First, terms: Computing, Computer Science, Information Technology, Information and Communication Technology, students, and even professionals often interfere.

According to ACM Computing Curricula 2005, defines five sub-disciplines of the computing field: Computer Science, Computer Engineering, Information Systems, Information Technology, and Software Engineering. Following a roughly 5-10 year cycle, the ACM and IEEE Computer Society jointly sponsor the development of a Computing Curricula volume on Computer Science, CS2013 Ironman v0.8 draft, Jan. 2013. In addition, in England, the Royal Society sponsored the manuscript “Shutdown or Restart: The way forward for computing in UK schools,” Nov. 2012.

Therefore the answer for "What are the bests methods for teaching Computer Sciences" is too complex! You need to be a little more precise...

About teaching programming language... In my opinion, programming languages are tools with which we solve problems, tasks, requests, etc. that are placed within our curriculum, regardless of whether it is CS or IT curriculum. Therefore, you have to decide which approach procedural or object-oriented want to use again in order that you set by your curriculum, i.e. what kind of professionals want to educate.

That is all for beginning of discussion…

I would give a very brief response - LEARNING BY DOING! would be the best method.

Regards

I'm not sure I entirely agree with the learning-by-doing thing (not entirely, I stress that - I think it is impossible to learn a large amount of programming without that component to it).

As a lecturer in mathematics, computing and technology, my view and experience of teaching programing is that teaching needs to begin with understanding the notion of Marr's three levels of understanding, i.e. goal, method, implementation.

I will comment more later, and be more practical-do-y about my next thoughts but not got time now.

Ah, now second thing - big debate taking place on this in my local 'academic community', and it would be good to discuss it further actually. We have been discussing:

a) The value of multi-modal education, e.g. using YouTube to support what we provide in material, and all the stuff that MIT, etc. offer

b) The value of high level analogy, such as talking about clocks in some detail, and other very familiar stuff, then basically just substituting elements so, for instance for the clocks analogy, the notions of discrete increments for a continuous variable (e.g. time) or as a 'count register' (from assembly)

c) The use of a reduced set natural language, a variant of 'structured English', perhaps, that bridges the gap between natural, high and low level language.

A work in progress!

Good points, Ljubomir!

I'm intrigued. I think I would agree that domain-agnostic knowledge, skills and attitudes are insufficient to fully address the demands of education, teaching and learning within 'computing' (as you say, imprecisely defined, and heterogeneous in nature as the question is put here). So I definitely think a hard perimeter (to think of it in Venn diagram terms) needs to be drawn around that 'blob' of computing sub-disciplines to say what is in and what is out when we are trying to answer this question.

I disagree with you then on only one point, and that is whether these individual sub-disciplines nested within the whole (which you express and reference excellently above, and is very helpful) have some 'extra' degree of domain-specific learning that need further resolution within that overall perimeter.

If so, I would actually argue that as a 'sub-discipline' of education (i.e. what is domain-specific to the teaching and learning of that part of the blob) may actually be part of something else entirely. The most obvious, at first glance, might be to group, e.g. number theory, Boolean logic, statistics, permutations, optimisation, etc. as being part of the special knowledge, skills and attitudes used in maths teaching rather than a super-specialism of a bigger set of 'computing teaching'.

If that's so, I would start to wonder how easy it is to actually specify the 'domain' at all. But that's a bigger philosophical issue.

Essentially, what I'm saying, Ljubomir - and I'd genuinely like to hear your view is: is it really deeply relevant to the question how we define the subject, as opposed to how we define the knowledge, skills and attitudes we need to teach and we use to teach.

Finally, my comments above specifically are focussed on the teaching of 'programming'. I am not sure if you think the notion of programming needs further division and analysis, or if it is 'flat' and homogeneous as a field/concept?

I agree with all of you, and i can add ( adding to the idea of Theodora Issa) i believe to proverb "Listen And Forget, Write and Remember, Practice and Understand", where the practice can facilitate understanding programming. In the process of teaching this area, i prefer to start by giving practical examples , then pass to theoretical courses. May be that you agree with this ideas.

I like to keep it simple when I've taught introductory classes. I generally start with algorithms, written in natural language. The one I've found most useful (for demonstrative purposes) is 'tell a 5yr old how to make a peanut butter and jelly sandwhich'. To illustrate, I will bring in a knife, a loaf of sliced bread, a jar of jelly, and a jar of peanut butter - and then use the descriptions (algorithms) proposed by the students - following them exactly, without deviation. 90%+ of the descriptions will fail, due to assumptions that humans make about behavior. The point of the exercise is to show how much detail an algorithm must have in order to be successfully executed.

I try to stay away from specific programming languages because I find that once a student starts writing code first (as opposed to algorithms), the limitations imposed by the language will directly affect the way the student solves the problem.

An important factor when teaching (younger) students is to make them driven and motivated. Some are motivated by working on a project that they consider"exciting", others are simply motivated by presentation. Especially the last category should be given regular grades. I think it of high importance to hold assessments (that count) regularly and always give feedback based on individual students work.

Personally I have best results with a mix of theory (short classes) and practical work.

While this all may seem logical I still see many examples where teachers wait until the end of the coarse, give an exam and don't give any feedback. Don't forget those kids live in the information age, where there is a lot of input and distraction. Sometimes they just need a bad degree at the start of the course to let them feel the urgency of really putting some effort into learning to program. Good programming has to be learned by spending a lot of time "doing" it and receiving feedback on your work. Sometimes feedback can also be given by peers or other students. The more feedback, the better!

If you want to learn the students C or C++, Python might make a good starter language. Since its easy to learn and has a consistent syntax.

At my university, students frequently complain that courses are to theoretical. Therefore I tried and I had some noticeable results by taking practical problems into “real world”. So, a balanced combination between theoretical courses and practical courses (mostly based on real life problem examples) seemed to be a good solution.

In addition, constant monitoring of student’s evolution is more than welcomed. If the student fails one topic, the teacher should not punish him for that by instantly giving a bad partial degree that could negatively affect student’s final degree. Instead, the teacher should explain what went wrong and should check, in the near future, if the student corrected that aspect. In computer programming this cross check is quite easy to do.

Out-of-the class counseling should also be continuous. Emails, forums or e-learning systems can be used. This could encourage the student to work by himself and try to improve the in-class performance.

I am a student, and I prefer the practical ways.

But I would love to give a chance to solve the problem before my doctor gives the solution to me.

Thanks for the discussion so far - indeed, you cannot learn something through practice if you have no idea about it!? I would concur with Boudhir "Listen And Forget, Write and Remember, Practice and Understand",

Regards

Theodora Issa