I would like to preface this historical overview of Bosch's EFI adventures by stating that I have the utmost respect and appreciation for the Bosch EFI engineers who pioneered to bring EFI to the masses and improve ICE operation for the betterment of society. We owe then a debt of gratitude.

The first large scale use of Bosch EFI was actually on the VW Type III air-cooled vehicles in the late 60's. This EFI system was a relatively low pressure (30 PSI), design that used minimal temp and throttle position inputs for the ECU to calculate the injector duty cycle. It was a pretty sophisticated design for it's time but had some issues. The system was based on variable resistance to signal the ECU as to what duty cycle was required. It also used a cyl. head temp sensor to control cold start enrichment. Not too long after these vehicles went into service customers started experiencing erratic operation. It was determined that oxidation of all of the ground wires was causing false resistance values to be seen resulting in improper AFRs. After a number of years fighting the good fight Bosch scraped this system and moved on to an airflow based swinging door EFI system.

The swinging door or barn door flow meters as they were often referred to had a spring load flapper door that opened based on engine manifold vacuum. This EFI system was adopted to many Euro models as auto makers needed better performance and lower emissions. In applications like the VW Bug engines, this flow based system worked reasonably well until cold weather arrived. The typical intake runners on an air-cooled VW meet in a center plenum area. A fifth injector was added to enrich the AFR on cold start-up. The problem was that the intake runners were ~ 15" long and when you cranked the engine to start it cold the air entering the cyl. was not enriched until the mixture from the plenum reached the cyl. As a result there were a lot of cold start back-fires from excessively lean AFRs. When the engine back-fired the force was sufficient to slam the swinging door against the stop and bend the door to the point where the flow meter would now bind making the vehicle undrivable. VW replace many thousands of these damaged flow meters in the first couple years.

When Bosch was told of the problem the engineers developed a revised swinging door with a spring loaded valve that could open in the case of engine back-fires releasing pressure. This worked OK for 5-10 back-fires as the lean cold start AFRs had never been corrected. After 5-10 back-fires the aluminum by-pass valve stem would break and the parts were sucked into the engine where they did internal damage many times. So after wrestling with those problems for a few years Bosch moved on to CIS injection.

CIS is a high pressure FI system that worked reasonably well in many applications. One drawback to CIS is that it restricts WOT airflow to the point of reducing peak HP similar to the swinging door airflow meters. At the time CIS was acceptable however in many applications. One common problem with CIS EFI was that the warm-up regulators that managed the high fuel pressure would fail way too frequently and with a cost of around $100 in the day, it was not a welcome problem. What turned out to be the downfall of CIS in many cases however was the high pressure electric fuel pump. Bosch went thru several years when their electric fuel pumps on CIS vehicles would self-destruct and pump minute metallic sediment thru the entire CIS fuel system plugging injectors, warm-up regulators, etc. The replacement hardware was typical close to $1000 plus labor so Bosch moved on to hot wire MAF.

Hot wire MAF was another step forward in EFI tech. What was unfortunately discovered after a period of customer operation was that PCV vapors could oxidize the hot wire causing erratic data input to the ECU. Cleaning the hot wire with electrical cleaner some times helped and other times required the replacement of the expensive hot wire element. Bosch also had some electric fuel pump issues in this era but nothing like with the CIS era.

Seeing as though the hot wire issue was not easily cured Bosch decided to move to a hot film MAF system. The hot film design has worked better than the hot wire system but even the hot film system is subject to contamination by PCV oil vapor. So while Bosch has made great progress with their EFI systems, none of them were defect free in design or operation. The latest systems are quite good but not perfect by any stretch of the imagination. Other brands of EFI based on speed-density for example suffer none of the shortcomings of the Bosch systems. While there is currently no perfect EFI system that I am aware of we have come a long ways from the days of Bosch's first efforts on EFI.

I thought Marelli was an illness? Is brother Marelli a cousin to Lucas electric?

All VAG gasoline Bosch ECU's after 1998 (ME7, MED9, MED17, MG1) are speed density.
MAF is a hack on them. In fact they are calibrated at the factory with a MAP sensor in the manifold, and then the model is calibrated to get load/MAF value from that.
Hence why it is fairly simple to go from MAF to MAP on them - replace MAF sensor with a MAP sensor. Set manifold pressure directly from sensor (as done during calibration), instead of the MAF. Add a filter on the MAP signal. That's it.

So the info presented here is valid until 1998.

On VAG HFM it calculates ps_w from delta "fresh" rl_w via integrator...
So it's like reverse speed density, that's tacked on there. If SY_DSS is 1, then all that crap is omitted, and it just takes ps_w from sensor directly, and then goes the normal path: ps_w -> load.
Basically they left the entire manifold model in there, but it does nothing. It goes from uhfm_w to load to pressure to load, and the rl_w you have in the end is basically "directly" from MAF, because the calculation cancels itself out.

That's why I said that HFM on Bosch ME7 and up is a hack. It wasn't designed to use it. They made a speed density ECU, and then basically added a module that additionally to the normal calculation also runs the same module in reverse to get ps_w from rlroh_w.
But they needed ps_w for the throttle control, to calculate delta pressure (manifold vs pre-throttle). So they could not omit the entire module... It also says in the FR, the application notes that to calibrate the manifold model a pressure sensor is needed in the manifold, so it makes an accurate pressure reading from the MAF signal, because it relies on that to do everything else.

That's why a certain somebody on here made himself laughing stock, when he said that going from HFM to MAP on Bosch ME7 is a hack. It's actually just deleting code/deleting the hack that was added to a SD ecu to use a MAF.

Yup (if you can find one that references it). Alfa/Fiat stuff works the way Bosch intended, so the way you know it is the way the ECU was designed to run.

alpha-n is the "secondary load" on ME7, for when primary load (HFM or DSS) fails.

I was specifically talking about speed density via DSS or MAP.
Stock no VAG Bosch ECU has a MAP sensor until MED9 used on RS3, TT-RS and RS6 V10.

Yeah they do. But I am not sure if they came out before the RS6 V10. I think after.
EDIT: They came before. '04.

They have both SY_HFM and SY_DSS set... to see what is used for what requires looking through the relevant code or FR.
MED9 on the RS3/TT-RS and RS6 V10 does not have any MAF though.