Amir, let me provide a few comments. I am a diy speaker guy and I've measured a few speakers in my day. I do them indoors using the quasi anechoic technique used by many hobbyists, Stereophile and so on.
First, you are completely correct that these indoor measurement techniques do have advantages over even anechoic measurement. This may be a difficult myth to dispel but it is important. If you know what you are doing, and integrate the near field and far field measurements, you have a good picture of frequency response. Also, regarding spurious noises - this is not actually a problem. You generally send a short measurement signal (I use MLS signals) of about 4 seconds to the speaker - then repeat 3 times. It is very easy for software to identify any spurious noises by comparing the 3 measurements.
Second, the Klippel measurement system is amazing stuff, precisely because it provides a very 3d view of the speaker. Most of the time, people measure horizontal directivity and vertical as well, but I think that a full sphere measurement provides a better view. Earl Geddes has been sharing his technique/code for a modal measurement technique on DiyAudio. I am not totally familiar with his measurement protocol, but basically he uses fewer measurements behind the speaker, where there is less treble, and more measurements in front. I can put you in touch with the people who are continuing this work. You seem to have experience with programming and audio, so you may be the ideal person to pick up Geddes' work.
Geddes uses a simple rotating platform to do his measurements.
If you want to invest in a good measurement setup, I would recommend buying or building a turntable which works with ARTA or some other measurement package. These are not nearly as advanced as the Klippel products but will provide an accurate view of speaker directivity.
https://outline.it/outline-products/measurement-systems/et-250-3d/
See this measurement platform. I believe ARTA works with this one out of the box.
Now, regarding the utility of measurements. Harman's conception of the ideal speaker probably does reflect most listener preferences, but it seems to me that if you want to really interrogate speaker performance (and value) there are other speaker measurements which need to be done. You mention that bass performance is an important predictor of listener preference, but bass performance can't be completely characterized by FR measurement, especially with small speakers. It's essential to know how the speakers sound when they play loud! This requires distortion measurement, which is difficult to do in a repeatable and comparable fashion, but the equipment involved is simple - calibrated mic capable of high SPL.
Please see these ARTA application notes:
http://artalabs.hr/support.htm
Especially:
AP4 - Loudspeaker free field response
AP5 - Chamber for LF mic callibration
AP6 - Directivity measurements
AP7 - Displacement estimation using STEPS (ARTA's distortion analysis component)