The problem with being "too honest" about your work is that you would be the only one doing it.
When you read a "scientific" paper, you read it expecting it to be a sweetened account of the reality. The test have obviously been performed using only few samples; out of the the ten described features of a system, only the two described in depth are currently implemented; the source is not available because it does not work on the famous really hard cases of the class of problems it should solve.
The system is so polluted that if you start being as honest as you should be, the reviewers will think that you have done even less.
When you read a "scientific" paper, you read it expecting it to be a sweetened account of the reality. The test have obviously been performed using only few samples; out of the the ten described features of a system, only the two described in depth are currently implemented; the source is not available because it does not work on the famous really hard cases of the class of problems it should solve.
The system is so polluted that if you start being as honest as you should be, the reviewers will think that you have done even less.