We did our homework. This was also a good chance to compare NoMachine with the competition (something we regularly do, as everybody should be doing in any business). Well, let me say… WOW. I mean, both software you mentioned are definitely fast by any measure, but… Well, you understand what I mean. I don’t know what the people here put in the software but let me say: good job ;-).
But let’s back at our scaling problem. Yes, I confirm that is a scaling issue. Both UltraVNC and TeamViewer make a better job at scaling. The reason is that at the present moment we use an implementation that is optimized for speed, rather than quality. This implementation was originally intended to be a fallback for when HW accelerated scaling is not available, but since HW accelerated scaling is proceeding more slowly than expected (people have been busy with other tasks) it’s basically the only scaling available at the moment. We can improve the quality of this SW scaling, of course. The problem is how to do that in real-time without impacting the rest. There are some ideas on the table. I promise you will see consistent improvements in the next releases.