Can I publish a paper that doesn't have a specific question to answer?

I have created a new data structure, and I intend to submit a paper about it to an ACM conference later this year. However, in writing the paper, I realize there is no specific "question" the research answers. I was simply curious if such a structure could be created. Can I simply write the paper, describe the structure and how it works, and then give a pseudocode example and perhaps comparisons to other data structures? The structure has some rather interesting properties, and I can see specific applications for it in statistical analysis. But beyond that, as far as I can tell, it's just a neat structure with some weird properties.

• Make a scenario shows when your structure is good/or better. – seteropere Oct 11 '13 at 5:44
• Doesn't your paper therefore answer the question "Can structure X actually be created/work?" – Carl Oct 13 '13 at 6:49
• The answers at a somewhat similar question ( academia.stackexchange.com/questions/28809/… ) could be of some help. – just-learning Nov 6 '14 at 2:27

I would suggest that you investigate those interesting properties further and then have a paper along the lines of "this structure has this useful property, more-so (or less-so) than these other structures." That is justify why someone would actually use this structure.

Think about it from the editors' point of view. Why should this paper be published? If you can't provide an answer with a straight face then maybe you should wait until there is a purpose for it.

I think "theoretical" papers are welcome! Just make sure to clearly present your contributions to the field, even if they are incremental. Ex: place a table with other related structures and show that your data structure improves on some operation, (retrieval, insertion, or whatever operations you have) operation on which other structures have a higher O().

State your advantages clear in the discussions section but mention them in abstract, introduction and conclusions.

Please make sure that your mathematical argument on why you get the specific O() is sound! It also could be helpful to collect some experimental data with your proposed structures and current "state of the art". The experiments doesn't need to work on some real problem. It is enough if you can generate some data having a distribution similar to what you might encounter in the real world.

Submit it and remember, reviewers are always right :( Don't give up if you get rejected, try a conference, or another journal instead.

Hope it helps!

If you can see some useful applications then I think you have your question: What is the best data structure in _____ (situation).

If it is not the best data structure in any situation then perhaps it is better to improve it so that it does something important better than anything else that is out there.