An English Rechtschreibung — a book systematically, descriptively laying out the current spelling-to-sound rules of English, noting all the exceptions — is, in my view, the first reasonable step towards English spelling reform. It is also the easiest step. Once the Rechtschreibung is done, the next step is to decide on what reforms to introduce that will simplify it, and then, in the hardest step of all, to get the reform implemented.
It would not be enough to simply study the spellings of all words in English. Native English speakers, coping with the insanity that is our orthographic ‘system’, do have something of an internal, presumably consistent, spelling-to-sound model which we use to guess the spelling of unfamiliar words. This comes from a mix of the ‘taught rules’ of spelling we learn in primary school, and also from the experience of reading and hearing the written word spoken aloud and thus acquiring the more subtle aspects of it (such that there are) by osmosis.
In order to make this model an object of study for the Rechtschreibung — so that reforms can be designed to fit within existing speakers’ conceptions of spelling–sound correspondences — a large group of test subjects, with a variety of educational, national, and cultural backgrounds, should be asked to read aloud nonsense words which are created to test the hypotheses the Rechtschreibung puts forward about rules. For instance, if the book claims that ɡ is the standard realization of the written letter ‹g› and that the ’soft’ dʒ is an exception, a variety of nonsense words employing ‹g› in various contexts should be used to find out how native speakers guess which pronunciation to use.