Allocation
The allocation process usually results in a many-to-one mapping from students to supervisors; associated with each link is a defined project. There are many ways of conducting this process, but it should be self-evident that as numbers increase, so does complexity. This brings with it concomitant opportunities for error and dissatisfaction.
Allocation is a many-faceted process - it may be construed as:
The consequences of getting allocation wrong, either wholesale or just for unlucky individuals, are usually very bad, particularly for the students. Misallocation (actual or perceived) can be extraordinarily de-motivating, and the consequences of this, given the characteristic weight of projects, are severe. Misallocation can lead to , inter alia, misunderstanding of the project, or supervisor, or dislike of the supervisor (or, unprofessionally, supervisor's dislike of the student). The day the allocation list is published - usually an open and public mechanism, well advertised in advance - there is often much distress and complaint. This is probably unavoidable, but it is important that students perceive the allocation process as largely "fair", since this will dilute resentment. There is often an understandable, and sometime accurate, suspicion that staff will cherry pick their "favourites", and all steps should be taken to prevent this happening, and demonstrate publicly that it doesn't happen.
It is natural behaviour among many students, given free will, to queue up in front of the "popular" members of staff, howsoever judged: this "beauty competition" is often going to be at variance with academic best advice. Since it is common practice to spread supervision load across the whole staff, it is highly likely that some supervisors will be quite unknown in name and face to students, which is clearly intimidating. This is especially the case when non-academic staff (Research Fellows, or support staff), or staff in other departments are drafted in to supervise [See 2.12, Pulling in the Bodies and Graduate Students as Supervisors]. This important fact is often forgotten.
Equally, many staff will be unaware what drives student choice; it might be expected that the topic, (especially for research focussed staff) is the prime motivator and it will not be perceived that impressions received during teaching experiences in early years count very heavily. At the same time, staff must perceive the allocation as "fair" - for example, allocating five students who are all expected to graduate with third-class degrees to one individual is unlikely to be received enthusiastically except by the most devoted teachers.
In most instances, the allocation process is managed by an individual - it is difficult to see an organisationally acceptable way of distributing the core of this task. Like may administrative jobs, this is onerous and will not be popular. The choice of this individual can be critical, since whoever is chosen must have an insight into broader staff loading issues, and must have sufficient weight (or kudos) to be able to persuade or coerce colleagues into doing things that they may well not want to. Normally, this individual will have a senior member of staff as back up for the occasions that issues of seniority interfere with preferred allocations. These can be delicate managerial issues, dependent on local personalities and practices, for which it is difficult to plan.
What a "reasonable" load may be is likely to be a matter for local resource allocation [see 2.12 Staff Deployment], but experience suggests that if one individual is simultaneously supervising more than six independent projects, the resulting context-switching begins to disadvantage the students. Some mechanisms exist for streamlining the supervisory task, but when the system starts to creak because of shortage of supervisor time, it is usually time to consider a major rethink.
Special cases may arise in the event of, for example, MSc projects conducted over the summer. These provoke a special tension in that they usually require supervisors to be specialist academic staff (precluding the use of peripheral staff ) at the precise time they are accustomed to uninterrupted research and conference time.
Do not forget that the whole allocation process usually has to be repeated, at least in part, in respect of second markers, or shadow supervisors [see 2.9(iv) Moderator and Supervisor plus Another and 2.11 Co-ordination Structures for Supervision ]. Often this is overlooked and done in an unseemly hurry; this is a process that requires matching of staff to staff, with all the concomitant problems of local personality issues.
Especially, remember that you must have a contingency strategy. In less than 1% of cases, something will go wrong and a student/supervisor partnership will have to be changed. You must have a strategy so that you can react in time [see Section 5, Stuff Happens,]. The sub-issue of allocation of students into teams and groups is considered in Section 8 "Team/Group projects", and particularly 8.2 and 8.3.
Much the most commonly experienced problem is the mismatch of staff to student "demand" areas or, conversely, the mismatch of student desires to topics offered by staff. The latter here may be at the coarse level ("I don't want to do AI") or finer grained, if the local practice is to publish lists of precise project possibilities ("I don't want to do a project on simulated annealing"). This is another aspect of the difficult cost- minimisation task faced by the project co-ordinator or manager, and is intertwined with local expectations: Do students expect to be able to work in their chosen area? Do staff expect only to have to supervise in their research or cognate areas?
In both circumstances, the issue of first loyalty must be spelled out - that is, the project is being conducted in pursuit of a degree and may result in a useful product, not the other way around.
Orthogonally, problems often surround strong students taking on projects which do not challenge them, or weak students attempting something just too difficult [see 2.1, Weighted Topics].
When it is possible to gauge in advance how tough a project is likely to be, then some care can be taken to prevent this - it is not, of course, always the case that this will be possible. Another aspect of this issue is the type of project - many, and often the majority, are of the design-and-build kind preferred by accreditors, but examples of pure research projects are not uncommon, and many examples exist of very worthwhile activity conducted under the project head whose connection with computer science may be arguable. Some especial care needs to be taken in matching students to things which are out of the ordinary.
Quality assurance in supervision (extracting, or hoping for, uniform treatment from the staff cohort) is a known problem [see section 6, Supervision]. If it is practice to use non-academic staff as supervisors this problem becomes more serious since often such people are unaware of the mechanisms mainstream teaching staff take for granted - this is particularly the case if they are pursuing "pet" projects of their own, with the danger of the project becoming solely product focused. This is, in fact, a special case of the general problem of non-uniform experience and expectations among supervisors.