1-Million Netflix Prize some people managed to de-anonymized part of the dataset) – Franck Dernoncourt Jul 12 '14 at 2:25
  • 1
    @FranckDernoncourt I suggest then that you write "anonymization is supposed to be" - supposedly has more of the original sense of it is supposed that. – StoneyB on hiatus Jul 12 '14 at 11:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.