For `Method of delivery' and `Encoding system', respondents' preferences were clearly stated, and may be summarised as follows:
With respect to the other items analysed, the results are less clear cut. To simplify presentation of the results from this survey, we adopted the following summation procedure. In the summary tables, a feature is rated as `mandatory' if it received a score for ` essential' greater than the three other scores combined, and as `desirable 'if it received a combined score for` essential ' and `if possible ' greater than its `no opinion' and `don't want' scores combined. In cases where this procedure resulted in a tied score (e.g. an item scoring 11 both for `essential' and for the other options combined), the feature was given the benefit of the doubt and the higher ranking category chosen (i.e. `mandatory' in this case).
In addition to these summary scores, the following tables also indicate the total number of votes cast for each option, since this varied from feature to feature. The second of the two figures below indicates the number of votes cast for options other than that indicated. For example, an item rated `mandatory' with a score of (21/4) indicates that 21 respondents out of 25 voted it essential, the remaining 4 voting for one of the other three possibilities. An item rated `desirable' with a score of (12/9) indicates that 12 out of 21 respondents voted it either `essential 'or ` if possible', while the remaining 9 voted it as either ` no opinion' or `don't want'.
With respect to the header features, respondents rated highly only a small number of those available. The summed scores obtained were as follows:
Within the Encoding description: only the Sampling description (18/4) and Project description (15/7) were rated as `desirable', no other features being rated higher.
Within the Primary Data, only Text body (13/9) and Gap (11/11) were rated `mandatory', while the following features were all rated as `desirable':
No other primary data feature was seen as being of high priority.
Few respondents considered morphosyntactic mark-up of any kind as essential. For those who did, the summation procedure outlined above gave the following results:
These responses seem a little capricious: it is difficult to imagine why marking a numeral for example should be more important than marking a noun. A closer examination suggests that several of the items rated `mandatory' here are only marginally so, with very close or equivalent scores to those rated `desirable' by our procedure. We conclude that for this particular type of mark-up the choice between` essential' and `if possible' has little significance.
As with morphosyntactic features, only a minority of respondents expressed a requirement for markup of syntactic features. Of those who did, the following features were all rated as `desirable':