A single definitive paper, or an initial manuscript and a followup?

Question

Suppose that someone wants to publish his idea and his first successful experiments in a venue and then perform some extra experiments (for a more powerful proof) and publish it in another venue.

1. Submitting the first to a conference hope to be accepted soon and then submit the second one after a while to a journal, referencing the first.
2. Submitting both of them to journals and write them in different manners.
3. Waiting until future experiments finish and combining them to be published as one journal paper.

I count, here, some negative and positive points of the ideas

• In the 1st idea, the first paper proves that your innovation has improved a method in the state of the art and the second one refers to the first and says that this innovation, even, is able to improve other more-powerful methods in state of the art. Of course both of them are contributions. But, lacking the powerful contribution of combining them in one paper.
• In the 1st idea, you are able to present your work in an international conference and take advantage of global feedback.
• In the 2nd idea, you should suffer side effects of deleting some parts from each paper, to avoid plagiarism.
• In the 2nd idea, you will have two journal papers which normally are more creditable than having one conference and one journal paper.
• In the 3rd idea, you should wait for finishing the experiments.
• In the 3rd idea, though you have one journal paper instead of two, this paper is published in a more creditable journal than others.

However, please let me know the optimal choice.

Answer 1

Regardless of your elaborate schemes, you will always need to take into consideration that you cannot "sell something" twice.

If the only novelty that you present in a second paper are additional experiments (and deriving conclusions from these), then you must never ever state something different as a contribution of the paper. And in many fields, this would be quite a weak contribution, basically only leaving your schemes 1 (conference+plus extended journal version) and 3 (single publication) as realistic, unless you want to settle for a very weak venue for the second paper.

Slicing your work is also frowned upon in academia, for several reasons. First of all it looks as if you are trying to inflate your CV. If it makes grasping your overall contribution harder, it should also be avoided. Also, there should always be some scientific reason for splitting up your work. For example, if the target audiences for the two parts of your work have a very very limited overlap, then splitting up your work may yield a better service to the community. If the parts of your work are mostly interesting in isolation, and actually only of interest to few people in their combination, then splitting up the work is also sometimes the best way to go. And finally, if the underlying techniques of the two parts of your work are so different that publishing the two parts together would lead to a paper that is only readable by the very few experts that have knowledge about all techniques, then the split may also be justified. In all cases, you have to reference your other work and explain the differences in contribution.

In some fields with published conference proceedings with page limits, it also became customary to to leave out some additional information in a conference paper and writing longer journal version later. This is ethically OK because the journal paper in some sense replaces the older paper. This seems to be precisely your "scheme 1".