Are we doing the real beta testing for Devialet?

[raph972]

Multi-channel 5.1 setups were promised when the original Phantoms were announced, five years ago. We're allowed to complain about how poorly Devialet have kept their promises. They have consistently over-promised and under-delivered, with no transparency as to their road map.

A perfect example of over promise, under deliver.

The software engineering team was asked what features could be implemented given the basically locked-in hardware architecture of Phantom, to which they responded exhaustively, as engineers do.
They did so fully aware that it would take years to implement all of these features, and made sure to convey that what they described was basically a roadmap for the foreseeable future. From their point of view, it guaranteed that the product would keep evolving in the eyes of the early buyers, with nice new features being introduced as they were developed, tested, and readied to be shipped, over the coming years.

However the marketing department, approved by the management, always in need of more selling points, decided on day one in the Phantom Whitepaper to disclose all of that roadmap to potential buyers. Multi-channel setups were announced in that phase.

What followed was perfectly predictable. Customers bought in on the promise of this golden do-it-all product, only to be disappointed and rightfully angered when it never materialized.

For the team after that, it was a constant rush to deliver on those company promises to which none of the engineers had actually committed. This is the worst way to build software. To achieve those goals, a lot of shortcuts were taken especially regarding testing, but also architecture. All those shortcuts tied the DOS 1 software stack to the Phantom/Dialog hardware architecture.

All the while, management willfully decided to ship untested features for the sake of showing progress and even gave way to new urgent marketing-gimmick features without a sustainable business model (Live on Phantom), while still refusing to allocate the resources to do proper software Quality Assurance.

When the time came to work on Core Infinity for Expert Pro, instead of leveraging all the software that had already been developed, they had to start almost from scratch on what would become DOS 2. At that time, management did not understand how that could be the case. It took them months explaining/showing that DOS1, given its development conditions, really was a one-off product and that it would be impossible moving forward to support Phantom Classic, Core Infinity and what would become Reactor on that stack.

To sum it up, the Devialet management at that time did not understand software.
Engineering there had been driven by executives from the electrical engineering industry, where the software is actually firmware running on as single embedded chip without any external software interfaces. Phantom in comparison is a complex distributed system with real-time audio delivery constraints operating on unreliable consumer-grade networking infrastructure.

Devialet is a hardware company. That's where their focus is. What clearly illustrates this point is the effort they put on hardware QA at their different industrial sites where each hardware part is thoroughly tested on one hand and the repeated refusal to hire software quality assurance technicians on the other. The R&D headcount allocation reflects this also.

Things have changed under new management and key people responsible for a lot of these decisions have been exited. Today they have a small QA team, but the damage is done when it comes to the reputation of Devialet's software.

DOS 2 is late, but it has sound foundations. It's limited in its feature set, but can only improve over time. It currently powers Phantom Classic, Core Infinity and Phantom Reactor, all very different architectures, proving its portability moving forward. We should reasonably expect more products in the future, all compatible with one another, in the unified Devialet ecosystem.

It's worth noting, however, that most of the team that brought DOS 2 up has now left the company, tired of all this repeated nonsense. So only the future will tell...