I'd say it's a mixed experience. However, I found it really helpful to me to have a second portable oscilloscope (with better siggen), so I kept it.
They fixed some of the major bugs and missing features in DHO900. Less UI clutter with full screen mode. FFT is less buggy. The record function has lower dead time, and you can finally save the recorded data.
But the Bode plot function is still pretty broken.
My biggest complaint is still the aliasing problem and how MHO98 "tries to deal with it". The digital filtering theory is pretty much proved for me when I attached a fast edge pulse generator to the scope and the waveform looks very similiar to the step response of a 1.2GHz brick wall low pass FIR filter, with the symmetric ringing pattern.
To sum it up, I think MHO98 is a usable 1GHz single channel oscilloscope if you are very sure that the bandwidth of the measured signal is properly limited to 1GHz or do not care about the time domain waveform fidelity very much, and a usable 250MHz dual channel oscilloscope as long as you remembered to have the 250MHz BW limit enabled, and a qual channel sampling theorem demonstrator.