This is a list for some last checks before submitting a large document (say a journal article, master thesis or even PhD thesis). Created by Pascal Bercher (https://bercher.net) Update history: - 14.8.2022 (not clear, probably the first 'stable' version) - 6.12.2022 (added hints for grant proposals) Advice for any larger works =========================== - Make sure that graphics are not unnecessarily large MB-wise, nobody wants to download a 15 MB thesis (or even larger) if not required. 2 to 3 MB max would be best. Maybe 5 MB if it's a dissertation. More is just annoying. - Make sure that all "person" names which are written "by hand" (rather than being inserted via \citeauthor{} etc.) are correctly written. Better yet: Never do that in the first place! \citeauthor{} / \shortcite{} etc. should *always* be chosen! (These commands should be part of most standard authorkits.) - References: the ususal! I.e., consistency (like spelling or abbreviating conferences, but even names: are the full names always given or just sometimes?), completeness, correctness, spelling, and capitalization of acronyms like "HTN" or "HDDL". Take a careful look also at the output, not just at the bibtex -- sometimes one sees more there. E.g., a single wrongly placed or missing comma can screw the author list completely thus fusing authors into a single long name; this might be more visible in the final product rather than in the bibtex file. - Find usages of "his [...]" --> "his or her [...]". (Or her or his -- as you prefer; I mostly use the latter, but that's just taste.) Similarly: "he" --> "he or she" - Are all headlines (chapter, sections, subsections, etc.) correctly and consistently capitalized? (There are online tools for that!) - Do you ever use just a single section/subsection? (In this case it might just be removed) - Check all error messages and warnings in the LaTeX compilation (open the .log file and investigate each "error" and "warning"). Note: overfull boxes can be identified via compiling the document with the optional argument [draft] in the documentclass. Those overfull boxes will then visually be shown via black boxes, making it easy to be identified. - If you have the time, learn the rules regarding when to use "that" and when to use "which", in particular in combination with commas. It's about "nonrestrictive" phrases. It's not completely trivial: There are TONS of web pages on it and one can invest hours. They are well-invested, though, but it requires a few hours of "free mind" to fully understand the rules and read up enough examples to fully get it. Well, and it takes again hours to check the entire thesis/document for the correct usage (though you can just search for all usages of "which" and "that") -- but it might be worth it if you want to use only correct English. - British English (BE) versus American English (AE). Whatever you choose, be consistent. Similarly: Be consistent about the Oxford comma. Either always use it, or never. (Personally I always use the (BE) Oxford comma despite writing with AE otherwise. I just like AE, whereas I find the Oxford comma very useful as it sometimes clarifies meaning.) - Use a spell checker and carefully go through each single source file to search for the red-underlined (potential) errors. Even if spell checking is activated one not always takes the time to go through everything... MS Word is really amazing in finding errors! It is REALLY worth exporting your PDF to MS word and just scrolling though that document! (Do it! :)) (I am not a Windows user, but for important documents like a dissertation this extra step is definitely worth it.) - Typical spelling errors: * "quiet" vs. "quite" (probably "quiet" should never be used in our scientific/math context) * "here" vs. "hear" (probably "hear" should never be used in our scientific/math context) * "there" vs. "their" * typical word repetitions: ~ "the the" ~ "for for" ~ "is is" (use can search for " is is" instead to get less false positives so that "this is" is not captured when searching/grepping) ~ "it it" ~ "in in" ~ "on on" ~ "a a" ~ "to to" ~ "an an" ~ "with with" ~ "and and" ~ "or or" ~ "be be" Since this was only implied, let me be explicit: Take some time and do a search for all the above. Either use full-text search in the PDF or better yet use the "grep" command. If you have too much time, you can write a script that searches for it. If you do, send it to me and I'll add it here. :) - Maybe take the time searching for all "its" and "it's". Though it is completely trivial, sometimes one just mistypes and adds a wrong apostrophe, similarly to sometimes writing quiet instead of quite or their instead of there. It's just at typo that sometimes happens. :( So check whether they are all right. It would be really embarrassing to write "it's" where "its" would be correct. :) (I really find it annoying when that happens to me, and yet it does sometimes! - If the "first person" is used (which is more likely in theses), make sure "I vs. me" is used *correctly*. According to Stackoverflow (https://ell.stackexchange.com/questions/10515/should-i-write-x-and-i-x-and-me-i-and-x-or-me-and-x-in-a-conjoined-ob), we get the following rule: "I corresponds to we and me corresponds to us" (this is how you can decide whether "me" or "I" should be chosen.) - When reporting numbers: Did you use always the dot and comma separators in the right way? Note that how it's used can be different for different countries, but in English texts, a comma should be used for every third digit, whereas a dot indicates fractions. E.g., * write "1,000" instead of "1000". * 3.14 means the integer 3 plus the fraction 0.14. * thus, 1,042.73 means 1,042 + 0.73 (I'm emphasizing that for all Europeans in particular, since "we" have it flipped around! I.e., "we" would write 3,14 -- but in English it must be 3.14.) - Think about who (definitely or potentially, including external witnesses or people in Award committees) is reading, marking, or nominating your work. Got a bunch of names? Take time and check their work (DBLP, scholar, etc.), including their most recent ones (for which you might need their private websites or the list of accepted papers of recent conferences, as DBLP and scholar need some time to update). If it's related to your work, you should cite it. Don't get me wrong: You should not cite them *because* they are reading/marking/awarding your work, but because -- by assumption -- their work is related and relevant to your own. So in an ideal world (i.e., if you had done your job right) they should have already been positioned adequately int he first place! But thinking about who will read/mark/etc. your work might help you identify other works which simply were not on your radar. This applies to conference papers: Think about potential reviewers. This might help you find new related work that you just have not known about yet or simply didn't think about yet since you got caught up too much in describing your own work -- yet positioning your work in the landscape is essential. - The advice given in this document is a checklist that I use for "large documents" to make sure I didn't miss anything before submitting the last version. However, if you are new to all this (say you are an undergraduate or young PhD student), I strongly recommend to read my template I created for research projects at the ANU. Note that this template isn't just a LaTeX template, but it's pre-compiled PDF contains tons of advice about writing reports/theses. It thus goes far beyond this document here (though it doesn't subsume this document here, in fact it complements it). Still, give it a try, I've only received positive feedback about it so far. :) You find the git here: https://gitlab.cecs.anu.edu.au/u1092535/project-latex-template However, you'll have to be an ANU member (student or staff) to access. If you are not, you can access the pre-compiled PDF from here: https://bercher.net/data/teaching/ANU-researchProject-template.pdf - AFTER EVERYTHING ELSE IS COMPLETELY FINAL: Yes, I am shouting here! :) Make sure that's the very last step! Check the pagebreaks. E.g., a page should not just contain just two lines, that's just really, really ugly. Use \pagebreak (and maybe \flushbottom) or just shorten down text to prevent this. Even if you personally do not care (maybe we just don't share the same sense for esthetics), keep in mind that others might. So make sure it looks good. Advice for grant proposals ========================== - [Abbreviations] In grant proposals, the level of language should be more formal. At least it *might* help in some cases and certainly never does any harm if you do it. The same does of course apply to any scientific work, but for grant proposals it might be more important. One way how more formality can be achieved is by not using abbreviations (something that is enfored even by most if not all edited journals). So, below I list all abbreviations that you could/should grep for. Use -i to ignore capitalization. * it's --> it is * that's --> that is * wasn't --> was not * won't --> would not * didn't --> did not * hasn't --> has not * I'd --> I would * we'd --> we would * I've --> I have * we've --> we have - [AE vs. BE] Depending on your funding agency, either American English (AE) or British English (BE) might be required. Find this out early! Even if none of them is demanded, you should make sure that you stay consistent! In case your LaTeX editor does not fully support AE vs. BE, grepping will help. Here is just a short list of words or patterns that you might look for: * modeling (AE) vs. modelling (BE) * anyting with z inside (AE) vs. anyting with s inside (BE), e.g., optimization/formalization (AE) vs. optimisation/formalisation (BE) I recommend to simply grep for the letter z (in case you usually write AE but not need to write BE), and then search/replace any hit.