suggest milestones

[BarbaraMcG]

Think about what we can realistically achieve to meet our target time frame, and which intermediate steps are needed to get there.

Submission deadline for ACL2021 is February 1, 2021https://2021.aclweb.org/calls/papers/

[BarbaraMcG]

• [ ] Submit paper to ACL2021 - 29/1/2021
• [ ] Send paper out for final feedback to colleagues - 14/1/2021
• [ ] Send paper draft out for initial feedback to colleagues - 15/12
• [ ] Turing NLP reading group presentation- 10/12
• [ ] complete first draft - 7/12
• [ ] third round of experiments- 30/11
• [ ] second round of experiments 23/11
• [ ] first round of experiments - 16/11
• [x] confirm initial set of experiments, baseline(s), and evaluation framework - 9/11
• [x] confirm milestones, meeting structure, basic architecture - 3/11

[BarbaraMcG]

I've suggested the milestones above. Let me know what you think!

• Do you feel they are realistic?
• What would you change?

[BarbaraMcG]

Assigning to Kaspar to start the round of comments

[kasparvonbeelen]

This looks, good. I would suggest to agree on the task and primary data next week. And keep the whole period till January for experiments. If we have to ask colleagues on feedback, that would be great, but time is tight. I think it's more realistic to have the experiments done and draft completed by 15/01. Writing the actual paper is something we could do in a week, if the results are promising.

[BarbaraMcG]

This looks, good. I would suggest to agree on the task and primary data next week. And keep the whole period till January for experiments. If we have to ask colleagues on feedback, that would be great, but time is tight. I think it's more realistic to have the experiments done and draft completed by 15/01. Writing the actual paper is something we could do in a week, if the results are promising.

If I can speak from my experience, it's really important to have early drafts because what we think is clear in our minds often isn't clear when we write it down. Also, only when we write down what we have done and we try to motivate it, we get ideas for what else we should do and what can be improved. I used to do the writing at the very end until I read this blog post, which completely changed my view. Since then, I've been using the "write the paper" first approach, and I must say it really works. Also, given the collaborative nature of this project, we won't be able to do all the writing in one week at the end because people will have other commitments.

[kasparvonbeelen]

Agreed on the need to start writing early, I prefer to do the same. I think the first draft should at least accurately describe the task and include a few preliminary experiments (to get a sense what we are doing and how well it works).

[BarbaraMcG]

I'm closing this now, and transferring the main milestones as separate tickets