Ask HN: Help Us Build a Tool to Simplify Paper Writing for Researchers | Hacker News Reader

Ask HN: Help Us Build a Tool to Simplify Paper Writing for Researchers

4 pointskzardar8 comments

Hi HackerNews community!

My name is Zardar, a biomedical engineer turned software engineer and together with my buddy Ahmed - who is a seasoned biomedical researcher, we're on a mission to build a tool that will help the process of writing research papers and journals.

We'd like to better understand the pain points and challenges researchers face when writing papers. We're looking to speak with people from all levels of experience (undergrad, post-grad and veteran researchers) to learn about their workflow.

Your time is valuable to us, we are happy to compensate you - whatever you think is fair - for 30 to 60 minutes of your time. We're looking to deep-dive into what has worked for you, what hasn't and where your pain points lie.

If you're interested in sharing your experience and helping us develop a solution, we'd love to hear from you.

View on HN

By far my biggest pain point is Microsoft Word. I've found that there are existing solutions for most steps in the research process (Git for version control, Rmarkdown/Quarto/Jupyter/Orgmode for integrating the code and the analyses, Zotero for citations, etc), but almost every research project that I've been a part of has eventually required the manuscript to end up in Word due to either journal requirements or some collaborators refusing to use other tools. At that point all the advantages of the other tools go out the window. It's easy for people to change tables or otherwise break the linkage between the manuscript and the analysis code, and you're stuck spending a ton of time auditing the paper and fiddling with the appearance of tables. Furthermore, once you're in Word you have to manually move any changes in the analysis into the manuscript rather than being able to rerun the code and run the output through Pandoc. I would love a tool that lets me work with collaborators in Word (including track changes) without losing the advantages of my preferred workflow.

I haven't used it enough to have an opinion about using it, but you can build a word template that you hand to pandoc when it is converting to word, and it will stuff all your content into the matching styles in the template. The table templates in word are annoying to use, but you can put nearly all of the formatting into the style and not have to mess with individual tables.

Mostly just throwing stuff out there in case you hadn't looked at it.

Thanks! I actually have not spent much time looking into that and will see if it helps.

First of all, thank you for sharing! Much appreciated for the added insight.

How often do you tend to re-run analysis code and end up fiddling with tables in Word?

I have had to do it multiple times for every paper I’ve published. I do health economics research, which involves a lot of collaboration with physicians and people from other less quantitative fields (eg public health, implementation science, communications / media). From my experience those collaborators are unwilling to look at results until they look “finished” (ie are in the format that the target journal requires) and are part of a Word document that includes the working draft of the paper. Since they want to use Track Changes as version control, I can’t simply generate the tables again as a Word document and end up having to copy and paste them in, which is a huge time suck and runs the risk of making errors.

What are your thoughts on using Google Docs to prototype this? Correct me if I'm wrong in my logic. You're looking for a way to link end results of your analysis to a 'Word' document that will auto-populate with each re-run. The physicians etc. want a 'clean' manuscript that mimics graphs in the format of the preferred paper + has version control.

I'm not looking to get on a call but my pain points are as follows:

I hate reviewing the literature, it's too hard to organize and my workflow is terrible. I've tried different managers (zotero, mendeley), their add-ons, and search websites (google scholar, paperscape, others ...). I can never keep things organized and my train of thought is always getting interrupted by context switches. It takes like 10 clicks to open a paper and by then I don't even remember why the current paper cited it. I wish I could just click any citation in any paper and it'd directly open a new tab with the cited paper. Also I hate navigating single papers, the back button doesn't always work after clicking on a citation link depending in what piece of software it's open, I often lose my place when looking back for specific information. I don't know that there is a technical solution to this.

I also feel that I'm always missing papers when I search for those relevant to a problem, it's like I'm manually doing a union of sets of citations weighed by how relevant they are but when I talk to colleagues with knowledge of the problem they always come back to me with some that I didn't know about. The search websites are all relatively slow and pretty basic, they don't make me confident that I'm finding everything I should be. I'll often look up conferences' accepted papers listings for keywords and find things I didn't find through google scholar for example.

I wish there was a way to more easily highlight labs and their specialties. I have no idea what the solution to this would look like. I find that through practice I built an intuition of what the good labs were and what they worked on and what their biases were but I wish there was something to help with this. Maybe something to colour code author names so I don't have to parse dozens of hard to read names looking for familiar ones.

I'm part of three different organizations and never found a sensible way to share sets of papers with everyone easily.

I hate latex: positioning and sizing figures is hard, my formulas always come out ugly, I never know whether my ten thousands curly-braces are closed correctly before I compile.

I hate live online discussions of mathematical formulas, everyone has their preferred white boards full of ads that always bug, and I have to screenshot the screen to save what we talk about which is inconvenient, also I suck at writing math with a mouse. I hate writing live latex also because it's slow.

I like a lot of what overleaf has to offer for writing notes and asynchronous writing as a team.

That's all that I can think of off the top of my head with regards to the writing process. For context I only just finished a PhD (on a somewhat biomedical subject), maybe a lot of those pain points resolve themselves with more experience and different institutional practices, either way hope some of it is useful. Best of luck with your project.

Wow, thanks so much for sharing your pain points!

It's really too bad Mendeley butchered public groups feature, it could've helped solve the "are there any more readings about this topic that I'm missing" problem. You're preaching to the choir re: latex.

If you had a choose 1 issue that really grinds your gears, which one would it be?

a side-note - Have you tried a solution like Evernote to organize your thoughts?

Reply on news.ycombinator.com