We are talking about two different ways to get a libre baseband. One way is to reverse engineer without any help from the manufacturer and another way is to get the manufacturer to release specs and possibly source code. For the latter, you will need to muster some clout with the manufacturer. That means huge number of chip orders. So yes, 3000 people willing to pay $600 is not going to get us there. I doubt even 100000 will.
Reverse engineering on the other hand will work if you have sufficient number of people with the technical knowledge and skills, have interest in freeing up the baseband and aren’t prohibited from doing so( because they are employed by the manufacturer or other reasons). I would say that the number of people having interest in a libre baseband is larger than the number using Lineage OS. However the other requirements reduces the number drastically. Even now Lineage OS has to include proprietary drivers and libraries. Look also at the hardware support offered by Replicant. So if its this difficult to free up drivers, how is freeing up baseband going to happen?
Nicole Faerber from the Purism team has said
Just for the records, the idea to develop a free baseband firmware is deeply rooted in our development and research. At the moment this goal is not feasible to target since there is too little information out there on existing chips and too little existing software to base on. To do everything from scratch would take way more than the money that we will have at hand for development (please keep in mind that the campaign money also has to pay for the devices, not only the development).
But while choosing the mobile modem chip(s) we will keep this goal in mind and choose a chipset that at least could be, later on, hacked and freed.
What I want to assure to you is that we will do all and everything we can possibly do within the constraints we have to build a device that is as free/open and as transparent as possible. And we also think about its future, i.e. that the parts that might not be free from day one can be freed later on.
Now, who is going to do the hacking and freeing? What is the timeline for this? The team obtained approx $2.3Mil from their crowdfunding. The stretch goals for the funding go to $32Mil but don’t target freeing baseband.
IMHO, its not going to happen, at least in the next few years. So then one can ask whether one wants the modem at all, especially if it means getting a device now and for a lower cost. We can keep arguing about simplicity of design and cost of Librem5 with or without a modem but its pointless unless someone from team answers. They are the ones doing the work. Are you a member of the team?
Now you can ask why anyone would be interested in a modem free version after the Librem5 is released. That would depend entirely on how much isolation they manage to get between the modem and the rest of the phone. For example the modem would need access to speakers and microphones for phone calls, camera for video calls, maybe GPS for emergency calls. But will it have direct access or will it get data packets from the i.MX? Will it share RAM and storage with the i.MX? On the software side, will it be possible to require explicit permission every time the modem wants access to the mic,camera,GPS etc? The exact details of isolation will tell me if Librem5 without baseband is same as Librem5 with baseband turned off. Its a topic to be revisited after the release and its not as obvious as you point out to be.
Regarding, govt backdoors, mass surveillance is what I can hope to avoid. If a govt wants to target you specifically, there’s lot many things they can do other than tracking your phone.
What is also concerning is stuff like CarrierIQ. The latest report shows them using Sierra Wireless MC7455. Another chip from the same manufacturer, EM7345 is reported to have the CarrierIQ spyware.
Regarding backdoors, I hope you mean hardware backdoors since all software is supposed to be available right from the boot loader stage. Other than analyzing all network traffic, what else can be done? Hopefully there will be enough users do so and they’ll get caught.
If you want to talk about ideals, I will personally need to do all of these:audit the source code, audit the hardware design, manufacture the hardware, compile and load the software. A single person doesn’t have the resources (knowledge, skills, money) for this. So lets not go there. At present, getting all the source code is a goal that can be possibly achieved (except for the baseband).
Regarding software design, I can make phone calls using an “Arduino phone”. Doesn’t mean that its great software. If everything is already there, why are they yet to implement phone calls?