Well. Rybka 3 is known to work in Aquarium :) So Rybka 3 IS an engine.

As far as I saw from Rybka's output, it's behaving quite like an UCI engine unless you start giving the extra non-UCI commands, and Shredder GUI obviously doesn't do it.

Zruty:
I will do more scientific analaysis soon. These are just my first impressions. For many purposes (engine vs. engine, etc.) Aquarium looks to be superb, but for straightforward infinite analysis, it is much harder to use GUI than other GUIs, e.g., cannot find keyboard shortcuts for basic functions like Deep Analysis, Stop Analysis, Delete variation, promote variation, Clear Hash, etc.

  Also, when stop analysis, the analysis is lost, so one has to write it down while it is going. And there is no point using Stop/Insert, since one typically wants to insert only 1 or at most 2 plies, not a very lengthy variation.

   Hopefully, all these basic glitches will be ironed out in the next iteration of Aquarium. Nevertheless, Rybka 3.0 seems much, much stronger than ver. 2.3.2a based on my work today in a game I had analyzed for several days with ver. 2.3.2a.

Victor,
  You are right that the overhead of Aquarium is a minor factor. The big problem is ease of use. As Zruty points out, Aquarium is a very powerful software product which can do much, much more than other GUIs. That is Aquarium's great strength and advantage, but also its great weakness.

  We see the same problem with CA9 vs. Chessbase. CA9 is a much more powerful program than Chessbase. But for the average user who simply wants to easily do the same routine functions (copy games between databases; promote, delete variations in a game;merge games etc.) CB is much faster and more intuitive, much easier to use. (Delete variation in a game is unnecessarily complicated in CA9.)

I have both CA9 and CB. Although I have spent much time learning and using CA9, and although I greatly appreciate Hugebase, I find that I end up using CB for my database maintenance.

  I am not criticising Aquarium. It is a superb achievement, a software product with great power. So my comments are like criticizing an apple for not tasting like an orange.

  You and the Convekta team obviously  made a marketing decision when you created a super powerful Aquarium. I certainly hope this decision works out for you, since Convekta has been (along with the Rybka team) the most innovative force in chess software in the last 5 years.

  However, for a (small/large?) part of the market for GUIs, Aquarium is much too complicated and too hard to use.

On an average day,  I use Rybka for 5-6 hours of interactive analysis of my ongoing games. That means that I do certain repetitive actions over and over again very quickly, e.g., start/stop analysis, clear hash, delete variations, promote variations, load games, delete games from databases, change databases, etc.

I do these things very fast and efficiently to find the best moves in 16 ongoing games. (I cannot post a position from an ongoing game.) I have been working with the Aquarium manual, which I have printed out. Each step described above is very slow compared to the very fast operation in Deep Shredder GUI, e.g., in Aquarium, I do not even know which engine was installed as the default or what its parameters are in Aquarium.
To take another example, if I want to do infinite analysis, I have  many choices (Analyze, Start Custom, Deep Analysis, Game Analysis, etc., and each one alters the listed variations in ways I don't want.

  From the beginning of Aquarium's development, you responded to the many requests on this board for more and more features. I often posted requests to keep Aquarium simple and small. However, I assume the market was for a large, complicated program. Perhaps someday you may do a lighter, smaller version of Aquarium ("Aquarium Lite"), a much smaller, faster GUI, quick and easy to use.  Just remove many of the features in Aquarium and you will have Aquarium Lite.