Thesis

If you would like to write a student research project, bachelor's or master's thesis in the field I represent, please send me an e-mail (opens your email program). Ideally, the work should have something to do with our/my current research projects or research topics. You are therefore welcome to ask me or my colleagues about current topics.

Time and again I receive inquiries that go something like this: "Dear Professor, can you supervise my compulsory internship at company XY?".

In most cases, these inquiries refer to the practical project which is described, for example, in §10.10 of the current Bachelor's examination regulations.

It is understandable that you want to have this point completed during the course of your studies - but it is also damn clumsy to simply tick this point off! Yes, the practical project is ungraded (there is only pass and fail) and is pretty low-hanging fruit: In order to fail it, you have to be pretty stupid and either not keep to the required times or not submit a sufficient practical report.

But: The practical project is usually the last academic achievement before the Bachelor's thesis and should, in my opinion, be used to prepare the Bachelor's thesis (BA). In your BA, you will be working on a new task, which always requires familiarization. The maximum processing time for the BA is 3 months, which begins when you register for the BA; you must therefore have familiarized yourself with the topic before you register for the BA. It is therefore obvious and practicable to use the practical project to familiarize yourself with the subject of the BA. You will need the same amount of time as you would otherwise, but in addition if you consider the practical project separately from the BA.

So: Don't look for a practical project, instead look for a Bachelor's thesis - and then in most cases you can define the practical project beforehand in a relaxed manner. This usually works just as well for external theses as it does for internal ones: you can also complete the practical project (which prepares you for a BA) internally at Ostfalia, for example in my working group.

So if you are looking for a thesis in the area I represent, please get in touch before you look for a practical project so that we can try to combine the two!

So that I don't always have to tell everyone the same thing, I'm collecting things and questions that come up from time to time here.

Content

As a scientific paper, a thesis should always include the following three important points:

1. Related work

In general, you should always take a look: What have others already done in this area? If there is nothing in particular, you can also broaden the focus. In principle, you could also include the prerequisites or circumstances here - i.e. what is the status quo before your work. This can then lead more or less logically to the next point, namely

2. Own implementation

This is about describing what you design, programme, implement, ... the actual core of your work. Ideally following / inferring from the previous section: "This doesn't exist yet, and that's why I'm doing it now in my work". Describe what you did, how you did it and why you made the individual design decisions.

3. Evaluation

Evaluation is about finding out: How well does your implementation work? Quantitatively, i.e. with recorded measured values. Ideally in comparison to the previous status quo. The aim is to show: "This work has made the world a better place, here and there by x percent". Sometimes this is not possible without further ado, as there is (still) nothing to compare with. The relevant metrics for evaluation often only emerge during implementation. You should only think about this in good time.

This three points can certainly not be applied 100% to every project - but they should provide a rough guideline. Also, this three points cannot necessarily be "matched" to individual chapters of the thesis. Sometimes it can make sense to describe several aspects of the implementation in several chapters; sometimes it makes sense to evaluate sub-steps directly.

Structure / chapters

The structure and number of chapters depends on the type of work, the task and your own preferences. Sometimes chapters shift in the course of the work - this is completely normal. Cheers to electronic data processing! :-)

At the beginning (after the table of contents) there is usually an introduction and motivation. This is preceded by an abstract of the entire paper (0.25 to 1 page)

At the end of each paper there is a conclusion, possibly followed by an outlook (future work), which then states how to proceed or what points are still open.

The bibliography comes at the end. A list of figures and tables is usually not necessary.

Tip: Write the introduction last! After the abstract - then you know exactly what you have done (and why) and how "the story begins". Otherwise you would have to adapt the introduction every time you make a (re)decision during the course of the work (often things don't work out the way you planned).

General structure

The paper is divided into the following sections:

1. Title page
2. Declaration of independent preparation - according to BPO
3. Abstract (German) & Abstract (English) - 1/4 to 1/1 page each
4. Table of contents
5. Text chapters of the thesis
6. Bibliography
7. Attachments

Created software as file attachment

If software and/or code (fragments) have been created during your thesis, it does not make sense to include these completely in the text section of the thesis. For some parts (e.g. interface definitions), it may make sense to include them in the source text within the individual chapters; as a rule, the resulting (and hopefully well-commented) code belongs in the appendix. To do this, please create a file (zip/tar/etc.) in which your entire commented source code (including other files, readme, configuration for IDE, tex files and images of the development, etc.) is stored.

Please refer to this file in the "Attachments" section. Ideally, you should also show the folder structure (usually the top level is sufficient) and briefly describe (keywords/sentence) what is hidden behind the respective directories.

Volume

The volume of the paper depends on the type of work (student research project < bachelor's thesis < master's thesis), but the scope of the paper should be appropriate to the content. Do not inflate your paper unnecessarily - I have to read it all. I cannot and will not set a fixed number of pages: I generally assess the content and not the number of words it takes someone to describe the content. I have read very good Master's theses with 35 pages and quite bad student research projects with more than 100 pages.

Form

A thesis can be a good reason to familiarise yourself with LaTeX and BibTeX or BibLaTex! I expect the use of LaTeX (including the templates provided) for thesis I supervise! I have created a short video tutorial (external link, opens in a new window) on this.

• You can find a LaTeX template here.
• You can simply import it into ShareLaTeX.

General format (if you are not yet using the LaTeX template)

• DIN A4 paper size
• Margins not less than 2 cm
• Text set in justified text
• Chapters numbered
• Page numbers on each page

In general, papers no longer need to be printed - assume a purely electronic submission as PDF!

References

Please use the IEEE format (external link, opens in a new window) for references and citations (numbered references, sorted by occurrence in the text). Ideally, you should use BibTeX (external link, opens in a new window) or BibLaTex for source management. If you prefer graphical source management, you can use JabRef (external link, opens in a new window), for example.

Make an effort to find and use scientific primary sources; these can be found, for example, in Google Scholar (external link, opens in a new window), IEEE Xplore (external link, opens in a new window) or in the ACM Digital Libary (external link, opens in a new window).

The individual entries in the bibliography should be complete and the referenced work must be clearly identifiable, including the names of the authors, title of the publication, publisher (with location if applicable) and date of publication, and in the case of conference papers, the name, location and date of the conference. If possible, your sources should have a Digital Object Identifier(DOI (external link, opens in a new window)). With the vast majority of academic publishers, you can download a BibTeX "snippet" containing all relevant information directly from the website of the primary source.

For web-only sources, try to include as much information as possible (author, page title or article title, publication date, date of access) - a stand-alone link is not sufficient for a literature source, but would be possible in footnotes.

If you use non-scientific sources, such as news or blog posts, make sure that the source is reputable. For example, I consider all media from the Axel Springer Group (Bild, Welt, etc.) to be tendentious (and therefore dubious) and have no place in a scientific paper.

Wikipedia (external link, opens in a new window) can provide a good introduction to the topic for many things. For more detailed information and citable sources, please refer to the references given there, which often also refer to the primary scientific literature.

Communication and agreements

The scope of the work and the exact task must be clarified before the work begins. The first step is always communication with the thesis supervisor. In the course of preparation, the student prepares a short exposé in which the tasks and objectives of the thesis are presented. It is advisable to deal with the above-mentioned content points (related work, implementation and evaluation) in the synopsis.

The Ostfalia e-mail account should always be used for e-mail communication.

Exposé

You should prepare an exposé before beginning your thesis. The aim of the exposé is to ensure that everyone involved knows what you will be working on in your thesis - and how. To this end, the exposé should outline all the points listed above under content. This means that your exposé should answer the following questions:

• Problem/question: what is the problem you will address in your thesis?
• Categorisation / context: In what context does your work take place? What have others already contributed to it (see also: Related Work)? What are you building on?
• Approach / Methods: How do you plan to implement your solution? What aids and tools do you use? (see also: Implementation)
• Evaluation: How do you plan to evaluate your implementation? What are useful metrics and how do you plan to assess the quality of your implementation? (see also: Evaluation)
• Timing: How do you plan the rough sequence of your work? What are any "critical paths", which things are interdependent?

A frequently practised procedure is that the exposé already shows the structure of the later elaboration (incl. bibliography) and is then expanded by you into a "finished work" in the course of processing.

The exposé is usually created in dialogue and in several iterations: You write a first draft, we talk about it, you revise the draft until everyone agrees... :-)

Confidentiality and embargoes

Professors are generally financed from public funds (i.e. taxpayers' money). In my opinion, as much as possible should therefore flow back to society - in our professional field, this usually happens through the generation of generally and freely accessible (i.e. publicised) knowledge. A general confidentiality agreement that withholds the results of certain work from the public as a whole and in the long term should therefore be viewed critically: My consulting and advisory services would, after all, only benefit a specific company as a result of a non-disclosure agreement. It is also difficult for me to understand why tasks whose publication could harm the company have to be dealt with in student theses of all things. With a little creativity, the vast majority of results can be abstracted to such an extent that no company-critical data is made public.

I am therefore only in favour of a non-disclosure agreement in very well-founded exceptions. Such exceptions could also be justified by an existing or imminent (research) co-operation.

Non-disclosure agreements that "block" the publication of results for a defined period of time are generally less critical, but here too it must be justified why (and for what period of time) a block is necessary.

Presentation and colloquium

The presentation of the results of the work is an independent achievement and should be planned as such. As a guideline, assume 20-25 minutes of pure presentation for a student research project and 25-30 minutes for a Bachelor's thesis. You will (hopefully) not be able to present your entire work of the past 3-6 months in 20-30 minutes, which is why you should not simply describe your work chronologically in the presentation. Instead, put yourself in your audience's shoes when preparing the presentation:

• What are the most important results/findings you want to present?
• What prior knowledge does your audience need to understand the results and how can you convey this briefly and concisely?

So build your presentation "from the back" - starting with what you want to convey. If you have implemented something (in hardware and/or software), demonstrate it live during the presentation.

After the presentation, you will be asked questions about the content of your work by the first and second examiners and the audience present. If you have written an internal thesis at our research group, the audience will consist of interested people from the faculty. If you have written the thesis in collaboration with a company, the colloquium will usually take place there - preferably with colleagues on site to get as much input as possible from there.

DOs and DONTs

No headings without text

New chapters in particular should be introduced briefly, even if only with a few sentences.

Labelling images and tables

Images have captions, tables have headings - both are numbered. If images come from a source (this usually applies to all images not created by the author), this source should be stated in the caption - ideally simply with a reference to the literature source: "Figure 23: The course of things (from [12])."

No unreferenced figures and tables

Neither tables nor graphics should be unreferenced. There should be at least one reference and a description of the content in the text. In practice, this means that there should be a reference to each image somewhere in your text: "Figure 47 shows that the actual course largely corresponds to the predicted course; the deviations in the upper part of the scale are due to inadequacies in the measurement setup, which are discussed in the following chapter...".

This list will (hopefully) be constantly expanded.