FOSDEM: more boring shit
Let's take a look at my annotated copy of the FOSDEM 2018 main talk schedule, shall we?
Twenty-five minutes you'll never get back.
Some bureaucrats arrive to explain to everyone how important bureaucracy is. One of them takes credit for introducing Java and XML to IBM, as though that is a praiseworthy achievement instead of grounds for a war crimes tribunal. The talk description focuses heavily on namedropping corporations known for ramming code into production whether there is consensus or not.
Some academics would like anyone at all to listen to them. Their thesis is that the European Union is the right organization to reinvent the internet, presumably based on the wild success experienced during the reinvention of Europe. Nobody involved appears to have actually done anything, either in the past or the present.
A Red Hat employee who is good at ARM processors would like to lecture everyone about something. We don't get to know what, because the talk description is blank, but we can guess: the speaker's pet technologies are the solution to all our problems, assuming we can just ignore the crippling problems of those pet technologies. This is merely speculation, but it doesn't really matter which PR fluff Red Hat chooses to excrete; they paid a lot to be a "cornerstone sponsor" and by God they're going to get a keynote slot.
Ten MORE minutes you'll never get back.
Bosch, an automotive manufacturer, sends a drone to ask that programmers line up behind Bosch. Bosch has decided to base their automated driving software on Eclipse, which allows us to infer some important facts: This software will never finish loading, much less begin to work. Your car will have the ask.com toolbar installed on every update. Very soon, nobody will ever hear of OpenADx ever again, but not before it makes the news for causing massive freeway-blocking traffic fatalities because someone's car crossed a time zone.
A programmer tells the harrowing tale of doing things other than programming, fucking them all up, and trying again. The talk description does not indicate that the speaker will present a solution to one of the oldest open problems in the software engineering discipline: "why anyone should trust software written by PHP programmers."
A bureaucratic parasite tries to convince the world that the United Nations should be taken seriously as a software project management organization. The talk describes the parasite's domain as "a multilateral participatory program" and "in collaboration with partners around the world," which is ancient United Nations code for "we are going to demand resources that we can use to demand further resources."
Another bureaucrat arrives to teach us how to like our jobs, providing those jobs are "cloning Borland Delphi and giving it to the Apache Software Foundation" or "cloning Microsoft Office and giving it to the Apache Software Foundation." No advice is offered for "living with what you've done, everyday."
An academic has spent some number of years shoving unix into a github repository, and now would like to read it aloud. Presumably some Greek student is furious that they had to pay money to attend this lecture last year.
A person nobody cares about will read the names of some software nobody cares about.
A self-described "industry thought leader" invites you to "experience his musings." In fact it's another academic attempting to convince us of the importance of his job.
Some students recreate an ancient computer without any of the things that made it interesting.
An Internet invites you to simplify your program's configuration file by importing a 130,000-line software grenade, over which you have no control. Wasteful and complicated constructs like "text files" can finally be discarded in favor of a globally-addressable nested key-value datastore which requires new software to be written for every type of configuration data.
Some academics arrive to pitch their latest invention: interprocess communications in the form of a packet-switched network. All you need to get started is a packet-switched network, on top of which they may pile untold complexity. Enthusiasts of doctoral thesis-defense trial runs will be sure to love this presentation.
A person who maintains software targeting hardware that does not in fact exist shows up to talk about the burgeoning field of hardware fanfiction. The speaker has successfully tricked several GNU projects into supporting this nonexistent architecture, which was a natural fit for their nonexistent Hurd operating system. This is the first time on record that a complete compute stack, from absent silicon to absent operating system to absent users, has ever been announced to be released Real Soon Now.
Obviously the only place that bugs cannot survive is within software that is not ever run, so the presenter would like to discuss the many different approaches to ensuring that unused software is subjected to a mind-numbing array of bureaucratic oversight, outdated standards documents, and half-assed formal verification procedures.
A webshit would like to brag about handing control of a software project over to a monte-carlo approximation of a project manager. The software project in question is webshit designed to expose root control of your computer to a web browser. This talk is the first multi-scale integrated model of terrible decision making and questionable practice.
Someone has shoved enough bullshit into the linux kernel, the Mesa graphical library, and the Android user space that they finally work together, if you break a lot of stuff and add a lot of otherwise-unneccesary software. The speaker is here to gloat about being involved with cramming so much garbage into so many disparate projects.
Glossing over the reasons they skipped "making LibreOffice work well anywhere", a company devoted to taking credit for other people's work is here to take credit for shoving a Microsoft Office clone into a web browser.
A comedy routine in which a professional database janitor pretends to honestly believe that MySQL is capable of doing anything quickly or scaling in a manner other than "run sixty of these and have fifty-nine of them lie to clients." After the talk, several spontaneous "how to migrate to PostgreSQL" talks will break out in the hallway outside, closely attended by sweating project managers who were not previously aware of what a trash pile MySQL and its advocates are.
Elasticsearch is a distributed customer-data exposure tool with a ransomware-friendly webshit interface. The company who charges money to clean up the mess has sent one of its less useful drones to drum up audience interest in the implementation details that make their product respond reliably and quickly to random internet assailants scanning AWS for data left unprotected by morons.
The Gluster team at Red Hat, bereft of customers, whiles away the hours by pretending it takes any work at all to reap performance benefits from faster hardware. The thesis seems to be that shitty software was acceptable when the hardware was shitty, but since storage platforms have improved, the bad programming and awful architecture of their project has become more obvious. Evidently it takes three people to apologize.
A Python arrives to tell us that Python 2 is irretrievably fucked and everyone should switch to Python 3, just like they've been telling us for the past decade. The primary products of the speaker's own employer rely entirely on Python 2, just like they have for the past decade. The talk will include plans for replacing Python 3 with Python 4, by the team who fucked up the previous transition so badly that Python 2 has been "deprecated" longer than the release interval between 2.0 and 3.0.
A monster uses a Python subset as a C++ templating language. The monster will be on hand to explain how to secure funding for similar monstrosities.
The speaker presents a horrible chimera of a programming language, wherein the drawbacks and limitations of Python are augmented by the drawbacks and limitations of C. The result is a language that introduces header files to Python and requires breathtaking amounts of boilerplate. The primary goal of Cython appears to be transforming the programming experience from "implementing a solution to a given problem" to "trying to guess when to turn off exception handling so that your code runs marginally faster."
Since nobody uses the eighteen million web services that Mozilla starts, ignores, deprecates, and discontinues each month, Mozilla has devoted actual resources to creating software devoted to making themselves feel like they have users. The speaker is willing to educate the audience on how to replace market penetration with a few hundred lines of code. The talk is in the Python track because of an implementation detail and because there is no "software nobody wants" track.
Security and Encryption
A buzzword enthusiast will talk about some software nobody will use, designed to run on hardware nobody wants. Another refugee from the "software nobody wants" track FOSDEM has once again failed to implement.
An IBM arrives to teach us how to use hardware he doesn't use. Lots of words will be expounded about integrating all kinds of software into this hardware chip, but a cursory glance at the speaker's own website reveals all this crap is such a pain in the ass that he just uses ssh-agent anyway.
Red Hat explains how they're fixing all their previous fuckups with linux disk encryption.
An LDAP programmer arrives to explain why we should all give a shit about the Bitcoin knockoff whose primary use is chewing through your processor whenever you watch videos on Youtube.
Someone has noticed that most of the problems with computers are caused by people.
The speaker seems to be confused regarding employment; he either works for Greenpeace or Mozilla, but since the software on which this talk focuses appears to function as intended, we can assume he does not work for Mozilla.
This talk focuses on the only hardware project at FOSDEM that actually exists in the physical world. This speaker does work for Mozilla, but his title is "Community Architect." Apparently Mozilla has automated their user-ignoring toolkit sufficiently that the people in charge of it have time to reach orbit, where they can pretend people haven't been doing this since the Kennedy administration.