3

My previous publication was about improving a known method A and I called my algorithm XYZ-methodA, where XYZ is a catchy word and methodA is the known method. Now I apply my general theory to method B, so I am thinking of naming my second algorithm XYZ-methodB. This second algorithm uses a sub-routine, that I also developed. I am thinking of giving it also some name like XYZ-methodC. However, this sub-routine is not that related to my general theory anymore, but is just used by XYZ-methodB to accomplish my goals.

Is this a good naming approach for methodB and methodC? Is it a good idea to use XYZ and put it everywhere like bumper sticker, indicating it is from me and increasing recognizability? That was kind of the idea.

What are good principles in naming my algorithms in general?

Naming them after myself is apparently not a good idea. But what is a good idea?

Make42
  • 979
  • 8
  • 21
  • How about following how they name cars... and the consequences such as « Nova » and the spanish market etc So, use names that are relevant - : « method B quick loop » etc.. – Solar Mike Apr 12 '18 at 12:09
  • Similar and the duplicate is worth reading : https://academia.stackexchange.com/q/108033/72855 – Solar Mike Apr 12 '18 at 12:10
  • @SolarMike: I don't understand your first comment. What do you mean with the cars? Can you elaborate? – Make42 Apr 12 '18 at 17:15
  • 1
    I'm going to vote for "Algorithm 1" and "Algorithm 2". Let other people name the algorithms if they think they deserve names. – JeffE Apr 12 '18 at 18:40

1 Answers1

1

The bumper sticker idea is quite smart, provided it is indeed the same general approach being applied. Otherwise it becomes gimmicky. Also, IMO, it would be better if XYZ has some functional meaning besides just being catchy, but I understand this is not always easy.

For instance, suppose A and B are heap sort and bubble sort respectively. Suppose your general approach has something to do with hashing, and makes the sort faster. A name like Hashed heap sort and Hashed bubble sort would make more sense and have better recall value/correct association than say, Snappy heap sort and snappy bubble sort. This may be an oversimplified example, but the point about functionality is worth remembering.

Ultimately, the line between gimmicky and legacy is not predefined, and it's not in your hands. Since you've already published XYZ-method A, you could see how the response to that nomenclature has been (are other authors using that name? Do your collegaues and co-workers call it by that name? Do other people see it catching on?). If the response is positive, and the above points are met, you could go for XYZ-method B. Else consider learning from this experience and make a fresh start with a new name.

AppliedAcademic
  • 13,178
  • 4
  • 36
  • 69
  • Yes, the same basic idea is behind the different algorithms. 2) The catchy name has a meaning besides being catchy. I originally had an analogy - although to be honest, since I don't have space to explain it, I am not 100% sure, readers get it :-/. 3) I should clarify: The paper is accepted, but the conference is still to be held, so I don't know reactions yet. The current paper is due today, so... ;-).
    • – Make42 Apr 12 '18 at 17:20
  • Going by that information, I would say try it out! With two names out, you get twice the feedback. – AppliedAcademic Apr 12 '18 at 18:04
  • That would inflate my bumper sticker idea though ;-). In any case: Here I am searching more for a name for method C than method B. I don't have good idea though. – Make42 Apr 12 '18 at 18:09