On Wed, Sep 25, 2013 at 01:11:22PM +0200, FRIGN wrote: > On 09/25/2013 02:26 PM, Martin Kopta wrote: > > Hi, I will be giving a talk about suckless project on 6th October [1] for > > small > > audience (from 10 to 100 people). Talk description is (translated): > > > > "Introduction to ideology of software that smells less and presentation of > > results of the suckless project. Demostration of dwm window manager, > > terminal > > emulator st, statically linked distribution stali, minimalist web browser > > surf > > and a few others. This talk is particularly suitable for minimalist hackers > > who crave efficient work environment." > > > > Is there something I absolutely shouldn't forget to mention? > > > > I will post here my presentation slides and maybe even some people questions > > after the talk. > > > > Thanks. > > > > [1] > > https://www.google.com/calendar/render?eid=bmJzYzU2dmo3Nm1ybjYzM2ExOWNrbGR0aGsgaHJ1c2Vja3kubmV0XzRwYWExYTNiY2hwdTZsOWVlczQxZ2c2dHQwQGc&ctz=Europe/Prague&sf=true&output=xml > > > > > > Hi Martin, > > in the interest of giving a clear presentation of what stali is about, > you could clear up common misconceptions about static linking and show > why it is better today and why it was disadvantageous in the past. > The key points are: > > * static linking in general * > _security_: The advantage of "higher security" with > dynamically-linked libraries is a fallacy. Today we're facing > versioned symbols and ldd-exploits on the one hand and blazing > fast computers able to recompile binaries in a flash on the > other. > > _speed_: Which one is faster? Refer to the material linked on > the stali-page. Don't forget to talk about the dynamic linker > in the Linux-Kernel, which has spread like cancer across other > Kernel sub-projects. > Also, don't forget to point out we are planning to use > non-bloated and segmented static-libraries for stali, which > allow the binaries to be leaner and faster in most cases. > > Make sure to be well-prepared in regard to this highly > controversial topic. > If you plan to introduce your audience to the concepts of > stali, there sure will be many hackers wondering why static > linking should be better than the "dynamic- > linking" de facto standard. > > * stali-project * > Generally speaking, the pages about stali are good enough. > However, make sure to bring these points in: > > _filesystem_: Deprecation of the /usr-dir[1] and why. Reasons > for building in a chroot-environment and how it works (broadly) > > _init-system_: Another big point. As it's still in concept- > phase, you could present the ideas we want to bring in. > An excellent information would be the talk about runit by > Christian Neukirchen[2]. > > * dwm, st * > No problem. For detailed matters, I'd need to see the slides > > * surf * > Not problematic, either. However, it could be interesting to > reflect on the future plans when it comes to the engine[3] > > > Generally speaking, I'm glad to hear from people like you who are > presenting and thus popularizing our ideals on software development. > > With best regards > > FRIGN > > [1]: <http://www.tldp.org/LDP/Linux-Filesystem-Hierarchy/html/usr.html> > [2]: <http://www.youtube.com/watch?v=wIziZlFHshE> > [3]: <http://lists.suckless.org/dev/1309/17648.html>
Thank you very much for your hints! I will surely mention all of the above in my talk. Good idea to prepare myself to defend static linking :-)
