Author: FRIGN <email@example.com>
Date: Mon, 26 Sep 2016 18:27:39 +0200
Cleanup and rewrite the slcon 2016 abstracts
They need to have a neutral tone and should not be too long.
1 file changed, 93 insertions(+), 131 deletions(-)
diff --git a/suckless.org/conferences/2016.md b/suckless.org/conferences/2016.md
@@ -1,200 +1,162 @@
suckless conference 2016, Hofheim am Taunus, Germany
-slcon 2016 was held in Hofheim am Taunus near Frankfurt/Main on 2016-09-(23-25).
-Please note that we will collect a conference fee for the conference room and
-lunch/drink package. If the preliminary schedule sticks to Saturday being the
-only day when we book the conference room, we need to expect around *45 EUR*
-conference fee per person.
+slcon 2016 was held in Hofheim am Taunus near Frankfurt/Main on
-(19:30) *Welcome* in the Biergarten "XXL Waldgeist", Anselm R Garbe
-> We will gather together and check out the XXL Waldgeist.
- Waldgeist Hofheim
- Schlossstr. 70
- 65719 Hofheim
+(19:30) *Gathering in the "Waldgeist"*, Anselm R Garbe
+> We gathered together in and checked out the
+ [Waldgeist](http://www.derwaldgeist.de/) restaurant that was within
+ walking distance of the hotel.
(10:00-10:05) *Welcome*, Anselm R Garbe
-> Anselm will open slcon 2016 talk session
+> Anselm opened the slcon 2016 talk session
(10:10-10:40) libzahl -- simple bignum arithmetic, Mattias Andrée
- Software dependent on algorithms inherently perform
- better the more complex algorithms. Can we despite
- this compete with GNU MP and still have a simple
- solution? This is a discussion about libzahl's design
- and future.
+> The quality of algorithms usually increases with their complexity.
+ Can we, despite that, compete with GNU MP and other libraries
+ and still provide suckless solution? This talk discussed libzahl's
+ design and future.
(10:50-11:15) XML damage control, Silvan Jegen
- XML is a horrendously abused file format that suffers severely from
- over-engineering. There arguably is one legitimate use case for (a subset
- of) XML however: text markup. In this presentation I will show why that
- is and how to best deal with XML when encountered in such circumstances.
- To ease the pain for people like me who have to work with XML every day,
- I will first give an overview of different parsing strategies employed in
- XML libraries and introduce the most bearable APIs of the ones available.
- The second part of the presentation is dedicated to the benchmarking of the
- most sensible XML libraries and to the discussion of possible alternative
- parsing approaches as well as their implementation.
+> XML is a horrendously abused file format that severely suffers from
+ over-engineering. However, there is an arguably legitimate use case for
+ a subset of it: Text markup. This talk gave reasons for that and showed
+ how to deal with XML in these circumstances.
+ The presenter, forced to work with XML every day, first gave an
+ overview of different XML parsing strategies and presented a few decent
+ libraries for this purpose. To allow comparison, he benchmarked the
+ most sensible XML libraries and discussed alternative parsing approaches
+ as well as their implementation.
(11:20-11:35) Stali Pi B+, Manu Raster
- Report on readying Stali for the Raspberry Pi B+ (32 bit). It is
- also a story about monsters and maiden when we take into comparison
- other piles of code (kernels, distros etc.) and their build
- systems. Portability conceptions encountered en route are
- reviewed and ranked from 1 to 3 (depraved, naive and not too bad).
+> This talk gave a report on readying stali for the Raspberry Pi B+
+ (32 bit) and is also a story about monsters and maiden, as it compared
+ the stali approach to other codebases (kernels, distros, etc.) and their
+ build systems, evaluating portability concerns encountered along the way
+ on a scale from "depraved" over "naive" to "not too bad".
(11:40-12:05) farbfeld and color spaces, Laslo Hunhold
+> This talk discussed the future role of color management in computing
+ in light of recent developments and gave a future perspective on
+ necessary changes to the handling of image formats like farbfeld.
(12:05-13:25) Lunch break
-(13:25-14:10) SCC and QBE for practical compilation, Roberto E. Vargas Caballero
+(13:25-14:10) scc and qbe for practical compilation, Roberto E. Vargas Caballero
- For now decades, we have witnessed a dramatic increase in the complexity
- of compilers. Mainstream compilers have become huge C++ programs that try
- to incorporate all the state-of-the-art research. SCC and QBE propose a
- new approach to compilation where simplicity and predictability prime over
- reckless efficiency. These new design constraints have security benefits
- (as advocated by the "boring crypto" movement) and allow bootstrapping
- new systems more easily.
+> For decades we have been witnessing a dramatic increase in compiler
+ complexity. Popular compilers have become huge C++ programs trying
+ to incorporate every aspect of state-of-the-art research.
+ This talk proposed scc and qbe as a new approach to compilation, where
+ simplicity and predictability are the primary objective over reckless
+ efficiency and optimization.
+ It presented the security benefits resulting from this alternative design
+ approach, that is also advocated by the "boring crypto" movement, and
+ demonstrated how it allows easier bootstrapping of new systems.
(14:10-15:15) The Myrddin Programming Language, Ori Bernstein
- Myrddin is a programming language that was written for a saner, easier to
- use programming environment that I find more pleasant than C. The language
- lives somewhere between C and ML in design, with parametric polymorphism,
- type inference, closures, and pattern matching. In some ways, it can be
- thought of as a suckless rust.
- It also comes with a complete replacement for many standard libraries on a
- number of popular (and less popular) platforms.
+> This talk presented the Myrddin programming language that was written
+ in the interest of a saner and easier to use programming environment
+ the presenter, inventor of the language, deems more pleasant than C.
+ The presenter elaborated on its position between C and the ML
+ programming language by design using concepts like parametric
+ polymorphism, type inference, closures and pattern matching, concluding
+ that it can be thought of as a suckless Rust. In this context,
+ examples were presented using a replacement for many standard libraries
+ on a number of popular (and less popular) platforms.
(15:20-15:50) ii-like chatting improvements, Jan Klemkow
- Since slcon 2015 I have made several improvements in my ii-like chat
- infrastructure. In this talk I will give an overview of my various
- activities in this area. I will present new features in my UCSPI tool
- chain, a new modular front end 'lchat' and the idea of runit integration.
+> Since the last conference, the presenter had made several improvements
+ in his ii-like chat infrastructure and in this talk gave an overview of
+ his various activities in this area, presenting new features of his
+ UCSPI toolchain, a new modular frontend "lchat" and the idea of runit
(15:50-16:25) text engine, Francesc Hervada-Sala
- The right approach to software systems was introduced by the Unix programming
- environment over forty years ago with a file system that opened access to data
- across all applications and with a universal interface model based on strings.
- This approach must be deepened. Let us define "text" as the semantic data
- structure that results from parsing strings. Imagine that a software system has
- a "text engine" at its kernel. When you edit a run control file and save it,
- the text engine parses it and stores its parse tree into the central
- repository. When you execute an instruction at the shell, the text engine
- parses the command line and sets all parameters in the repository before
- executing the command. An application retrieves a configuration parameter by
- querying the text engine. In such a system parsers for particular formats or
- languages would not be embedded into particular applications or commands,
- they would be stand-alone software units with the sole purpose of converting
- a string into a semantic representation in the text engine's repository.
- Applications would not have to deal with and not be tied to particular file
- formats and languages, and the user would be able to choose the format or
- language to use for any configuration statement. The user would change the
- name of a file and would not need to adjust all references to it inside
- scripts and other files, because the reference in the repository (say as
- inode) would remain unchanged. To experiment with this new
- approach I am beginning to develop a text engine
+> The right approach to software systems was introduced by the Unix
+ programming environment over forty years ago with a file system that
+ opened access to data across all applications and with a universal
+ interface model based on strings.
+ This talk claimed that this approach should be deepened by defining
+ "text" as the semantic data structure that results from parsing
+ strings, theorizing a software system with a "text engine" at its
+ kernel and discussing improvements it can bring to modern data
+ processing and user experience.
(16:25-16:35) Coffee break
(16:35-17:15) Suckless Image Processing, EML
- An image is an array of numbers. Since arrays of numbers are
- natively represented in C, this language is well-suited for image
- processing algorithms. Moreover, many complex image processing
- algorithms are pipelines of simple, independent steps. Thus,
- the UNIX programming environment is an ideal platform for image
- processing, where several simple programs written in C share
- information through UNIX pipes. In this talk, we will see many
- examples of simple image processing utilities (blur, local maxima
- detection, interpolation, warping, pointwise algebraic expressions,
- morphological filters, block matching, optical flow, image fusion)
- and how they are naturally combined to perform advanced imaging
- techniques; in the selected example, we show a reflection removal
- method based on the combination of several images. An important
- issue, with far-reaching philosophical consequences, is the data
- format written through the pipes. We explain why the farbfeld image
- format---proposed by the suckless community---is inappropriate in
- general, because of the need to represent negative and floating-point
- pixel values.
+> This talk picked up the conception of an image being an array of numbers
+ and deduced that C is well suited for image processing algorithms,
+ given arrays of numbers are natively representable in it.
+ It also elaborated on and exemplified how many complex image processing
+ algorithms are pipelines of simple, independent steps, making the Unix
+ programming environment an ideal platform for image processing with
+ several simple programs written in C sharing information through
+ In this context, the problem of selecting a proper file format for
+ pipe interchange is discussed and claimed that the farbfeld image
+ format is not suitable for this task and general image processing.
(17:20-17:45) shared farbfeld, Jan Klemkow
- Farbfeld is a good base for general purpose image processing. The
- tools that have been made so far demonstrate that sophisticated image
- processing can be achieved using cooperative programs. In this talk I
- will present a shared memory interface that speeds up the usage of
- farbfeld tools. Furthermore, I will sketch the architecture of a
- suckless-style Photoshop-like graphical image editor, and present a PoC
- implementation together with some benchmarks.
+> This talk explored farbfeld as a basis for general purpose image
+ processing based on the idea of cooperative image processing tools,
+ presenting a shared memory interface that speeds up the usage of
+ these tools. Based on this foundation, the architecture of a
+ Photoshop-like image editor is presented with a proof of concept
+ implementation and benchmarks supporting this idea.
(17:45-17:50) Display servers, Mattias Andrée
(17:50-18:30) stali learnings and beehive observation, Anselm R Garbe
- Anselm will present his learnings with stali since the last slcon. In
- particular he will discuss his new goals with stali and reasoning behind why
- making stali self-bootstrappable is a bad idea. He will also
- demonstrate stali as platform for observing his beehives.
+> This talk presented the experience gained with stali since the last
+ slcon, discussing new goals and why self-bootstrappability is a bad
+ idea in particular. As a proof of concept, the presenter demonstrated
+ stali as a platform for observing his beehives.
Official slcon 2016 talk session end.
-(18:55-19:20) suckless.org e.V. Mitgliederversammlung (yearly suckless.org e.V. member meeting)
+(18:55-19:20) suckless.org e.V. Mitgliederversammlung (suckless.org e.V. general assembly)
- - Report of the chairs, Anselm R Garbe, Laslo Hunhold
- - Report of the treasurer, Jan Klemkow
- - Other topics (members can raise topics until the begin of the Mitgliederversammlung)
+> - Report of the chairs, Anselm R Garbe and Laslo Hunhold
+ - Report of the treasurer, Jan Klemkow
+ - Other topics raised by members
-(19:30-) Social event in Frankfurt City
+(19:30-) Social event in Frankfurt/Main City
No talk session plans, hack sessions and departure day.
-You can arrange your booking at the [venue](http://www.hotelamrosenberg.de) if you like:
- Hotel am Rosenberg
- Wielandstraße 24
- 65719 Hofheim am Taunus
- +49 (0) 6192 292-0
-When booking a room, use the booking code: slcon or refer to the suckless.org e.V. event.
+We kindly thank [![genua GmbH](genua.png)](http://www.genua.eu) for
+borrowing us the equipment to record the slcon 2016 conference videos.
* [slcon 2015](2015)
* [slcon 2013](2013)