Some Tips and Tricks for Surviving the ASE Seminar

Transcript

1. software evolution & architecture lab Dr. Philipp Leitner University of Zurich, Switzerland Seminar: Advanced Software Engineering Some Remarks, Tips, and Tricks for Surviving the Seminar
2. Overview • Finding, Accessing, and Evaluating Papers • Writing your Paper • Conducting a Review
3. Types of Literature What’s academic literature? • Journals • Conference proceedings • Workshop proceedings • (Some) books Rule: • Needs to be peer-reviewed (no Wikipedia / blog posts) • Should be the original source of a claim / result
4. How to Find Literature Three major ways: • Searching in digital libraries • Actively looking through relevant venues • Reference crawling
5. Digital Libraries in CS • Google Scholar: https://scholar.google.ch • IEEEXplorer: http://ieeexplore.ieee.org/Xplore/home.jsp • ACM DL: http://dl.acm.org • Good: gives you access to lots of literature, easy keyword search, ranks papers by citations (may be good or bad) • Bad: hard to find something useful if you don’t know yet what you are looking for, much crap can be found, tends to prefer older sources
6. Some Tips • Build up a collection of keywords that you use to search (and re- execute your search periodically) • Have a combination of generic and specific keywords • When you read papers, update your keyword list with the terminology used in the paper • Don’t assume that different authors use the same terms for the same concepts
7. Venues • Actively look through previous iterations of relevant conferences and journals • Good index for CS: http://dblp.uni-trier.de • Again: build up an index of relevant venues as you read • Check where papers you already found have been published • Good: does not overwhelm you so much, more quality control, newer literature, more “explorative” (easier to find something you didn’t think of so far) • Bad: more cumbersome, you need to know what the relevant venues are
8. Ranking Venues • Finding out whether a given journal or conference is good or bad is a science in itself • There are a few sanity checks: • Check whether it’s indexed by DBLP (most good venues are) • Check the CORE ranking (or other rankings) • Conferences: http://portal.core.edu.au/conf-ranks/ • Journals: http://portal.core.edu.au/jnl-ranks/ • Google the acceptance rate (< 25% is quite competitive) • All of those are only rough (biblio-)metrics. Trust your gut!
9. Reference Crawling • Finding other relevant work by going over the references of the papers you already read • Backwards — just look over the references of the paper • Forwards — use Google Scholar’s cited by feature • Good: comprehensive and low-effort way to observe a body of literature • Bad: easy to get stuck in a specific community; won’t find the lesser-known “pearls”
10. Using GS For Forward Reference Crawling
11. Accessing Literature • Publications can be either open access or behind a paywall • Publications behind a paywall can usually be accessed freely from the UZH or ETH network • Either physically sit at UZH or use the VPN services • Eduroam will not work :( • For some important venues, UZH does not have a subscription. • Try googling the paper. Many “preprint” versions are available freely on the Web. • Otherwise: check ETH, check ResearchGate (http:// researchgate.net/), or mail the authors
12. Paywall Behind a paywall Same URL from within UZH
13. Writing your Paper • Length: 12 - 15 pages • Written in Latex (or, if you really want, in Word) using the LNCS style • http://www.springer.com/computer/lncs? SGWID=0-164-6-793341-0 • English
14. Makings of a Strong Seminar Paper • Seminar paper does not require original research • —> you are not expected to collect your own data • However, you should collect, integrate, summarize, and compare existing work • The paper is then a synopsis or survey of the existing works
15. Bad Outline • Section 1 - Introduction • Section 2 - Summary of Paper 1 • Section 3 - Summary of Paper 2 • … • Section n - Conclusions
16. Much Better Outline • Section 1 - Introduction • Section 2 - Overview of Research • Section 3 - Synopsis of Topical Cluster 1 • Section 4 - Synopsis of Topical Cluster 2 • … • Section n-1 - Discussion and Open Issues • Section n - Conclusions
17. Comparison Table • A lot of your “brain power” will go into finding the best way to group and discuss your area of research • A good way to emphasize this is to come up with a comparison table (or figure), where you sketch which papers address which aspects of your field • This table may have “holes” • —> Identified open research questions
18. Writing Style • Reflect how the good papers you are reading are written: • Avoid unnecessary and unwarranted superlatives • Avoid claims that are not supported by data • Avoid colloquial language • Prefer passive over active voice • If you use figures or screenshots: • Make sure they look reasonably professional and visually pleasing • Check the resolution (esp. of screenshots)
19. Academic Honesty / Plagiarism Rules of academic honesty apply! • No plagiarism! • You can’t copy anything. All text / figures need to be your own. • Exception: direct quotations (use very sparingly, visually distinct from text, immediately followed by citation) • No misrepresentation! • Never claim something that is not actually supported by your references. • Make clear who is the source of what! • Don’t just cite a source randomly on a page and assume that the reader infers that the rest of this part of the text is based on this source.
20. Conducting your Reviews Reviews should comment on: Technical quality Logical structure Presentation Style References Each category should be graded on an A to D scale: A: An excellent work. B: A good work with just a couple of small weaknesses. C: An average work with clear weaknesses. D: Insufficient work with many substantial weaknesses.
21. Good Reviews A good review has a fisheye view • Comments on positive and negative aspects • Comments on macro issues (selection of literature, general approach, etc.) … • … but also gives actionable, detailed comments • Does not only criticise, but also gives concrete suggestions for improvement • You can comment on spelling or grammar, but don’t make your review primarily about that • Never make it personal. Stay professional at all costs.
22. Bad Reviews “Great paper. I have no comments.” “Terrible paper. I have no comments.” “Here are some typos I have found.” “Rather than doing A and B I would have looked at C and D.” “I am personally offended that you are doing A, and I think you are an idiot for it.”
23. Structure of a Good Review < Short summary of paper > < List of positive aspects > < List of negative aspects > < List of other things that should be mentioned > < Short conclusions >
24. “I don’t have comments.” If you carefully read a 10+ page paper, some comments (good or bad, usually both) should come to mind. If you literally have nothing to say about the paper, I recommend reading the paper again.
25. Further Timeline
26. Final Comments http://phdcomics.com/comics.php