We have seen an explosion in machine learning in the past decade, alongside an explosion in the popularity of free software. At the same time as FOSS has come to dominate software and found its place in almost all new software products, machine learning has increased dramatically in sophistication, facilitating more natural interactions between humans and computers. However, despite their parallel rise in computing, these two domains remain philosophically distant.

Though some audaciously-named companies might suggest otherwise, the machine learning space has enjoyed almost none of the freedoms forwarded by the free and open source software movement. Much of the actual code related to machine learning is publicly available, and there are many public access research papers available for anyone to read. However, the key to machine learning is access to a high-quality dataset and heaps of computing power to process that data, and these two resources are still kept under lock and key by almost all participants in the space.1

The essential barrier to entry for machine learning projects is overcoming these two problems, which are often very costly to secure. A high-quality, well tagged data set generally requires thousands of hours of labor to produce,2 a task which can potentially cost millions of dollars. Any approach which lowers this figure is thus very desirable, even if the cost is making ethical compromises. With Amazon, it takes the form of gig economy exploitation. With GitHub, it takes the form of disregarding the terms of free software licenses. In the process, they built a tool which facilitates the large-scale laundering of free software into non-free software by their customers, who GitHub offers plausible deniability through an inscrutable algorithm.

Personally, I've been doing things with Debian and free software for almost thirty years. I was really shocked when I heard that Debian funds were being used to try and shut down independent, volunteer-run web sites publishing news about Debian itself.

I had a closer look at the situation myself and realized that nobody has registered a Debian trademark in Switzerland. Therefore, the Software Freedom Institute submitted an application for the mark.

The application was submitted on 14 May 2022 and granted on 8 June 2022.

Software Freedom Institute SA immediately published a statement authorizing legitimate use of the trademark in domain names.

It appears really bizarre that some rogue members of Debian have collaborated for months with an expensive lawyer and yet none of them bothered to ensure they were holding a registration in Switzerland before filing their attacks at WIPO. The Swiss Institute for Intellectual Property charges a fee of just CHF 550 to register a trademark. That is less than what Debian pays for two hours with their lawyer. Einstein himself used to work there but you don't need to be Einstein to realize who got better value for their money in this case.

Also: Louis-Philippe Véronneau: Montreal's Debian & Stuff - June 2022

Debian Community News is sharing our response to the WIPO UDRP harassment from Jonathan Cohen, Charles Fussell.

The response contains a number of confidential emails, including one from the missing Arjen Kamphuis and we hoped to send it directly to the administrative panel.

We believe our response is so strong that Debian Project Leader Jonathan Carter will totally capitulate and withdraw the complaint from WIPO. If he withdraws the complaint, we believe he must resign as DPL.

Corruption in the WIPO email systems

Unfortunately, there are problems with the WIPO mail servers. When we previously tried to communicate with WIPO by email, we received errors like this:

: host mx10.unicc.org[193.223.57.49] refused to talk
    to me: 554 Blocked - see https://ipcheck.proofpoint.com/?ip=78.46.198.153

Do we have to go to Geneva and pay somebody a bribe just to accept our response?

The latest phase of the Amber Heard/Johnny Depp saga is a defamation trial in the United States. Heard falsely accused Depp of pushing his ex-girlfriend down the stairs. The ex-girlfriend, Kate Moss, joined the trial by video link to tell the world that she actually fell. In other words, Heard had tried to deceive the court and public opinion.

Lies like that belong in Molly de Blanc's infamous whisper network. Maybe Heard can volunteer to join the Debian anti-harassment team?

Nonetheless, the Heard/Depp trial demonstrates the foolishness of going to court. Debian is spending vast sums of money on a greedy London lawyer to pursue a malicious claim at WIPO. Today we expose another one of the lies in their legal documents.